diff --git "a/exp/log/log-train-2023-03-25-21-42-34-4" "b/exp/log/log-train-2023-03-25-21-42-34-4" new file mode 100644--- /dev/null +++ "b/exp/log/log-train-2023-03-25-21-42-34-4" @@ -0,0 +1,23260 @@ +2023-03-25 21:42:34,511 INFO [finetune.py:1046] (4/7) Training started +2023-03-25 21:42:34,511 INFO [finetune.py:1056] (4/7) Device: cuda:4 +2023-03-25 21:42:34,514 INFO [finetune.py:1065] (4/7) {'frame_shift_ms': 10.0, 'allowed_excess_duration_ratio': 0.1, 'best_train_loss': inf, 'best_valid_loss': inf, 'best_train_epoch': -1, 'best_valid_epoch': -1, 'batch_idx_train': 0, 'log_interval': 50, 'reset_interval': 200, 'valid_interval': 3000, 'feature_dim': 80, 'subsampling_factor': 4, 'warm_step': 2000, 'env_info': {'k2-version': '1.23.4', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '62e404dd3f3a811d73e424199b3408e309c06e1a', 'k2-git-date': 'Mon Jan 30 02:26:16 2023', 'lhotse-version': '1.12.0.dev+git.3ccfeb7.clean', 'torch-version': '1.13.0', 'torch-cuda-available': True, 'torch-cuda-version': '11.7', 'python-version': '3.8', 'icefall-git-branch': 'master', 'icefall-git-sha1': 'd74822d-dirty', 'icefall-git-date': 'Tue Mar 21 21:35:32 2023', 'icefall-path': '/home/lishaojie/icefall', 'k2-path': '/home/lishaojie/.conda/envs/env_lishaojie/lib/python3.8/site-packages/k2/__init__.py', 'lhotse-path': '/home/lishaojie/.conda/envs/env_lishaojie/lib/python3.8/site-packages/lhotse/__init__.py', 'hostname': 'cnc533', 'IP address': '127.0.1.1'}, 'world_size': 7, 'master_port': 18181, 'tensorboard': True, 'num_epochs': 30, 'start_epoch': 1, 'start_batch': 0, 'exp_dir': PosixPath('pruned_transducer_stateless7_streaming/exp1'), 'bpe_model': 'data/lang_bpe_500/bpe.model', 'base_lr': 0.004, 'lr_batches': 100000.0, 'lr_epochs': 100.0, 'context_size': 2, 'prune_range': 5, 'lm_scale': 0.25, 'am_scale': 0.0, 'simple_loss_scale': 0.5, 'seed': 42, 'print_diagnostics': False, 'inf_check': False, 'save_every_n': 2000, 'keep_last_k': 30, 'average_period': 200, 'use_fp16': True, 'num_encoder_layers': '2,4,3,2,4', 'feedforward_dims': '1024,1024,2048,2048,1024', 'nhead': '8,8,8,8,8', 'encoder_dims': '384,384,384,384,384', 'attention_dims': '192,192,192,192,192', 'encoder_unmasked_dims': '256,256,256,256,256', 'zipformer_downsampling_factors': '1,2,4,8,2', 'cnn_module_kernels': '31,31,31,31,31', 'decoder_dim': 512, 'joiner_dim': 512, 'do_finetune': True, 'init_modules': 'encoder', 'finetune_ckpt': '/home/lishaojie/icefall/egs/commonvoice/ASR/pruned_transducer_stateless7_streaming/exp/english_pretrain/epoch-30.pt', 'manifest_dir': PosixPath('data/fbank'), 'max_duration': 200, 'bucketing_sampler': True, 'num_buckets': 30, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': True, 'drop_last': True, 'return_cuts': True, 'num_workers': 2, 'enable_spec_aug': True, 'spec_aug_time_warp_factor': 80, 'enable_musan': True, 'input_strategy': 'PrecomputedFeatures', 'blank_id': 0, 'vocab_size': 500} +2023-03-25 21:42:34,514 INFO [finetune.py:1067] (4/7) About to create model +2023-03-25 21:42:34,862 INFO [zipformer.py:405] (4/7) At encoder stack 4, which has downsampling_factor=2, we will combine the outputs of layers 1 and 3, with downsampling_factors=2 and 8. +2023-03-25 21:42:34,871 INFO [finetune.py:1071] (4/7) Number of model parameters: 70369391 +2023-03-25 21:42:34,871 INFO [finetune.py:626] (4/7) Loading checkpoint from /home/lishaojie/icefall/egs/commonvoice/ASR/pruned_transducer_stateless7_streaming/exp/english_pretrain/epoch-30.pt +2023-03-25 21:42:35,501 INFO [finetune.py:647] (4/7) Loading parameters starting with prefix encoder +2023-03-25 21:42:36,945 INFO [finetune.py:1093] (4/7) Using DDP +2023-03-25 21:42:37,686 INFO [commonvoice_fr.py:392] (4/7) About to get train cuts +2023-03-25 21:42:37,689 INFO [commonvoice_fr.py:218] (4/7) Enable MUSAN +2023-03-25 21:42:37,689 INFO [commonvoice_fr.py:219] (4/7) About to get Musan cuts +2023-03-25 21:42:39,413 INFO [commonvoice_fr.py:243] (4/7) Enable SpecAugment +2023-03-25 21:42:39,413 INFO [commonvoice_fr.py:244] (4/7) Time warp factor: 80 +2023-03-25 21:42:39,414 INFO [commonvoice_fr.py:254] (4/7) Num frame mask: 10 +2023-03-25 21:42:39,414 INFO [commonvoice_fr.py:267] (4/7) About to create train dataset +2023-03-25 21:42:39,414 INFO [commonvoice_fr.py:294] (4/7) Using DynamicBucketingSampler. +2023-03-25 21:42:42,294 INFO [commonvoice_fr.py:309] (4/7) About to create train dataloader +2023-03-25 21:42:42,294 INFO [commonvoice_fr.py:399] (4/7) About to get dev cuts +2023-03-25 21:42:42,295 INFO [commonvoice_fr.py:340] (4/7) About to create dev dataset +2023-03-25 21:42:42,700 INFO [commonvoice_fr.py:357] (4/7) About to create dev dataloader +2023-03-25 21:42:42,700 INFO [finetune.py:1289] (4/7) Sanity check -- see if any of the batches in epoch 1 would cause OOM. +2023-03-25 21:46:46,134 INFO [finetune.py:1317] (4/7) Maximum memory allocated so far is 5247MB +2023-03-25 21:46:46,827 INFO [finetune.py:1317] (4/7) Maximum memory allocated so far is 5735MB +2023-03-25 21:46:48,051 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=96.36 vs. limit=5.0 +2023-03-25 21:46:48,914 INFO [finetune.py:1317] (4/7) Maximum memory allocated so far is 5735MB +2023-03-25 21:46:49,575 INFO [finetune.py:1317] (4/7) Maximum memory allocated so far is 5735MB +2023-03-25 21:46:50,266 INFO [finetune.py:1317] (4/7) Maximum memory allocated so far is 5735MB +2023-03-25 21:46:50,961 INFO [finetune.py:1317] (4/7) Maximum memory allocated so far is 5735MB +2023-03-25 21:46:59,843 INFO [finetune.py:976] (4/7) Epoch 1, batch 0, loss[loss=7.521, simple_loss=6.826, pruned_loss=6.933, over 4904.00 frames. ], tot_loss[loss=7.521, simple_loss=6.826, pruned_loss=6.933, over 4904.00 frames. ], batch size: 38, lr: 2.00e-03, grad_scale: 2.0 +2023-03-25 21:46:59,843 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-25 21:47:14,955 INFO [finetune.py:1010] (4/7) Epoch 1, validation: loss=7.294, simple_loss=6.606, pruned_loss=6.863, over 2265189.00 frames. +2023-03-25 21:47:14,956 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 5735MB +2023-03-25 21:47:19,866 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-25 21:47:26,752 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=21.15 vs. limit=5.0 +2023-03-25 21:47:30,297 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 21:47:42,791 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.6825, 3.2619, 3.6061, 3.5960, 2.7788, 2.9176, 3.6304, 1.2620], + device='cuda:4'), covar=tensor([0.2014, 0.3528, 0.2058, 0.1831, 0.3112, 0.2226, 0.1961, 0.5159], + device='cuda:4'), in_proj_covar=tensor([0.0366, 0.0239, 0.0251, 0.0290, 0.0341, 0.0278, 0.0307, 0.0298], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 21:48:00,212 INFO [finetune.py:976] (4/7) Epoch 1, batch 50, loss[loss=2.619, simple_loss=2.48, pruned_loss=1.381, over 4909.00 frames. ], tot_loss[loss=4.252, simple_loss=3.819, pruned_loss=4.148, over 214893.73 frames. ], batch size: 46, lr: 2.20e-03, grad_scale: 0.000244140625 +2023-03-25 21:48:12,369 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=43.31 vs. limit=5.0 +2023-03-25 21:48:32,996 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 21:48:42,905 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6649, 2.1074, 1.7511, 2.3216, 3.0049, 3.1771, 1.9797, 1.7674], + device='cuda:4'), covar=tensor([0.0239, 0.0226, 0.0230, 0.0159, 0.0215, 0.0133, 0.0240, 0.0193], + device='cuda:4'), in_proj_covar=tensor([0.0257, 0.0239, 0.0229, 0.0209, 0.0272, 0.0204, 0.0235, 0.0209], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 21:48:53,451 WARNING [finetune.py:966] (4/7) Grad scale is small: 0.000244140625 +2023-03-25 21:48:53,451 INFO [finetune.py:976] (4/7) Epoch 1, batch 100, loss[loss=3.406, simple_loss=3.237, pruned_loss=1.721, over 4796.00 frames. ], tot_loss[loss=3.441, simple_loss=3.171, pruned_loss=2.629, over 376935.06 frames. ], batch size: 51, lr: 2.40e-03, grad_scale: 0.00048828125 +2023-03-25 21:49:07,289 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=32.24 vs. limit=5.0 +2023-03-25 21:49:13,204 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 7.539e+02 2.791e+03 6.484e+03 1.700e+04 1.722e+07, threshold=1.297e+04, percent-clipped=0.0 +2023-03-25 21:49:27,832 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8288, 2.1614, 2.8524, 0.6671, 2.7541, 2.3918, 1.7431, 2.9528], + device='cuda:4'), covar=tensor([0.0238, 0.0279, 0.0281, 0.0590, 0.0333, 0.0374, 0.0638, 0.0222], + device='cuda:4'), in_proj_covar=tensor([0.0175, 0.0188, 0.0210, 0.0189, 0.0211, 0.0205, 0.0218, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 21:49:28,872 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=144.0, num_to_drop=2, layers_to_drop={1, 3} +2023-03-25 21:49:37,440 INFO [finetune.py:976] (4/7) Epoch 1, batch 150, loss[loss=1.57, simple_loss=1.426, pruned_loss=1.174, over 4843.00 frames. ], tot_loss[loss=2.834, simple_loss=2.618, pruned_loss=2.054, over 506320.52 frames. ], batch size: 47, lr: 2.60e-03, grad_scale: 0.00048828125 +2023-03-25 21:49:49,133 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=18.94 vs. limit=5.0 +2023-03-25 21:50:01,667 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=2.21 vs. limit=2.0 +2023-03-25 21:50:15,717 WARNING [finetune.py:966] (4/7) Grad scale is small: 0.00048828125 +2023-03-25 21:50:15,717 INFO [finetune.py:976] (4/7) Epoch 1, batch 200, loss[loss=1.457, simple_loss=1.261, pruned_loss=1.365, over 4806.00 frames. ], tot_loss[loss=2.345, simple_loss=2.145, pruned_loss=1.774, over 608461.24 frames. ], batch size: 45, lr: 2.80e-03, grad_scale: 0.0009765625 +2023-03-25 21:50:29,330 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 2.018e+02 7.406e+02 1.293e+03 3.197e+03 6.754e+04, threshold=2.586e+03, percent-clipped=12.0 +2023-03-25 21:50:54,578 INFO [finetune.py:976] (4/7) Epoch 1, batch 250, loss[loss=1.489, simple_loss=1.27, pruned_loss=1.403, over 4923.00 frames. ], tot_loss[loss=2.032, simple_loss=1.836, pruned_loss=1.612, over 686093.48 frames. ], batch size: 38, lr: 3.00e-03, grad_scale: 0.0009765625 +2023-03-25 21:51:43,808 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=296.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 21:51:45,811 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=300.0, num_to_drop=2, layers_to_drop={0, 3} +2023-03-25 21:51:46,256 WARNING [finetune.py:966] (4/7) Grad scale is small: 0.0009765625 +2023-03-25 21:51:46,256 INFO [finetune.py:976] (4/7) Epoch 1, batch 300, loss[loss=1.234, simple_loss=1.035, pruned_loss=1.174, over 4773.00 frames. ], tot_loss[loss=1.832, simple_loss=1.634, pruned_loss=1.51, over 747264.60 frames. ], batch size: 28, lr: 3.20e-03, grad_scale: 0.001953125 +2023-03-25 21:51:49,975 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=3.68 vs. limit=2.0 +2023-03-25 21:51:58,578 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 2.075e+01 5.781e+01 1.827e+02 5.788e+02 1.230e+04, threshold=3.655e+02, percent-clipped=4.0 +2023-03-25 21:52:39,153 INFO [finetune.py:976] (4/7) Epoch 1, batch 350, loss[loss=1.351, simple_loss=1.121, pruned_loss=1.274, over 4814.00 frames. ], tot_loss[loss=1.69, simple_loss=1.488, pruned_loss=1.433, over 792538.57 frames. ], batch size: 39, lr: 3.40e-03, grad_scale: 0.001953125 +2023-03-25 21:52:47,113 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=357.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-25 21:52:47,362 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=2.71 vs. limit=2.0 +2023-03-25 21:53:10,160 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=387.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 21:53:16,125 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.5915, 3.2947, 4.2465, 4.1450, 3.1492, 3.2508, 3.8313, 1.6269], + device='cuda:4'), covar=tensor([0.8776, 1.6533, 0.4369, 0.3403, 1.7748, 2.0162, 1.2145, 1.4417], + device='cuda:4'), in_proj_covar=tensor([0.0365, 0.0238, 0.0251, 0.0290, 0.0340, 0.0277, 0.0306, 0.0297], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 21:53:27,704 WARNING [finetune.py:966] (4/7) Grad scale is small: 0.001953125 +2023-03-25 21:53:27,704 INFO [finetune.py:976] (4/7) Epoch 1, batch 400, loss[loss=1.231, simple_loss=1.007, pruned_loss=1.158, over 4834.00 frames. ], tot_loss[loss=1.58, simple_loss=1.373, pruned_loss=1.368, over 829272.44 frames. ], batch size: 44, lr: 3.60e-03, grad_scale: 0.00390625 +2023-03-25 21:53:39,884 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.702e+01 2.277e+01 3.517e+01 1.113e+02 1.032e+03, threshold=7.035e+01, percent-clipped=3.0 +2023-03-25 21:53:40,553 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=2.45 vs. limit=2.0 +2023-03-25 21:53:51,259 WARNING [optim.py:389] (4/7) Scaling gradients by 0.06621765345335007, model_norm_threshold=70.34587860107422 +2023-03-25 21:53:51,343 INFO [optim.py:451] (4/7) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.67, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=7.539e+05, grad_sumsq = 2.933e+06, orig_rms_sq=2.571e-01 +2023-03-25 21:54:00,702 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=439.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 21:54:00,709 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6827, 1.3871, 2.0284, 3.7458, 2.7127, 2.6283, 0.6916, 2.6419], + device='cuda:4'), covar=tensor([0.0903, 0.1015, 0.0856, 0.0468, 0.0465, 0.0990, 0.1106, 0.0524], + device='cuda:4'), in_proj_covar=tensor([0.0104, 0.0120, 0.0138, 0.0159, 0.0109, 0.0145, 0.0130, 0.0116], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-25 21:54:05,302 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=448.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-25 21:54:06,748 INFO [finetune.py:976] (4/7) Epoch 1, batch 450, loss[loss=1.104, simple_loss=0.8876, pruned_loss=1.046, over 4801.00 frames. ], tot_loss[loss=1.475, simple_loss=1.266, pruned_loss=1.3, over 857592.73 frames. ], batch size: 29, lr: 3.80e-03, grad_scale: 0.00390625 +2023-03-25 21:54:43,454 WARNING [finetune.py:966] (4/7) Grad scale is small: 0.00390625 +2023-03-25 21:54:43,455 INFO [finetune.py:976] (4/7) Epoch 1, batch 500, loss[loss=1.082, simple_loss=0.864, pruned_loss=1.003, over 4924.00 frames. ], tot_loss[loss=1.376, simple_loss=1.166, pruned_loss=1.226, over 879677.80 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 0.0078125 +2023-03-25 21:54:57,599 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.430e+01 1.676e+01 1.950e+01 4.114e+01 1.062e+03, threshold=3.899e+01, percent-clipped=11.0 +2023-03-25 21:55:02,885 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=23.77 vs. limit=5.0 +2023-03-25 21:55:28,670 INFO [finetune.py:976] (4/7) Epoch 1, batch 550, loss[loss=0.973, simple_loss=0.7683, pruned_loss=0.8922, over 4897.00 frames. ], tot_loss[loss=1.289, simple_loss=1.08, pruned_loss=1.155, over 896980.74 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 0.0078125 +2023-03-25 21:55:39,561 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=562.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 21:55:42,629 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=568.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 21:55:58,116 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=590.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 21:56:09,550 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=600.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 21:56:09,984 WARNING [finetune.py:966] (4/7) Grad scale is small: 0.0078125 +2023-03-25 21:56:09,984 INFO [finetune.py:976] (4/7) Epoch 1, batch 600, loss[loss=1.125, simple_loss=0.8817, pruned_loss=1.014, over 4835.00 frames. ], tot_loss[loss=1.226, simple_loss=1.015, pruned_loss=1.102, over 910374.19 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 0.015625 +2023-03-25 21:56:22,614 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.472e+01 1.758e+01 2.024e+01 2.271e+01 8.528e+01, threshold=4.048e+01, percent-clipped=5.0 +2023-03-25 21:56:27,287 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=623.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-25 21:56:36,100 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=629.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-25 21:56:48,207 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=50.92 vs. limit=5.0 +2023-03-25 21:56:54,358 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=648.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 21:56:55,845 INFO [finetune.py:976] (4/7) Epoch 1, batch 650, loss[loss=1.056, simple_loss=0.8238, pruned_loss=0.932, over 4843.00 frames. ], tot_loss[loss=1.189, simple_loss=0.973, pruned_loss=1.067, over 920221.84 frames. ], batch size: 44, lr: 4.00e-03, grad_scale: 0.015625 +2023-03-25 21:56:55,935 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=651.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 21:56:56,432 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=652.0, num_to_drop=2, layers_to_drop={1, 3} +2023-03-25 21:57:00,749 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=2.79 vs. limit=2.0 +2023-03-25 21:57:31,099 INFO [finetune.py:976] (4/7) Epoch 1, batch 700, loss[loss=1.076, simple_loss=0.8427, pruned_loss=0.9164, over 4767.00 frames. ], tot_loss[loss=1.154, simple_loss=0.9336, pruned_loss=1.031, over 928418.11 frames. ], batch size: 59, lr: 4.00e-03, grad_scale: 0.03125 +2023-03-25 21:57:38,293 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.810e+01 2.037e+01 2.232e+01 2.628e+01 5.516e+01, threshold=4.463e+01, percent-clipped=4.0 +2023-03-25 21:57:59,237 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=739.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-25 21:58:01,228 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=743.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-25 21:58:05,816 INFO [finetune.py:976] (4/7) Epoch 1, batch 750, loss[loss=0.825, simple_loss=0.6233, pruned_loss=0.7278, over 4225.00 frames. ], tot_loss[loss=1.119, simple_loss=0.8966, pruned_loss=0.9952, over 933086.58 frames. ], batch size: 18, lr: 4.00e-03, grad_scale: 0.03125 +2023-03-25 21:58:29,192 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=787.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 21:58:36,509 INFO [finetune.py:976] (4/7) Epoch 1, batch 800, loss[loss=1.024, simple_loss=0.78, pruned_loss=0.8699, over 4815.00 frames. ], tot_loss[loss=1.093, simple_loss=0.8663, pruned_loss=0.9646, over 937971.24 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 0.0625 +2023-03-25 21:58:45,189 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 2.059e+01 2.266e+01 2.508e+01 2.744e+01 4.199e+01, threshold=5.016e+01, percent-clipped=0.0 +2023-03-25 21:59:19,867 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=2.05 vs. limit=2.0 +2023-03-25 21:59:20,572 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=847.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 21:59:22,545 INFO [finetune.py:976] (4/7) Epoch 1, batch 850, loss[loss=0.933, simple_loss=0.7076, pruned_loss=0.7783, over 4837.00 frames. ], tot_loss[loss=1.061, simple_loss=0.8334, pruned_loss=0.9294, over 942549.77 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 0.0625 +2023-03-25 21:59:41,265 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=2.26 vs. limit=2.0 +2023-03-25 22:00:12,170 INFO [finetune.py:976] (4/7) Epoch 1, batch 900, loss[loss=0.8717, simple_loss=0.6428, pruned_loss=0.7394, over 4741.00 frames. ], tot_loss[loss=1.031, simple_loss=0.8026, pruned_loss=0.8953, over 945913.09 frames. ], batch size: 27, lr: 4.00e-03, grad_scale: 0.125 +2023-03-25 22:00:16,310 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=908.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-25 22:00:25,565 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 2.101e+01 2.406e+01 2.575e+01 3.027e+01 5.726e+01, threshold=5.150e+01, percent-clipped=1.0 +2023-03-25 22:00:28,255 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=918.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-25 22:00:32,445 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=924.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-25 22:00:32,984 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6946, 1.9727, 1.5340, 2.4075, 2.1302, 2.0774, 1.5516, 1.4145], + device='cuda:4'), covar=tensor([0.1156, 0.0989, 0.1155, 0.0692, 0.0798, 0.0654, 0.1347, 0.1050], + device='cuda:4'), in_proj_covar=tensor([0.0257, 0.0239, 0.0229, 0.0209, 0.0272, 0.0204, 0.0235, 0.0209], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:00:53,629 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=946.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 22:00:56,170 INFO [finetune.py:976] (4/7) Epoch 1, batch 950, loss[loss=0.943, simple_loss=0.7001, pruned_loss=0.7753, over 4815.00 frames. ], tot_loss[loss=1.011, simple_loss=0.7801, pruned_loss=0.8692, over 948328.97 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 0.125 +2023-03-25 22:00:56,751 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=952.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-25 22:01:43,865 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1000.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:01:44,324 INFO [finetune.py:976] (4/7) Epoch 1, batch 1000, loss[loss=1.112, simple_loss=0.8183, pruned_loss=0.9048, over 4849.00 frames. ], tot_loss[loss=1.006, simple_loss=0.7687, pruned_loss=0.8564, over 948801.12 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 0.25 +2023-03-25 22:01:50,090 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=2.34 vs. limit=2.0 +2023-03-25 22:01:58,291 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 2.382e+01 2.890e+01 3.153e+01 3.664e+01 7.462e+01, threshold=6.306e+01, percent-clipped=2.0 +2023-03-25 22:02:21,751 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1043.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 22:02:31,288 INFO [finetune.py:976] (4/7) Epoch 1, batch 1050, loss[loss=1.015, simple_loss=0.7437, pruned_loss=0.8146, over 4816.00 frames. ], tot_loss[loss=1.008, simple_loss=0.7628, pruned_loss=0.8486, over 947931.98 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 0.25 +2023-03-25 22:02:48,367 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=63.50 vs. limit=5.0 +2023-03-25 22:03:07,640 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1091.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:03:18,239 INFO [finetune.py:976] (4/7) Epoch 1, batch 1100, loss[loss=0.8321, simple_loss=0.6053, pruned_loss=0.6607, over 4731.00 frames. ], tot_loss[loss=1.003, simple_loss=0.7533, pruned_loss=0.8354, over 948883.97 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 0.5 +2023-03-25 22:03:24,695 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-25 22:03:30,770 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 2.698e+01 3.337e+01 3.640e+01 4.251e+01 7.174e+01, threshold=7.279e+01, percent-clipped=4.0 +2023-03-25 22:04:04,831 INFO [finetune.py:976] (4/7) Epoch 1, batch 1150, loss[loss=0.9911, simple_loss=0.7252, pruned_loss=0.7679, over 4893.00 frames. ], tot_loss[loss=0.9978, simple_loss=0.7443, pruned_loss=0.8204, over 949100.47 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 0.5 +2023-03-25 22:04:06,016 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1153.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:04:43,459 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=5.42 vs. limit=5.0 +2023-03-25 22:04:45,292 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=24.51 vs. limit=5.0 +2023-03-25 22:04:46,447 INFO [finetune.py:976] (4/7) Epoch 1, batch 1200, loss[loss=0.9172, simple_loss=0.6758, pruned_loss=0.6938, over 4810.00 frames. ], tot_loss[loss=0.9839, simple_loss=0.7306, pruned_loss=0.7982, over 950857.58 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 1.0 +2023-03-25 22:04:47,540 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1203.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-25 22:04:59,185 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 3.248e+01 4.460e+01 5.563e+01 6.854e+01 1.013e+02, threshold=1.113e+02, percent-clipped=20.0 +2023-03-25 22:04:59,303 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1214.0, num_to_drop=2, layers_to_drop={2, 3} +2023-03-25 22:05:01,660 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1218.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 22:05:02,360 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=5.65 vs. limit=5.0 +2023-03-25 22:05:06,567 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1224.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:05:24,114 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1246.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:05:26,685 INFO [finetune.py:976] (4/7) Epoch 1, batch 1250, loss[loss=0.9534, simple_loss=0.7035, pruned_loss=0.7089, over 4828.00 frames. ], tot_loss[loss=0.9638, simple_loss=0.7142, pruned_loss=0.7699, over 951837.27 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 1.0 +2023-03-25 22:05:46,032 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1266.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:05:49,136 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1272.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:06:08,475 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1294.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:06:14,925 INFO [finetune.py:976] (4/7) Epoch 1, batch 1300, loss[loss=0.8994, simple_loss=0.6732, pruned_loss=0.6494, over 4756.00 frames. ], tot_loss[loss=0.9399, simple_loss=0.6968, pruned_loss=0.7383, over 952079.50 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 1.0 +2023-03-25 22:06:18,912 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=34.98 vs. limit=5.0 +2023-03-25 22:06:19,127 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-25 22:06:23,505 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 5.599e+01 8.403e+01 9.999e+01 1.262e+02 2.600e+02, threshold=2.000e+02, percent-clipped=40.0 +2023-03-25 22:06:57,098 INFO [finetune.py:976] (4/7) Epoch 1, batch 1350, loss[loss=0.7519, simple_loss=0.5703, pruned_loss=0.5285, over 4809.00 frames. ], tot_loss[loss=0.9211, simple_loss=0.6845, pruned_loss=0.7101, over 953948.48 frames. ], batch size: 25, lr: 4.00e-03, grad_scale: 1.0 +2023-03-25 22:07:21,103 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.59 vs. limit=2.0 +2023-03-25 22:07:50,288 INFO [finetune.py:976] (4/7) Epoch 1, batch 1400, loss[loss=0.8206, simple_loss=0.6255, pruned_loss=0.5671, over 4765.00 frames. ], tot_loss[loss=0.9067, simple_loss=0.6767, pruned_loss=0.6853, over 955706.42 frames. ], batch size: 27, lr: 4.00e-03, grad_scale: 1.0 +2023-03-25 22:07:58,272 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.277e+01 1.400e+02 1.610e+02 1.980e+02 2.974e+02, threshold=3.221e+02, percent-clipped=23.0 +2023-03-25 22:08:09,181 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1434.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:08:20,139 INFO [finetune.py:976] (4/7) Epoch 1, batch 1450, loss[loss=0.8104, simple_loss=0.6114, pruned_loss=0.5589, over 4873.00 frames. ], tot_loss[loss=0.8846, simple_loss=0.664, pruned_loss=0.6554, over 956216.53 frames. ], batch size: 31, lr: 4.00e-03, grad_scale: 1.0 +2023-03-25 22:08:47,668 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1495.0, num_to_drop=2, layers_to_drop={1, 3} +2023-03-25 22:08:51,786 INFO [finetune.py:976] (4/7) Epoch 1, batch 1500, loss[loss=0.7057, simple_loss=0.5614, pruned_loss=0.4577, over 4819.00 frames. ], tot_loss[loss=0.8537, simple_loss=0.6459, pruned_loss=0.6197, over 956895.87 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 1.0 +2023-03-25 22:08:52,965 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1503.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:09:02,827 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1509.0, num_to_drop=2, layers_to_drop={2, 3} +2023-03-25 22:09:05,946 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.968e+01 1.844e+02 2.293e+02 2.711e+02 4.587e+02, threshold=4.586e+02, percent-clipped=13.0 +2023-03-25 22:09:17,181 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-25 22:09:42,480 INFO [finetune.py:976] (4/7) Epoch 1, batch 1550, loss[loss=0.6499, simple_loss=0.5268, pruned_loss=0.4103, over 4774.00 frames. ], tot_loss[loss=0.8217, simple_loss=0.6275, pruned_loss=0.5841, over 956835.98 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 1.0 +2023-03-25 22:09:42,538 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1551.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:09:52,836 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1566.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:10:33,765 INFO [finetune.py:976] (4/7) Epoch 1, batch 1600, loss[loss=0.6459, simple_loss=0.5276, pruned_loss=0.4016, over 4799.00 frames. ], tot_loss[loss=0.7868, simple_loss=0.6068, pruned_loss=0.5478, over 957515.11 frames. ], batch size: 29, lr: 4.00e-03, grad_scale: 2.0 +2023-03-25 22:10:40,878 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1611.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:10:42,942 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.367e+02 1.965e+02 2.441e+02 2.819e+02 5.041e+02, threshold=4.882e+02, percent-clipped=1.0 +2023-03-25 22:10:52,756 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2090, 1.9834, 1.3923, 1.2283, 1.9946, 2.5109, 2.3145, 1.8009], + device='cuda:4'), covar=tensor([0.0108, 0.0175, 0.0430, 0.0302, 0.0160, 0.0133, 0.0106, 0.0227], + device='cuda:4'), in_proj_covar=tensor([0.0092, 0.0120, 0.0148, 0.0116, 0.0111, 0.0115, 0.0095, 0.0120], + device='cuda:4'), out_proj_covar=tensor([7.2287e-05, 9.4712e-05, 1.2103e-04, 9.1564e-05, 8.8032e-05, 8.6713e-05, + 7.3578e-05, 9.4637e-05], device='cuda:4') +2023-03-25 22:10:55,976 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1627.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-25 22:11:18,727 INFO [finetune.py:976] (4/7) Epoch 1, batch 1650, loss[loss=0.5723, simple_loss=0.4843, pruned_loss=0.3415, over 4906.00 frames. ], tot_loss[loss=0.7513, simple_loss=0.5855, pruned_loss=0.5126, over 958713.35 frames. ], batch size: 36, lr: 4.00e-03, grad_scale: 2.0 +2023-03-25 22:11:41,983 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1672.0, num_to_drop=2, layers_to_drop={0, 3} +2023-03-25 22:11:48,527 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3490, 1.4857, 1.1744, 1.4175, 1.4650, 1.3108, 2.2388, 1.3739], + device='cuda:4'), covar=tensor([0.5534, 0.9299, 0.8106, 0.8034, 0.5623, 0.4831, 0.2969, 0.9333], + device='cuda:4'), in_proj_covar=tensor([0.0179, 0.0209, 0.0246, 0.0272, 0.0228, 0.0191, 0.0195, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:11:50,174 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9510, 1.5822, 2.6235, 3.9878, 2.9161, 2.5392, 0.9120, 3.1766], + device='cuda:4'), covar=tensor([0.2503, 0.2045, 0.1477, 0.0439, 0.1088, 0.1558, 0.2440, 0.0921], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0116, 0.0129, 0.0150, 0.0104, 0.0137, 0.0122, 0.0108], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-25 22:12:02,515 INFO [finetune.py:976] (4/7) Epoch 1, batch 1700, loss[loss=0.6506, simple_loss=0.5347, pruned_loss=0.3966, over 4822.00 frames. ], tot_loss[loss=0.7192, simple_loss=0.5663, pruned_loss=0.4812, over 957155.54 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 2.0 +2023-03-25 22:12:14,554 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.227e+02 2.187e+02 2.736e+02 3.197e+02 8.210e+02, threshold=5.471e+02, percent-clipped=2.0 +2023-03-25 22:12:31,386 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=5.34 vs. limit=5.0 +2023-03-25 22:12:53,682 INFO [finetune.py:976] (4/7) Epoch 1, batch 1750, loss[loss=0.6793, simple_loss=0.5519, pruned_loss=0.4155, over 4183.00 frames. ], tot_loss[loss=0.7004, simple_loss=0.5582, pruned_loss=0.4587, over 956947.96 frames. ], batch size: 66, lr: 4.00e-03, grad_scale: 2.0 +2023-03-25 22:13:24,638 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.2544, 1.3133, 1.3767, 0.3834, 1.2634, 1.6442, 1.4617, 1.3404], + device='cuda:4'), covar=tensor([0.0662, 0.0271, 0.0223, 0.0538, 0.0278, 0.0188, 0.0219, 0.0299], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0156, 0.0119, 0.0129, 0.0132, 0.0126, 0.0151, 0.0160], + device='cuda:4'), out_proj_covar=tensor([1.0029e-04, 1.1607e-04, 8.7273e-05, 9.4390e-05, 9.5948e-05, 9.3977e-05, + 1.1337e-04, 1.1964e-04], device='cuda:4') +2023-03-25 22:13:29,981 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1790.0, num_to_drop=1, layers_to_drop={3} +2023-03-25 22:13:32,189 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4700, 1.1970, 1.0767, 0.6484, 1.2420, 1.2830, 0.9439, 1.2616], + device='cuda:4'), covar=tensor([0.0491, 0.0908, 0.1031, 0.1671, 0.0866, 0.1088, 0.1879, 0.0876], + device='cuda:4'), in_proj_covar=tensor([0.0147, 0.0161, 0.0177, 0.0161, 0.0179, 0.0174, 0.0189, 0.0175], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:13:36,052 INFO [finetune.py:976] (4/7) Epoch 1, batch 1800, loss[loss=0.6237, simple_loss=0.5304, pruned_loss=0.3647, over 4819.00 frames. ], tot_loss[loss=0.6888, simple_loss=0.555, pruned_loss=0.4423, over 955977.20 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 2.0 +2023-03-25 22:13:40,474 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1809.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:13:43,023 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.190e+02 2.215e+02 2.629e+02 3.291e+02 5.990e+02, threshold=5.258e+02, percent-clipped=1.0 +2023-03-25 22:13:59,239 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1838.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:14:06,687 INFO [finetune.py:976] (4/7) Epoch 1, batch 1850, loss[loss=0.6003, simple_loss=0.5226, pruned_loss=0.3425, over 4869.00 frames. ], tot_loss[loss=0.6696, simple_loss=0.5456, pruned_loss=0.422, over 956358.83 frames. ], batch size: 34, lr: 4.00e-03, grad_scale: 2.0 +2023-03-25 22:14:10,080 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=1857.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:14:10,121 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1857.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:14:17,837 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1870.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:14:38,305 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-03-25 22:14:39,265 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4285, 1.1800, 1.4655, 1.0800, 1.4346, 1.6349, 1.1193, 1.8191], + device='cuda:4'), covar=tensor([0.0890, 0.1407, 0.0936, 0.1149, 0.0580, 0.0712, 0.2012, 0.0532], + device='cuda:4'), in_proj_covar=tensor([0.0180, 0.0198, 0.0190, 0.0178, 0.0157, 0.0196, 0.0203, 0.0176], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:14:51,530 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1899.0, num_to_drop=2, layers_to_drop={0, 3} +2023-03-25 22:14:52,529 INFO [finetune.py:976] (4/7) Epoch 1, batch 1900, loss[loss=0.549, simple_loss=0.4995, pruned_loss=0.3003, over 4820.00 frames. ], tot_loss[loss=0.652, simple_loss=0.537, pruned_loss=0.4039, over 955252.85 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 2.0 +2023-03-25 22:15:03,950 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.381e+02 2.208e+02 2.560e+02 3.227e+02 6.450e+02, threshold=5.121e+02, percent-clipped=1.0 +2023-03-25 22:15:11,549 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1918.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-25 22:15:14,238 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1922.0, num_to_drop=1, layers_to_drop={2} +2023-03-25 22:15:20,180 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1931.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 22:15:36,617 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.7068, 1.7652, 1.7726, 0.6774, 1.8192, 2.2093, 1.7385, 1.6234], + device='cuda:4'), covar=tensor([0.0813, 0.0324, 0.0295, 0.0665, 0.0250, 0.0174, 0.0320, 0.0304], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0157, 0.0120, 0.0130, 0.0133, 0.0126, 0.0152, 0.0160], + device='cuda:4'), out_proj_covar=tensor([1.0124e-04, 1.1708e-04, 8.7950e-05, 9.5378e-05, 9.6338e-05, 9.4152e-05, + 1.1406e-04, 1.1976e-04], device='cuda:4') +2023-03-25 22:15:37,064 INFO [finetune.py:976] (4/7) Epoch 1, batch 1950, loss[loss=0.5734, simple_loss=0.493, pruned_loss=0.3281, over 4916.00 frames. ], tot_loss[loss=0.6299, simple_loss=0.5245, pruned_loss=0.3839, over 954783.84 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 2.0 +2023-03-25 22:15:44,362 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.7377, 3.2497, 3.2504, 3.6125, 3.3898, 3.3164, 3.8816, 1.2754], + device='cuda:4'), covar=tensor([0.1304, 0.1312, 0.1139, 0.1454, 0.2233, 0.1657, 0.1167, 0.6230], + device='cuda:4'), in_proj_covar=tensor([0.0365, 0.0240, 0.0256, 0.0289, 0.0344, 0.0283, 0.0305, 0.0299], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:15:46,010 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1967.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:16:03,928 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=5.37 vs. limit=5.0 +2023-03-25 22:16:12,899 INFO [finetune.py:976] (4/7) Epoch 1, batch 2000, loss[loss=0.5145, simple_loss=0.4551, pruned_loss=0.2869, over 4792.00 frames. ], tot_loss[loss=0.607, simple_loss=0.51, pruned_loss=0.3648, over 954643.46 frames. ], batch size: 29, lr: 4.00e-03, grad_scale: 4.0 +2023-03-25 22:16:22,846 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.403e+02 2.183e+02 2.758e+02 3.285e+02 7.843e+02, threshold=5.515e+02, percent-clipped=1.0 +2023-03-25 22:16:57,270 INFO [finetune.py:976] (4/7) Epoch 1, batch 2050, loss[loss=0.5142, simple_loss=0.4532, pruned_loss=0.2876, over 4484.00 frames. ], tot_loss[loss=0.5809, simple_loss=0.4941, pruned_loss=0.3437, over 954025.11 frames. ], batch size: 20, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:17:31,385 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2090.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 22:17:41,737 INFO [finetune.py:976] (4/7) Epoch 1, batch 2100, loss[loss=0.4913, simple_loss=0.4379, pruned_loss=0.2724, over 4755.00 frames. ], tot_loss[loss=0.5624, simple_loss=0.4842, pruned_loss=0.328, over 954523.33 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:17:55,063 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.336e+02 2.022e+02 2.484e+02 2.961e+02 6.695e+02, threshold=4.968e+02, percent-clipped=1.0 +2023-03-25 22:18:12,540 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2138.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:18:12,645 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=5.00 vs. limit=5.0 +2023-03-25 22:18:22,670 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.1940, 4.4808, 4.6808, 5.0174, 4.8417, 4.6028, 5.2923, 1.5948], + device='cuda:4'), covar=tensor([0.0747, 0.0811, 0.0589, 0.0806, 0.1327, 0.1140, 0.0488, 0.5017], + device='cuda:4'), in_proj_covar=tensor([0.0366, 0.0241, 0.0256, 0.0290, 0.0346, 0.0285, 0.0306, 0.0300], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:18:29,607 INFO [finetune.py:976] (4/7) Epoch 1, batch 2150, loss[loss=0.4933, simple_loss=0.4659, pruned_loss=0.2603, over 4793.00 frames. ], tot_loss[loss=0.5527, simple_loss=0.4816, pruned_loss=0.3179, over 953370.13 frames. ], batch size: 29, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:18:36,847 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-03-25 22:18:44,335 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5488, 3.7839, 3.8180, 1.6061, 4.0828, 2.9782, 1.0370, 2.6350], + device='cuda:4'), covar=tensor([0.2028, 0.1182, 0.1131, 0.3081, 0.0741, 0.0728, 0.3450, 0.1153], + device='cuda:4'), in_proj_covar=tensor([0.0144, 0.0144, 0.0153, 0.0120, 0.0142, 0.0108, 0.0132, 0.0111], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:19:03,504 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2194.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:19:08,499 INFO [finetune.py:976] (4/7) Epoch 1, batch 2200, loss[loss=0.4319, simple_loss=0.404, pruned_loss=0.2299, over 4666.00 frames. ], tot_loss[loss=0.5446, simple_loss=0.4797, pruned_loss=0.3094, over 953754.78 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:19:12,555 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5764, 1.8086, 1.4501, 1.7372, 1.1009, 3.2562, 1.1379, 1.7860], + device='cuda:4'), covar=tensor([0.3421, 0.2359, 0.2201, 0.2137, 0.1911, 0.0243, 0.2609, 0.1308], + device='cuda:4'), in_proj_covar=tensor([0.0114, 0.0099, 0.0107, 0.0103, 0.0093, 0.0083, 0.0080, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-25 22:19:17,018 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2213.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:19:17,476 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.568e+02 2.355e+02 2.819e+02 3.325e+02 5.172e+02, threshold=5.637e+02, percent-clipped=1.0 +2023-03-25 22:19:22,622 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2222.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 22:19:28,124 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2226.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:19:30,421 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9487, 1.7445, 2.1772, 3.6534, 2.8425, 2.4771, 0.7549, 2.9702], + device='cuda:4'), covar=tensor([0.1675, 0.1465, 0.1308, 0.0427, 0.0788, 0.1520, 0.1944, 0.0624], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0112, 0.0126, 0.0147, 0.0100, 0.0135, 0.0118, 0.0103], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-25 22:19:49,077 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7074, 3.9546, 3.9777, 2.0903, 4.2087, 3.1303, 1.1505, 2.8976], + device='cuda:4'), covar=tensor([0.1966, 0.1046, 0.1090, 0.2652, 0.0671, 0.0698, 0.3504, 0.1057], + device='cuda:4'), in_proj_covar=tensor([0.0143, 0.0143, 0.0152, 0.0119, 0.0142, 0.0108, 0.0132, 0.0111], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:19:56,641 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=9.59 vs. limit=5.0 +2023-03-25 22:19:57,055 INFO [finetune.py:976] (4/7) Epoch 1, batch 2250, loss[loss=0.5776, simple_loss=0.5093, pruned_loss=0.323, over 4817.00 frames. ], tot_loss[loss=0.5364, simple_loss=0.4766, pruned_loss=0.3017, over 954505.57 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:20:18,371 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2267.0, num_to_drop=1, layers_to_drop={2} +2023-03-25 22:20:20,121 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2270.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:20:28,290 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8042, 1.5097, 0.9618, 1.6957, 1.9525, 1.4302, 1.3456, 1.8148], + device='cuda:4'), covar=tensor([0.1801, 0.1905, 0.2065, 0.1259, 0.2650, 0.1843, 0.1468, 0.1771], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0092, 0.0108, 0.0089, 0.0119, 0.0087, 0.0094, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-25 22:21:00,800 INFO [finetune.py:976] (4/7) Epoch 1, batch 2300, loss[loss=0.3848, simple_loss=0.3926, pruned_loss=0.1885, over 4751.00 frames. ], tot_loss[loss=0.5217, simple_loss=0.4693, pruned_loss=0.2899, over 955814.47 frames. ], batch size: 27, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:21:15,884 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.417e+02 2.050e+02 2.425e+02 2.921e+02 4.362e+02, threshold=4.850e+02, percent-clipped=0.0 +2023-03-25 22:21:21,995 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2315.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:21:28,809 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3334, 1.3853, 1.4146, 0.6160, 1.3748, 1.8732, 1.5294, 1.5345], + device='cuda:4'), covar=tensor([0.1053, 0.0501, 0.0460, 0.0856, 0.0357, 0.0250, 0.0340, 0.0476], + device='cuda:4'), in_proj_covar=tensor([0.0128, 0.0151, 0.0116, 0.0127, 0.0128, 0.0121, 0.0146, 0.0154], + device='cuda:4'), out_proj_covar=tensor([9.6785e-05, 1.1272e-04, 8.4803e-05, 9.3095e-05, 9.2578e-05, 8.9902e-05, + 1.0919e-04, 1.1494e-04], device='cuda:4') +2023-03-25 22:21:42,838 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2340.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:21:54,803 INFO [finetune.py:976] (4/7) Epoch 1, batch 2350, loss[loss=0.4464, simple_loss=0.4312, pruned_loss=0.2308, over 4739.00 frames. ], tot_loss[loss=0.5053, simple_loss=0.4589, pruned_loss=0.278, over 955410.94 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:22:38,925 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4915, 1.2596, 1.3797, 1.4928, 2.0044, 1.3936, 1.1236, 1.1504], + device='cuda:4'), covar=tensor([0.4498, 0.5010, 0.3964, 0.4113, 0.3569, 0.3071, 0.6099, 0.3714], + device='cuda:4'), in_proj_covar=tensor([0.0225, 0.0208, 0.0196, 0.0182, 0.0233, 0.0181, 0.0205, 0.0183], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:22:57,442 INFO [finetune.py:976] (4/7) Epoch 1, batch 2400, loss[loss=0.428, simple_loss=0.4081, pruned_loss=0.224, over 4767.00 frames. ], tot_loss[loss=0.4903, simple_loss=0.4489, pruned_loss=0.2675, over 954866.15 frames. ], batch size: 27, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:22:57,565 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2401.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-25 22:23:00,298 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4166, 1.7502, 1.7319, 0.8863, 1.9269, 1.8583, 1.3146, 1.8473], + device='cuda:4'), covar=tensor([0.0407, 0.1143, 0.1377, 0.2371, 0.0786, 0.1330, 0.2355, 0.1084], + device='cuda:4'), in_proj_covar=tensor([0.0144, 0.0159, 0.0173, 0.0159, 0.0174, 0.0170, 0.0185, 0.0171], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:23:05,866 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.408e+02 1.953e+02 2.427e+02 2.971e+02 6.309e+02, threshold=4.853e+02, percent-clipped=1.0 +2023-03-25 22:23:28,132 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7724, 1.3063, 2.0940, 1.3491, 1.8753, 1.9597, 1.3447, 2.1563], + device='cuda:4'), covar=tensor([0.1574, 0.2077, 0.1200, 0.2052, 0.0931, 0.1381, 0.2713, 0.0869], + device='cuda:4'), in_proj_covar=tensor([0.0183, 0.0196, 0.0192, 0.0180, 0.0159, 0.0200, 0.0204, 0.0177], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:23:32,014 INFO [finetune.py:976] (4/7) Epoch 1, batch 2450, loss[loss=0.4134, simple_loss=0.4053, pruned_loss=0.2107, over 4927.00 frames. ], tot_loss[loss=0.4768, simple_loss=0.4398, pruned_loss=0.2582, over 954466.95 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:23:49,467 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6420, 1.3487, 1.0517, 1.1452, 1.5677, 1.8724, 1.5712, 1.0589], + device='cuda:4'), covar=tensor([0.0236, 0.0479, 0.0784, 0.0499, 0.0292, 0.0245, 0.0393, 0.0499], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0119, 0.0142, 0.0116, 0.0109, 0.0109, 0.0093, 0.0118], + device='cuda:4'), out_proj_covar=tensor([7.0148e-05, 9.4186e-05, 1.1559e-04, 9.2339e-05, 8.6493e-05, 8.2032e-05, + 7.1925e-05, 9.2885e-05], device='cuda:4') +2023-03-25 22:24:03,840 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6321, 4.0338, 4.0799, 1.9642, 4.3197, 2.9924, 0.9620, 2.7157], + device='cuda:4'), covar=tensor([0.1869, 0.1037, 0.1071, 0.2537, 0.0661, 0.0766, 0.3414, 0.1160], + device='cuda:4'), in_proj_covar=tensor([0.0145, 0.0145, 0.0153, 0.0120, 0.0143, 0.0109, 0.0134, 0.0112], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:24:05,633 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.7048, 1.5798, 1.5965, 0.8306, 1.6506, 2.1940, 1.6200, 1.6688], + device='cuda:4'), covar=tensor([0.1103, 0.0596, 0.0562, 0.0822, 0.0385, 0.0263, 0.0437, 0.0576], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0153, 0.0117, 0.0128, 0.0129, 0.0122, 0.0148, 0.0156], + device='cuda:4'), out_proj_covar=tensor([9.7771e-05, 1.1416e-04, 8.5985e-05, 9.4006e-05, 9.3696e-05, 9.0775e-05, + 1.1057e-04, 1.1607e-04], device='cuda:4') +2023-03-25 22:24:21,668 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2494.0, num_to_drop=1, layers_to_drop={2} +2023-03-25 22:24:25,636 INFO [finetune.py:976] (4/7) Epoch 1, batch 2500, loss[loss=0.5704, simple_loss=0.517, pruned_loss=0.3119, over 4730.00 frames. ], tot_loss[loss=0.4733, simple_loss=0.4388, pruned_loss=0.255, over 953295.98 frames. ], batch size: 59, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:24:34,904 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2513.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:24:35,371 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.353e+02 2.241e+02 2.593e+02 3.079e+02 4.323e+02, threshold=5.185e+02, percent-clipped=0.0 +2023-03-25 22:24:44,448 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2526.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:24:54,543 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2542.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:24:57,929 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.6800, 4.0810, 4.2024, 4.4817, 4.3757, 4.1656, 4.7935, 1.6355], + device='cuda:4'), covar=tensor([0.0700, 0.0788, 0.0630, 0.0885, 0.1237, 0.1143, 0.0546, 0.4652], + device='cuda:4'), in_proj_covar=tensor([0.0372, 0.0245, 0.0262, 0.0296, 0.0351, 0.0289, 0.0311, 0.0304], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:25:00,195 INFO [finetune.py:976] (4/7) Epoch 1, batch 2550, loss[loss=0.5497, simple_loss=0.496, pruned_loss=0.3016, over 4286.00 frames. ], tot_loss[loss=0.475, simple_loss=0.4439, pruned_loss=0.2539, over 955466.67 frames. ], batch size: 65, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:25:05,759 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2987, 1.4425, 1.0924, 1.3725, 1.4951, 1.2260, 2.0385, 1.3522], + device='cuda:4'), covar=tensor([0.3424, 0.5806, 0.5448, 0.5961, 0.3442, 0.3246, 0.3115, 0.5378], + device='cuda:4'), in_proj_covar=tensor([0.0159, 0.0186, 0.0221, 0.0242, 0.0201, 0.0171, 0.0176, 0.0180], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0002], + device='cuda:4') +2023-03-25 22:25:09,688 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2561.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:25:18,673 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=2574.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:25:42,500 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-25 22:25:48,340 INFO [finetune.py:976] (4/7) Epoch 1, batch 2600, loss[loss=0.3996, simple_loss=0.4078, pruned_loss=0.1957, over 4765.00 frames. ], tot_loss[loss=0.4696, simple_loss=0.4419, pruned_loss=0.2493, over 955089.16 frames. ], batch size: 27, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:25:55,823 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.651e+02 2.205e+02 2.587e+02 2.996e+02 4.228e+02, threshold=5.174e+02, percent-clipped=0.0 +2023-03-25 22:26:06,645 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=5.44 vs. limit=5.0 +2023-03-25 22:26:19,956 INFO [finetune.py:976] (4/7) Epoch 1, batch 2650, loss[loss=0.435, simple_loss=0.4254, pruned_loss=0.2223, over 4831.00 frames. ], tot_loss[loss=0.4633, simple_loss=0.4392, pruned_loss=0.2442, over 954635.08 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:26:38,612 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.64 vs. limit=2.0 +2023-03-25 22:26:54,467 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2685.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:27:06,878 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2696.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:27:15,138 INFO [finetune.py:976] (4/7) Epoch 1, batch 2700, loss[loss=0.4665, simple_loss=0.4193, pruned_loss=0.2569, over 4105.00 frames. ], tot_loss[loss=0.4548, simple_loss=0.4343, pruned_loss=0.238, over 954176.77 frames. ], batch size: 65, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:27:28,255 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.346e+02 2.127e+02 2.493e+02 3.058e+02 5.200e+02, threshold=4.985e+02, percent-clipped=1.0 +2023-03-25 22:28:14,554 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2746.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-25 22:28:17,274 INFO [finetune.py:976] (4/7) Epoch 1, batch 2750, loss[loss=0.4199, simple_loss=0.4167, pruned_loss=0.2115, over 4824.00 frames. ], tot_loss[loss=0.4444, simple_loss=0.4267, pruned_loss=0.2313, over 954320.34 frames. ], batch size: 41, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:28:50,938 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5282, 1.9423, 1.9654, 0.8744, 2.0037, 1.9391, 1.5838, 2.0870], + device='cuda:4'), covar=tensor([0.0548, 0.0924, 0.1176, 0.2102, 0.0989, 0.1386, 0.1718, 0.0870], + device='cuda:4'), in_proj_covar=tensor([0.0142, 0.0157, 0.0171, 0.0156, 0.0172, 0.0170, 0.0180, 0.0168], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:28:58,717 INFO [finetune.py:976] (4/7) Epoch 1, batch 2800, loss[loss=0.3537, simple_loss=0.3664, pruned_loss=0.1705, over 4912.00 frames. ], tot_loss[loss=0.4341, simple_loss=0.4191, pruned_loss=0.2248, over 955989.91 frames. ], batch size: 32, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:29:06,644 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.489e+02 2.264e+02 2.537e+02 3.001e+02 5.007e+02, threshold=5.073e+02, percent-clipped=1.0 +2023-03-25 22:29:12,563 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2824.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:29:40,283 INFO [finetune.py:976] (4/7) Epoch 1, batch 2850, loss[loss=0.3788, simple_loss=0.3702, pruned_loss=0.1937, over 4720.00 frames. ], tot_loss[loss=0.4252, simple_loss=0.4126, pruned_loss=0.2191, over 953665.80 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:29:53,492 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3272, 1.4512, 1.3369, 1.4554, 0.8422, 2.7178, 0.8405, 1.5537], + device='cuda:4'), covar=tensor([0.4077, 0.2552, 0.2457, 0.2501, 0.2394, 0.0305, 0.2984, 0.1563], + device='cuda:4'), in_proj_covar=tensor([0.0116, 0.0100, 0.0107, 0.0105, 0.0096, 0.0085, 0.0082, 0.0081], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004], + device='cuda:4') +2023-03-25 22:30:03,461 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2885.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-25 22:30:15,603 INFO [finetune.py:976] (4/7) Epoch 1, batch 2900, loss[loss=0.3647, simple_loss=0.3814, pruned_loss=0.174, over 4823.00 frames. ], tot_loss[loss=0.428, simple_loss=0.4158, pruned_loss=0.2203, over 952608.70 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:30:21,056 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-25 22:30:23,164 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.426e+02 2.100e+02 2.461e+02 2.914e+02 6.574e+02, threshold=4.923e+02, percent-clipped=3.0 +2023-03-25 22:30:37,165 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2937.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:30:50,913 INFO [finetune.py:976] (4/7) Epoch 1, batch 2950, loss[loss=0.3974, simple_loss=0.4017, pruned_loss=0.1965, over 4776.00 frames. ], tot_loss[loss=0.4272, simple_loss=0.4175, pruned_loss=0.2185, over 951928.19 frames. ], batch size: 27, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:31:31,753 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2996.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:31:33,404 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2998.0, num_to_drop=2, layers_to_drop={2, 3} +2023-03-25 22:31:35,185 INFO [finetune.py:976] (4/7) Epoch 1, batch 3000, loss[loss=0.4474, simple_loss=0.4433, pruned_loss=0.2257, over 4911.00 frames. ], tot_loss[loss=0.4249, simple_loss=0.417, pruned_loss=0.2165, over 951458.83 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:31:35,185 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-25 22:31:56,384 INFO [finetune.py:1010] (4/7) Epoch 1, validation: loss=0.4228, simple_loss=0.4589, pruned_loss=0.1933, over 2265189.00 frames. +2023-03-25 22:31:56,384 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 5735MB +2023-03-25 22:32:03,994 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-25 22:32:17,072 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.364e+02 2.092e+02 2.490e+02 2.940e+02 5.162e+02, threshold=4.980e+02, percent-clipped=2.0 +2023-03-25 22:32:25,844 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3019.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:32:32,904 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2454, 3.6783, 3.7823, 4.1349, 3.9344, 3.7830, 4.3515, 1.3916], + device='cuda:4'), covar=tensor([0.0669, 0.0679, 0.0626, 0.0827, 0.1134, 0.1196, 0.0601, 0.4784], + device='cuda:4'), in_proj_covar=tensor([0.0370, 0.0244, 0.0262, 0.0294, 0.0347, 0.0288, 0.0310, 0.0301], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:32:39,245 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3041.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:32:40,986 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3044.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:32:45,001 INFO [finetune.py:976] (4/7) Epoch 1, batch 3050, loss[loss=0.4466, simple_loss=0.4305, pruned_loss=0.2314, over 4806.00 frames. ], tot_loss[loss=0.4242, simple_loss=0.4181, pruned_loss=0.2153, over 952908.68 frames. ], batch size: 55, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:33:09,814 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3080.0, num_to_drop=2, layers_to_drop={0, 3} +2023-03-25 22:33:16,459 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3246, 1.0903, 0.8981, 1.0321, 1.0641, 0.9375, 0.9324, 1.6091], + device='cuda:4'), covar=tensor([10.0803, 13.3670, 9.7667, 17.3186, 10.2973, 6.7188, 14.3871, 3.6323], + device='cuda:4'), in_proj_covar=tensor([0.0230, 0.0226, 0.0204, 0.0262, 0.0219, 0.0189, 0.0226, 0.0168], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-25 22:33:26,545 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=6.30 vs. limit=5.0 +2023-03-25 22:33:28,605 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.84 vs. limit=5.0 +2023-03-25 22:33:38,269 INFO [finetune.py:976] (4/7) Epoch 1, batch 3100, loss[loss=0.4536, simple_loss=0.4165, pruned_loss=0.2454, over 4836.00 frames. ], tot_loss[loss=0.4157, simple_loss=0.4119, pruned_loss=0.2098, over 954371.55 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:33:51,990 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.371e+02 2.009e+02 2.458e+02 3.052e+02 4.298e+02, threshold=4.916e+02, percent-clipped=0.0 +2023-03-25 22:34:11,528 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.6326, 4.0578, 4.1920, 4.3907, 4.2663, 4.1460, 4.7167, 1.7538], + device='cuda:4'), covar=tensor([0.0757, 0.0794, 0.0726, 0.1039, 0.1430, 0.1261, 0.0628, 0.4762], + device='cuda:4'), in_proj_covar=tensor([0.0371, 0.0244, 0.0264, 0.0295, 0.0350, 0.0289, 0.0311, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:34:39,799 INFO [finetune.py:976] (4/7) Epoch 1, batch 3150, loss[loss=0.3853, simple_loss=0.3779, pruned_loss=0.1964, over 4805.00 frames. ], tot_loss[loss=0.4091, simple_loss=0.4066, pruned_loss=0.2059, over 955162.09 frames. ], batch size: 29, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:35:16,506 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3180.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:35:39,622 INFO [finetune.py:976] (4/7) Epoch 1, batch 3200, loss[loss=0.3959, simple_loss=0.4024, pruned_loss=0.1947, over 4714.00 frames. ], tot_loss[loss=0.3997, simple_loss=0.3989, pruned_loss=0.2002, over 953487.53 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:35:52,730 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.131e+02 1.973e+02 2.320e+02 2.787e+02 5.091e+02, threshold=4.641e+02, percent-clipped=1.0 +2023-03-25 22:36:29,141 INFO [finetune.py:976] (4/7) Epoch 1, batch 3250, loss[loss=0.5332, simple_loss=0.5066, pruned_loss=0.2799, over 4817.00 frames. ], tot_loss[loss=0.3974, simple_loss=0.3973, pruned_loss=0.1987, over 953004.36 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:37:12,261 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3288.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:37:19,996 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3293.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:37:24,069 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3766, 1.6441, 1.7669, 1.9624, 1.7840, 3.8422, 1.3385, 1.8914], + device='cuda:4'), covar=tensor([0.1148, 0.1580, 0.1369, 0.1086, 0.1512, 0.0185, 0.1539, 0.1762], + device='cuda:4'), in_proj_covar=tensor([0.0072, 0.0075, 0.0070, 0.0073, 0.0087, 0.0074, 0.0081, 0.0075], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-25 22:37:24,557 INFO [finetune.py:976] (4/7) Epoch 1, batch 3300, loss[loss=0.3777, simple_loss=0.3985, pruned_loss=0.1785, over 4858.00 frames. ], tot_loss[loss=0.3992, simple_loss=0.4003, pruned_loss=0.1991, over 953786.60 frames. ], batch size: 44, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:37:32,705 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.553e+02 2.210e+02 2.512e+02 3.057e+02 4.555e+02, threshold=5.024e+02, percent-clipped=0.0 +2023-03-25 22:38:03,265 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3341.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:38:13,347 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3349.0, num_to_drop=2, layers_to_drop={1, 3} +2023-03-25 22:38:14,400 INFO [finetune.py:976] (4/7) Epoch 1, batch 3350, loss[loss=0.3402, simple_loss=0.3501, pruned_loss=0.1651, over 4828.00 frames. ], tot_loss[loss=0.3996, simple_loss=0.4022, pruned_loss=0.1985, over 954087.69 frames. ], batch size: 30, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:38:44,447 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3375.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:38:59,534 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3389.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:39:06,746 INFO [finetune.py:976] (4/7) Epoch 1, batch 3400, loss[loss=0.3573, simple_loss=0.3851, pruned_loss=0.1647, over 4818.00 frames. ], tot_loss[loss=0.3996, simple_loss=0.4034, pruned_loss=0.198, over 954197.04 frames. ], batch size: 41, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:39:20,796 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.307e+02 1.988e+02 2.392e+02 2.720e+02 4.202e+02, threshold=4.784e+02, percent-clipped=0.0 +2023-03-25 22:39:29,602 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2147, 3.5973, 3.7880, 4.1001, 3.9763, 3.7414, 4.2890, 1.3994], + device='cuda:4'), covar=tensor([0.0730, 0.0808, 0.0758, 0.0927, 0.1224, 0.1256, 0.0724, 0.4571], + device='cuda:4'), in_proj_covar=tensor([0.0373, 0.0246, 0.0266, 0.0298, 0.0352, 0.0292, 0.0312, 0.0305], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:39:38,517 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3121, 2.6903, 1.9390, 1.6606, 3.0239, 3.0392, 2.4467, 2.3488], + device='cuda:4'), covar=tensor([0.0946, 0.0663, 0.1060, 0.1203, 0.0395, 0.0767, 0.1013, 0.1159], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0132, 0.0133, 0.0122, 0.0108, 0.0131, 0.0137, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:40:08,846 INFO [finetune.py:976] (4/7) Epoch 1, batch 3450, loss[loss=0.3653, simple_loss=0.3921, pruned_loss=0.1692, over 4811.00 frames. ], tot_loss[loss=0.3971, simple_loss=0.4018, pruned_loss=0.1962, over 953125.13 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:40:40,372 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3480.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:41:03,280 INFO [finetune.py:976] (4/7) Epoch 1, batch 3500, loss[loss=0.391, simple_loss=0.3935, pruned_loss=0.1942, over 4929.00 frames. ], tot_loss[loss=0.39, simple_loss=0.3961, pruned_loss=0.192, over 953511.89 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:41:03,388 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1162, 1.5868, 2.3154, 3.7301, 2.7832, 2.6201, 0.6700, 2.9150], + device='cuda:4'), covar=tensor([0.1734, 0.1619, 0.1381, 0.0456, 0.0843, 0.1454, 0.2072, 0.0720], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0113, 0.0130, 0.0150, 0.0101, 0.0138, 0.0121, 0.0104], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-25 22:41:14,917 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.301e+02 2.271e+02 2.832e+02 3.824e+02 1.123e+03, threshold=5.664e+02, percent-clipped=12.0 +2023-03-25 22:41:30,605 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3528.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:41:48,803 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3546.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:41:57,050 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.58 vs. limit=5.0 +2023-03-25 22:41:57,306 INFO [finetune.py:976] (4/7) Epoch 1, batch 3550, loss[loss=0.415, simple_loss=0.4082, pruned_loss=0.2109, over 4847.00 frames. ], tot_loss[loss=0.3872, simple_loss=0.3927, pruned_loss=0.1908, over 954520.16 frames. ], batch size: 44, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:42:45,285 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3593.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:42:50,743 INFO [finetune.py:976] (4/7) Epoch 1, batch 3600, loss[loss=0.3441, simple_loss=0.3652, pruned_loss=0.1615, over 4768.00 frames. ], tot_loss[loss=0.3811, simple_loss=0.3879, pruned_loss=0.1871, over 957162.07 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:42:59,658 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3607.0, num_to_drop=2, layers_to_drop={2, 3} +2023-03-25 22:43:09,363 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.842e+02 2.557e+02 2.866e+02 3.769e+02 9.044e+02, threshold=5.732e+02, percent-clipped=5.0 +2023-03-25 22:43:33,250 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-03-25 22:43:44,103 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3641.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:43:46,916 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3644.0, num_to_drop=1, layers_to_drop={3} +2023-03-25 22:43:52,549 INFO [finetune.py:976] (4/7) Epoch 1, batch 3650, loss[loss=0.3657, simple_loss=0.3874, pruned_loss=0.172, over 4867.00 frames. ], tot_loss[loss=0.3799, simple_loss=0.3873, pruned_loss=0.1863, over 955984.27 frames. ], batch size: 31, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:43:53,277 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0803, 0.9874, 0.8888, 1.1832, 1.4296, 0.9492, 1.6819, 1.0114], + device='cuda:4'), covar=tensor([2.2289, 4.8701, 3.1381, 4.2606, 2.2874, 1.6755, 2.4114, 3.1999], + device='cuda:4'), in_proj_covar=tensor([0.0158, 0.0186, 0.0224, 0.0239, 0.0200, 0.0171, 0.0177, 0.0181], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0002], + device='cuda:4') +2023-03-25 22:44:20,324 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3675.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:44:21,466 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3676.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:44:40,051 INFO [finetune.py:976] (4/7) Epoch 1, batch 3700, loss[loss=0.3351, simple_loss=0.3485, pruned_loss=0.1608, over 4079.00 frames. ], tot_loss[loss=0.3841, simple_loss=0.3923, pruned_loss=0.1879, over 953871.94 frames. ], batch size: 17, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:44:52,808 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.548e+02 2.567e+02 2.980e+02 3.536e+02 5.905e+02, threshold=5.959e+02, percent-clipped=1.0 +2023-03-25 22:45:09,752 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3723.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:45:23,614 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3737.0, num_to_drop=2, layers_to_drop={2, 3} +2023-03-25 22:45:43,728 INFO [finetune.py:976] (4/7) Epoch 1, batch 3750, loss[loss=0.3965, simple_loss=0.4123, pruned_loss=0.1904, over 4819.00 frames. ], tot_loss[loss=0.3822, simple_loss=0.3925, pruned_loss=0.186, over 954049.91 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:46:33,629 INFO [finetune.py:976] (4/7) Epoch 1, batch 3800, loss[loss=0.3429, simple_loss=0.3709, pruned_loss=0.1574, over 4889.00 frames. ], tot_loss[loss=0.3819, simple_loss=0.393, pruned_loss=0.1854, over 954933.95 frames. ], batch size: 32, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:46:36,135 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=2.00 vs. limit=2.0 +2023-03-25 22:46:42,976 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=5.64 vs. limit=5.0 +2023-03-25 22:46:47,077 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.580e+02 2.141e+02 2.900e+02 3.620e+02 1.043e+03, threshold=5.800e+02, percent-clipped=4.0 +2023-03-25 22:47:22,191 INFO [finetune.py:976] (4/7) Epoch 1, batch 3850, loss[loss=0.4123, simple_loss=0.4029, pruned_loss=0.2109, over 4873.00 frames. ], tot_loss[loss=0.3761, simple_loss=0.3888, pruned_loss=0.1817, over 955954.85 frames. ], batch size: 34, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:48:09,989 INFO [finetune.py:976] (4/7) Epoch 1, batch 3900, loss[loss=0.337, simple_loss=0.3656, pruned_loss=0.1543, over 4933.00 frames. ], tot_loss[loss=0.3704, simple_loss=0.3835, pruned_loss=0.1787, over 954222.00 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:48:10,643 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3902.0, num_to_drop=1, layers_to_drop={3} +2023-03-25 22:48:20,247 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.580e+02 2.264e+02 2.673e+02 3.196e+02 5.181e+02, threshold=5.346e+02, percent-clipped=0.0 +2023-03-25 22:48:30,356 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3926.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:48:50,576 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3944.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:48:54,666 INFO [finetune.py:976] (4/7) Epoch 1, batch 3950, loss[loss=0.3284, simple_loss=0.3564, pruned_loss=0.1502, over 4824.00 frames. ], tot_loss[loss=0.364, simple_loss=0.378, pruned_loss=0.175, over 955468.08 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:49:03,415 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=5.09 vs. limit=5.0 +2023-03-25 22:49:45,870 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3987.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:49:53,766 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=3992.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:50:05,835 INFO [finetune.py:976] (4/7) Epoch 1, batch 4000, loss[loss=0.4031, simple_loss=0.4092, pruned_loss=0.1985, over 4771.00 frames. ], tot_loss[loss=0.3618, simple_loss=0.3757, pruned_loss=0.174, over 952242.68 frames. ], batch size: 54, lr: 4.00e-03, grad_scale: 8.0 +2023-03-25 22:50:14,896 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2588, 1.7097, 1.4846, 1.2014, 1.8726, 2.7720, 2.3378, 1.6660], + device='cuda:4'), covar=tensor([0.0218, 0.0404, 0.0479, 0.0494, 0.0332, 0.0137, 0.0167, 0.0374], + device='cuda:4'), in_proj_covar=tensor([0.0084, 0.0112, 0.0132, 0.0110, 0.0104, 0.0100, 0.0087, 0.0111], + device='cuda:4'), out_proj_covar=tensor([6.5884e-05, 8.8847e-05, 1.0672e-04, 8.7582e-05, 8.2507e-05, 7.4793e-05, + 6.7383e-05, 8.6981e-05], device='cuda:4') +2023-03-25 22:50:18,333 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.331e+02 2.072e+02 2.562e+02 2.941e+02 5.028e+02, threshold=5.123e+02, percent-clipped=0.0 +2023-03-25 22:50:31,342 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4032.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:50:50,288 INFO [finetune.py:976] (4/7) Epoch 1, batch 4050, loss[loss=0.4677, simple_loss=0.4588, pruned_loss=0.2383, over 4159.00 frames. ], tot_loss[loss=0.3633, simple_loss=0.3781, pruned_loss=0.1742, over 950798.99 frames. ], batch size: 65, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:51:10,555 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1662, 1.4679, 1.1097, 1.5080, 1.4797, 2.8101, 1.2399, 1.5179], + device='cuda:4'), covar=tensor([0.1236, 0.1757, 0.1536, 0.1145, 0.1615, 0.0340, 0.1557, 0.1822], + device='cuda:4'), in_proj_covar=tensor([0.0073, 0.0076, 0.0072, 0.0075, 0.0088, 0.0076, 0.0082, 0.0075], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-25 22:51:30,270 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.83 vs. limit=5.0 +2023-03-25 22:51:46,309 INFO [finetune.py:976] (4/7) Epoch 1, batch 4100, loss[loss=0.3749, simple_loss=0.4014, pruned_loss=0.1742, over 4895.00 frames. ], tot_loss[loss=0.3642, simple_loss=0.3802, pruned_loss=0.1742, over 950767.56 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:52:00,028 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.481e+02 2.010e+02 2.495e+02 2.957e+02 5.246e+02, threshold=4.990e+02, percent-clipped=1.0 +2023-03-25 22:52:34,824 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=4141.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:52:43,038 INFO [finetune.py:976] (4/7) Epoch 1, batch 4150, loss[loss=0.3761, simple_loss=0.3918, pruned_loss=0.1802, over 4728.00 frames. ], tot_loss[loss=0.3626, simple_loss=0.38, pruned_loss=0.1726, over 950363.67 frames. ], batch size: 59, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:52:43,791 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1155, 2.3102, 1.9956, 1.5159, 2.7392, 2.5616, 2.2595, 2.1454], + device='cuda:4'), covar=tensor([0.0800, 0.0588, 0.0809, 0.1070, 0.0324, 0.0702, 0.0809, 0.1040], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0130, 0.0132, 0.0121, 0.0106, 0.0131, 0.0137, 0.0158], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:53:24,477 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-25 22:53:31,299 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9948, 2.3378, 2.4455, 1.1930, 2.3790, 2.2596, 1.8042, 2.2142], + device='cuda:4'), covar=tensor([0.1108, 0.1298, 0.1876, 0.3281, 0.1817, 0.2308, 0.2676, 0.1609], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0169, 0.0183, 0.0169, 0.0188, 0.0187, 0.0193, 0.0181], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:53:43,097 INFO [finetune.py:976] (4/7) Epoch 1, batch 4200, loss[loss=0.3345, simple_loss=0.3719, pruned_loss=0.1485, over 4811.00 frames. ], tot_loss[loss=0.3602, simple_loss=0.379, pruned_loss=0.1707, over 949402.62 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:53:43,826 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4202.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 22:53:43,865 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=4202.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:54:02,772 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.292e+02 2.089e+02 2.432e+02 2.936e+02 5.530e+02, threshold=4.864e+02, percent-clipped=1.0 +2023-03-25 22:54:44,473 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4250.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:54:45,015 INFO [finetune.py:976] (4/7) Epoch 1, batch 4250, loss[loss=0.3092, simple_loss=0.3273, pruned_loss=0.1455, over 4802.00 frames. ], tot_loss[loss=0.3535, simple_loss=0.3733, pruned_loss=0.1668, over 951940.76 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:55:23,158 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4282.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:55:27,291 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=4288.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:55:35,147 INFO [finetune.py:976] (4/7) Epoch 1, batch 4300, loss[loss=0.3078, simple_loss=0.328, pruned_loss=0.1438, over 4829.00 frames. ], tot_loss[loss=0.348, simple_loss=0.3687, pruned_loss=0.1636, over 953201.80 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:55:45,446 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.425e+02 1.950e+02 2.267e+02 2.860e+02 4.056e+02, threshold=4.534e+02, percent-clipped=0.0 +2023-03-25 22:56:13,411 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4332.0, num_to_drop=1, layers_to_drop={2} +2023-03-25 22:56:35,737 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=4349.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:56:36,848 INFO [finetune.py:976] (4/7) Epoch 1, batch 4350, loss[loss=0.2933, simple_loss=0.3221, pruned_loss=0.1322, over 4773.00 frames. ], tot_loss[loss=0.3417, simple_loss=0.3633, pruned_loss=0.16, over 954206.63 frames. ], batch size: 26, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:57:17,558 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4380.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 22:57:36,805 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1378, 1.5801, 2.5159, 1.6030, 2.1082, 2.1325, 1.5849, 2.3163], + device='cuda:4'), covar=tensor([0.2048, 0.2470, 0.1700, 0.2504, 0.1177, 0.2025, 0.2713, 0.1266], + device='cuda:4'), in_proj_covar=tensor([0.0194, 0.0199, 0.0196, 0.0185, 0.0168, 0.0211, 0.0205, 0.0186], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:57:44,939 INFO [finetune.py:976] (4/7) Epoch 1, batch 4400, loss[loss=0.3403, simple_loss=0.3761, pruned_loss=0.1523, over 4843.00 frames. ], tot_loss[loss=0.3426, simple_loss=0.3647, pruned_loss=0.1603, over 954133.26 frames. ], batch size: 44, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:57:57,029 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.139e+02 1.978e+02 2.430e+02 2.895e+02 4.966e+02, threshold=4.860e+02, percent-clipped=1.0 +2023-03-25 22:58:14,218 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6021, 2.0870, 1.9856, 1.4185, 2.2610, 3.1770, 2.7011, 2.0684], + device='cuda:4'), covar=tensor([0.0196, 0.0433, 0.0440, 0.0521, 0.0275, 0.0133, 0.0234, 0.0411], + device='cuda:4'), in_proj_covar=tensor([0.0084, 0.0113, 0.0132, 0.0111, 0.0104, 0.0100, 0.0087, 0.0111], + device='cuda:4'), out_proj_covar=tensor([6.5800e-05, 8.9432e-05, 1.0675e-04, 8.8102e-05, 8.2723e-05, 7.4576e-05, + 6.7425e-05, 8.6967e-05], device='cuda:4') +2023-03-25 22:58:24,338 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-25 22:58:28,111 INFO [finetune.py:976] (4/7) Epoch 1, batch 4450, loss[loss=0.3499, simple_loss=0.3977, pruned_loss=0.1511, over 4837.00 frames. ], tot_loss[loss=0.3481, simple_loss=0.3696, pruned_loss=0.1633, over 953995.26 frames. ], batch size: 49, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:58:49,012 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8664, 1.5591, 2.0819, 1.4758, 1.8042, 1.9474, 1.5168, 2.1428], + device='cuda:4'), covar=tensor([0.1981, 0.2144, 0.1570, 0.2142, 0.1154, 0.1743, 0.2635, 0.0980], + device='cuda:4'), in_proj_covar=tensor([0.0194, 0.0199, 0.0196, 0.0185, 0.0167, 0.0211, 0.0206, 0.0186], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 22:59:09,844 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4497.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 22:59:12,225 INFO [finetune.py:976] (4/7) Epoch 1, batch 4500, loss[loss=0.3812, simple_loss=0.406, pruned_loss=0.1782, over 4792.00 frames. ], tot_loss[loss=0.3466, simple_loss=0.3692, pruned_loss=0.162, over 953631.07 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 22:59:29,321 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.227e+02 2.111e+02 2.516e+02 2.889e+02 5.762e+02, threshold=5.032e+02, percent-clipped=1.0 +2023-03-25 23:00:15,077 INFO [finetune.py:976] (4/7) Epoch 1, batch 4550, loss[loss=0.3582, simple_loss=0.3927, pruned_loss=0.1618, over 4846.00 frames. ], tot_loss[loss=0.3483, simple_loss=0.3713, pruned_loss=0.1627, over 953672.29 frames. ], batch size: 44, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:00:55,801 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4582.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:01:25,593 INFO [finetune.py:976] (4/7) Epoch 1, batch 4600, loss[loss=0.3167, simple_loss=0.3501, pruned_loss=0.1417, over 4854.00 frames. ], tot_loss[loss=0.3445, simple_loss=0.3688, pruned_loss=0.1601, over 955852.98 frames. ], batch size: 31, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:01:38,111 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.209e+02 2.082e+02 2.456e+02 3.064e+02 5.977e+02, threshold=4.911e+02, percent-clipped=1.0 +2023-03-25 23:01:59,353 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4630.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:02:18,020 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4644.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:02:28,981 INFO [finetune.py:976] (4/7) Epoch 1, batch 4650, loss[loss=0.304, simple_loss=0.3257, pruned_loss=0.1411, over 4887.00 frames. ], tot_loss[loss=0.3394, simple_loss=0.3636, pruned_loss=0.1575, over 953864.36 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:03:06,337 INFO [finetune.py:976] (4/7) Epoch 1, batch 4700, loss[loss=0.3055, simple_loss=0.3341, pruned_loss=0.1384, over 4826.00 frames. ], tot_loss[loss=0.3333, simple_loss=0.3578, pruned_loss=0.1543, over 953542.16 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:03:20,654 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.265e+02 1.870e+02 2.224e+02 2.796e+02 5.273e+02, threshold=4.448e+02, percent-clipped=2.0 +2023-03-25 23:03:59,288 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.94 vs. limit=5.0 +2023-03-25 23:03:59,441 INFO [finetune.py:976] (4/7) Epoch 1, batch 4750, loss[loss=0.3147, simple_loss=0.3503, pruned_loss=0.1395, over 4864.00 frames. ], tot_loss[loss=0.3293, simple_loss=0.3543, pruned_loss=0.1521, over 955573.60 frames. ], batch size: 34, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:04:36,392 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4797.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:04:39,223 INFO [finetune.py:976] (4/7) Epoch 1, batch 4800, loss[loss=0.347, simple_loss=0.371, pruned_loss=0.1616, over 4768.00 frames. ], tot_loss[loss=0.3336, simple_loss=0.3585, pruned_loss=0.1543, over 956592.12 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:04:56,835 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.406e+02 2.096e+02 2.556e+02 3.186e+02 5.883e+02, threshold=5.111e+02, percent-clipped=4.0 +2023-03-25 23:05:32,161 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4845.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:05:41,984 INFO [finetune.py:976] (4/7) Epoch 1, batch 4850, loss[loss=0.3906, simple_loss=0.4004, pruned_loss=0.1904, over 4303.00 frames. ], tot_loss[loss=0.3352, simple_loss=0.3617, pruned_loss=0.1544, over 953968.67 frames. ], batch size: 66, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:06:28,850 INFO [finetune.py:976] (4/7) Epoch 1, batch 4900, loss[loss=0.2937, simple_loss=0.3316, pruned_loss=0.1279, over 4780.00 frames. ], tot_loss[loss=0.3349, simple_loss=0.3616, pruned_loss=0.1541, over 953146.22 frames. ], batch size: 29, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:06:45,503 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.521e+02 2.056e+02 2.408e+02 2.893e+02 5.886e+02, threshold=4.817e+02, percent-clipped=2.0 +2023-03-25 23:06:47,355 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6847, 1.2565, 1.1122, 0.2372, 1.2575, 1.5031, 1.3092, 1.4211], + device='cuda:4'), covar=tensor([0.0726, 0.0940, 0.1326, 0.2118, 0.1237, 0.2076, 0.2206, 0.0915], + device='cuda:4'), in_proj_covar=tensor([0.0158, 0.0175, 0.0189, 0.0174, 0.0196, 0.0195, 0.0199, 0.0186], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:07:07,585 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4551, 1.5154, 1.1107, 1.5793, 1.6288, 1.1859, 1.9816, 1.3885], + device='cuda:4'), covar=tensor([0.3560, 0.6894, 0.6664, 0.7604, 0.4255, 0.3336, 0.5061, 0.5362], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0180, 0.0220, 0.0233, 0.0195, 0.0167, 0.0179, 0.0175], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:4') +2023-03-25 23:07:15,535 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4944.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:07:19,810 INFO [finetune.py:976] (4/7) Epoch 1, batch 4950, loss[loss=0.29, simple_loss=0.3365, pruned_loss=0.1217, over 4865.00 frames. ], tot_loss[loss=0.334, simple_loss=0.3613, pruned_loss=0.1534, over 954241.59 frames. ], batch size: 31, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:08:09,425 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5723, 1.4475, 1.1672, 1.1187, 1.6441, 1.9061, 1.5318, 1.1631], + device='cuda:4'), covar=tensor([0.0259, 0.0443, 0.0564, 0.0542, 0.0307, 0.0279, 0.0370, 0.0525], + device='cuda:4'), in_proj_covar=tensor([0.0083, 0.0111, 0.0131, 0.0110, 0.0103, 0.0098, 0.0087, 0.0109], + device='cuda:4'), out_proj_covar=tensor([6.5075e-05, 8.7937e-05, 1.0595e-04, 8.7490e-05, 8.2058e-05, 7.3751e-05, + 6.6998e-05, 8.5565e-05], device='cuda:4') +2023-03-25 23:08:11,782 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=4992.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:08:22,748 INFO [finetune.py:976] (4/7) Epoch 1, batch 5000, loss[loss=0.3761, simple_loss=0.3942, pruned_loss=0.179, over 4913.00 frames. ], tot_loss[loss=0.3307, simple_loss=0.3588, pruned_loss=0.1513, over 956446.20 frames. ], batch size: 37, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:08:32,725 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 2.131e+02 2.461e+02 3.038e+02 5.796e+02, threshold=4.923e+02, percent-clipped=4.0 +2023-03-25 23:08:59,418 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2780, 0.8058, 1.1295, 0.9196, 0.9657, 0.9385, 0.9369, 1.1099], + device='cuda:4'), covar=tensor([2.8411, 5.7528, 3.7985, 4.6111, 4.9753, 3.3129, 6.1394, 3.4002], + device='cuda:4'), in_proj_covar=tensor([0.0207, 0.0236, 0.0221, 0.0250, 0.0231, 0.0203, 0.0258, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-25 23:09:20,252 INFO [finetune.py:976] (4/7) Epoch 1, batch 5050, loss[loss=0.2717, simple_loss=0.3183, pruned_loss=0.1125, over 4763.00 frames. ], tot_loss[loss=0.3269, simple_loss=0.3549, pruned_loss=0.1495, over 955175.88 frames. ], batch size: 27, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:09:21,058 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-25 23:09:43,836 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8199, 1.7182, 1.4339, 1.8782, 2.0395, 1.5633, 2.2581, 1.6373], + device='cuda:4'), covar=tensor([0.3703, 0.6434, 0.7504, 0.7184, 0.4672, 0.3716, 0.4949, 0.5675], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0181, 0.0221, 0.0235, 0.0195, 0.0168, 0.0180, 0.0176], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:4') +2023-03-25 23:10:01,167 INFO [finetune.py:976] (4/7) Epoch 1, batch 5100, loss[loss=0.3203, simple_loss=0.3461, pruned_loss=0.1472, over 4817.00 frames. ], tot_loss[loss=0.322, simple_loss=0.3502, pruned_loss=0.1469, over 954597.78 frames. ], batch size: 38, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:10:09,455 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.918e+02 2.379e+02 2.966e+02 8.444e+02, threshold=4.758e+02, percent-clipped=2.0 +2023-03-25 23:10:34,835 INFO [finetune.py:976] (4/7) Epoch 1, batch 5150, loss[loss=0.296, simple_loss=0.3386, pruned_loss=0.1267, over 4911.00 frames. ], tot_loss[loss=0.3197, simple_loss=0.3477, pruned_loss=0.1458, over 953173.44 frames. ], batch size: 36, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:11:14,837 INFO [finetune.py:976] (4/7) Epoch 1, batch 5200, loss[loss=0.4033, simple_loss=0.4275, pruned_loss=0.1895, over 4902.00 frames. ], tot_loss[loss=0.3231, simple_loss=0.3517, pruned_loss=0.1473, over 952563.57 frames. ], batch size: 43, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:11:24,773 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.601e+02 2.261e+02 2.570e+02 3.078e+02 5.221e+02, threshold=5.140e+02, percent-clipped=2.0 +2023-03-25 23:11:57,507 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4424, 1.5491, 1.7164, 1.8606, 1.8295, 4.1020, 1.2428, 1.8687], + device='cuda:4'), covar=tensor([0.1409, 0.2268, 0.1662, 0.1415, 0.1943, 0.0223, 0.2237, 0.2385], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0077, 0.0074, 0.0076, 0.0090, 0.0078, 0.0083, 0.0076], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-25 23:12:06,969 INFO [finetune.py:976] (4/7) Epoch 1, batch 5250, loss[loss=0.3853, simple_loss=0.4062, pruned_loss=0.1822, over 4746.00 frames. ], tot_loss[loss=0.3236, simple_loss=0.3528, pruned_loss=0.1472, over 952648.05 frames. ], batch size: 59, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:12:55,137 INFO [finetune.py:976] (4/7) Epoch 1, batch 5300, loss[loss=0.3544, simple_loss=0.3747, pruned_loss=0.1671, over 4811.00 frames. ], tot_loss[loss=0.3245, simple_loss=0.3538, pruned_loss=0.1477, over 951963.38 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:13:08,331 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.321e+02 2.032e+02 2.465e+02 2.907e+02 4.480e+02, threshold=4.930e+02, percent-clipped=0.0 +2023-03-25 23:13:49,138 INFO [finetune.py:976] (4/7) Epoch 1, batch 5350, loss[loss=0.2779, simple_loss=0.328, pruned_loss=0.1139, over 4849.00 frames. ], tot_loss[loss=0.3254, simple_loss=0.3548, pruned_loss=0.148, over 952887.45 frames. ], batch size: 49, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:14:48,109 INFO [finetune.py:976] (4/7) Epoch 1, batch 5400, loss[loss=0.2948, simple_loss=0.3271, pruned_loss=0.1312, over 4815.00 frames. ], tot_loss[loss=0.322, simple_loss=0.3513, pruned_loss=0.1463, over 953996.42 frames. ], batch size: 40, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:14:49,463 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5668, 1.4844, 1.1608, 1.5954, 1.8766, 1.3215, 2.0908, 1.4570], + device='cuda:4'), covar=tensor([0.3715, 0.7744, 0.7641, 0.8199, 0.4397, 0.3626, 0.5989, 0.5428], + device='cuda:4'), in_proj_covar=tensor([0.0155, 0.0183, 0.0224, 0.0237, 0.0198, 0.0170, 0.0183, 0.0178], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-25 23:14:51,824 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5107, 1.2492, 1.3499, 1.4951, 2.0263, 1.4377, 1.0886, 1.2030], + device='cuda:4'), covar=tensor([0.3091, 0.3059, 0.2437, 0.2273, 0.2625, 0.1775, 0.3849, 0.2251], + device='cuda:4'), in_proj_covar=tensor([0.0211, 0.0197, 0.0182, 0.0168, 0.0217, 0.0166, 0.0196, 0.0171], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:14:55,692 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.63 vs. limit=5.0 +2023-03-25 23:14:55,970 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.226e+02 1.939e+02 2.339e+02 2.729e+02 4.650e+02, threshold=4.678e+02, percent-clipped=0.0 +2023-03-25 23:15:36,521 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0835, 1.8361, 1.5325, 0.5901, 1.5720, 1.7537, 1.4802, 1.8584], + device='cuda:4'), covar=tensor([0.0863, 0.0859, 0.1466, 0.2134, 0.1231, 0.2206, 0.2241, 0.0859], + device='cuda:4'), in_proj_covar=tensor([0.0158, 0.0176, 0.0189, 0.0173, 0.0196, 0.0195, 0.0199, 0.0187], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:15:39,347 INFO [finetune.py:976] (4/7) Epoch 1, batch 5450, loss[loss=0.2553, simple_loss=0.2995, pruned_loss=0.1056, over 4704.00 frames. ], tot_loss[loss=0.3175, simple_loss=0.3472, pruned_loss=0.1439, over 954483.21 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:16:08,562 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.58 vs. limit=2.0 +2023-03-25 23:16:31,086 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5914, 1.6943, 1.5755, 1.0916, 2.0231, 1.7397, 1.6840, 1.5711], + device='cuda:4'), covar=tensor([0.0821, 0.0748, 0.0903, 0.1087, 0.0451, 0.0893, 0.0894, 0.1283], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0132, 0.0137, 0.0127, 0.0108, 0.0136, 0.0142, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:16:31,588 INFO [finetune.py:976] (4/7) Epoch 1, batch 5500, loss[loss=0.3119, simple_loss=0.3373, pruned_loss=0.1432, over 4822.00 frames. ], tot_loss[loss=0.3135, simple_loss=0.3435, pruned_loss=0.1417, over 954565.94 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:16:42,830 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7053, 1.0107, 1.2938, 1.2445, 1.1812, 1.2747, 1.2571, 1.3627], + device='cuda:4'), covar=tensor([3.5238, 8.1659, 4.8718, 6.2373, 6.5604, 3.9786, 8.4658, 4.8217], + device='cuda:4'), in_proj_covar=tensor([0.0210, 0.0240, 0.0225, 0.0254, 0.0234, 0.0206, 0.0262, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-25 23:16:45,506 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3731, 1.2085, 0.9531, 0.9510, 1.1417, 1.0720, 1.1161, 1.9337], + device='cuda:4'), covar=tensor([3.5522, 3.4358, 2.8751, 4.6883, 2.7684, 2.1002, 3.5751, 1.0189], + device='cuda:4'), in_proj_covar=tensor([0.0224, 0.0213, 0.0195, 0.0248, 0.0207, 0.0177, 0.0211, 0.0158], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-25 23:16:45,961 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.302e+02 2.026e+02 2.277e+02 2.875e+02 1.009e+03, threshold=4.553e+02, percent-clipped=5.0 +2023-03-25 23:17:20,561 INFO [finetune.py:976] (4/7) Epoch 1, batch 5550, loss[loss=0.3085, simple_loss=0.3522, pruned_loss=0.1324, over 4877.00 frames. ], tot_loss[loss=0.3186, simple_loss=0.3479, pruned_loss=0.1446, over 954776.09 frames. ], batch size: 34, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:17:46,981 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8208, 2.2309, 2.3370, 1.1797, 2.3349, 2.0112, 1.6352, 2.1035], + device='cuda:4'), covar=tensor([0.0868, 0.1214, 0.1988, 0.3087, 0.1741, 0.2556, 0.2559, 0.1638], + device='cuda:4'), in_proj_covar=tensor([0.0158, 0.0176, 0.0189, 0.0173, 0.0197, 0.0195, 0.0199, 0.0187], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:18:01,091 INFO [finetune.py:976] (4/7) Epoch 1, batch 5600, loss[loss=0.4054, simple_loss=0.3949, pruned_loss=0.208, over 4191.00 frames. ], tot_loss[loss=0.3205, simple_loss=0.3508, pruned_loss=0.1451, over 953702.64 frames. ], batch size: 65, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:18:19,454 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.468e+02 1.839e+02 2.287e+02 2.793e+02 4.099e+02, threshold=4.573e+02, percent-clipped=0.0 +2023-03-25 23:18:59,388 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5650.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 23:18:59,874 INFO [finetune.py:976] (4/7) Epoch 1, batch 5650, loss[loss=0.3205, simple_loss=0.357, pruned_loss=0.142, over 4869.00 frames. ], tot_loss[loss=0.3226, simple_loss=0.3536, pruned_loss=0.1458, over 953727.87 frames. ], batch size: 31, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:19:26,175 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0858, 2.0088, 1.7488, 2.1721, 2.3095, 1.7905, 2.5411, 1.9952], + device='cuda:4'), covar=tensor([0.2739, 0.5409, 0.5678, 0.5147, 0.3506, 0.2782, 0.4362, 0.4136], + device='cuda:4'), in_proj_covar=tensor([0.0156, 0.0185, 0.0225, 0.0238, 0.0200, 0.0171, 0.0186, 0.0179], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:19:35,521 INFO [finetune.py:976] (4/7) Epoch 1, batch 5700, loss[loss=0.2144, simple_loss=0.2541, pruned_loss=0.0873, over 4598.00 frames. ], tot_loss[loss=0.3179, simple_loss=0.3477, pruned_loss=0.1441, over 933538.94 frames. ], batch size: 20, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:19:35,651 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7912, 1.6064, 1.3741, 1.4709, 1.5974, 1.4399, 1.4493, 2.3822], + device='cuda:4'), covar=tensor([4.1848, 3.5361, 3.0763, 5.0015, 2.9280, 2.3059, 4.1345, 1.0321], + device='cuda:4'), in_proj_covar=tensor([0.0225, 0.0214, 0.0196, 0.0249, 0.0209, 0.0178, 0.0213, 0.0159], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-25 23:19:41,538 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5711.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 23:19:43,180 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.818e+02 2.245e+02 2.685e+02 4.321e+02, threshold=4.489e+02, percent-clipped=0.0 +2023-03-25 23:19:46,353 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.67 vs. limit=2.0 +2023-03-25 23:20:08,413 INFO [finetune.py:976] (4/7) Epoch 2, batch 0, loss[loss=0.2991, simple_loss=0.3445, pruned_loss=0.1269, over 4760.00 frames. ], tot_loss[loss=0.2991, simple_loss=0.3445, pruned_loss=0.1269, over 4760.00 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:20:08,413 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-25 23:20:25,000 INFO [finetune.py:1010] (4/7) Epoch 2, validation: loss=0.2224, simple_loss=0.2847, pruned_loss=0.08, over 2265189.00 frames. +2023-03-25 23:20:25,000 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6026MB +2023-03-25 23:20:47,072 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-25 23:20:56,859 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5755.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:20:59,309 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0585, 2.1802, 1.8979, 1.3859, 2.5169, 2.2190, 2.0400, 1.9308], + device='cuda:4'), covar=tensor([0.0803, 0.0592, 0.0974, 0.1209, 0.0356, 0.0868, 0.0892, 0.1111], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0132, 0.0138, 0.0126, 0.0107, 0.0136, 0.0142, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:21:22,830 INFO [finetune.py:976] (4/7) Epoch 2, batch 50, loss[loss=0.3272, simple_loss=0.3497, pruned_loss=0.1524, over 4902.00 frames. ], tot_loss[loss=0.3171, simple_loss=0.3504, pruned_loss=0.1419, over 213945.89 frames. ], batch size: 46, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:21:27,864 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6530, 1.5215, 1.8958, 1.3806, 1.6962, 1.8154, 1.5306, 1.9354], + device='cuda:4'), covar=tensor([0.1176, 0.1521, 0.1014, 0.1311, 0.0702, 0.1078, 0.1971, 0.0725], + device='cuda:4'), in_proj_covar=tensor([0.0199, 0.0200, 0.0199, 0.0190, 0.0172, 0.0217, 0.0208, 0.0192], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:21:54,356 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.238e+02 1.870e+02 2.317e+02 2.912e+02 7.564e+02, threshold=4.633e+02, percent-clipped=3.0 +2023-03-25 23:21:55,698 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5816.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:22:11,076 INFO [finetune.py:976] (4/7) Epoch 2, batch 100, loss[loss=0.2955, simple_loss=0.3298, pruned_loss=0.1306, over 4755.00 frames. ], tot_loss[loss=0.3071, simple_loss=0.3403, pruned_loss=0.137, over 378173.61 frames. ], batch size: 27, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:22:36,213 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5858.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:22:49,648 INFO [finetune.py:976] (4/7) Epoch 2, batch 150, loss[loss=0.2669, simple_loss=0.321, pruned_loss=0.1064, over 4831.00 frames. ], tot_loss[loss=0.3008, simple_loss=0.3337, pruned_loss=0.134, over 506002.54 frames. ], batch size: 30, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:23:18,195 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.293e+02 1.883e+02 2.329e+02 2.858e+02 5.160e+02, threshold=4.657e+02, percent-clipped=2.0 +2023-03-25 23:23:21,831 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5919.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:23:28,109 INFO [finetune.py:976] (4/7) Epoch 2, batch 200, loss[loss=0.3296, simple_loss=0.3616, pruned_loss=0.1487, over 4739.00 frames. ], tot_loss[loss=0.2988, simple_loss=0.3319, pruned_loss=0.1328, over 606844.66 frames. ], batch size: 59, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:23:44,975 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5363, 1.5935, 1.4543, 1.6205, 0.9044, 3.3801, 1.2635, 1.8577], + device='cuda:4'), covar=tensor([0.3716, 0.2350, 0.2131, 0.2254, 0.2256, 0.0181, 0.3046, 0.1598], + device='cuda:4'), in_proj_covar=tensor([0.0124, 0.0106, 0.0112, 0.0113, 0.0107, 0.0091, 0.0094, 0.0091], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0003, 0.0004, 0.0004], + device='cuda:4') +2023-03-25 23:24:01,220 INFO [finetune.py:976] (4/7) Epoch 2, batch 250, loss[loss=0.3048, simple_loss=0.3322, pruned_loss=0.1387, over 4906.00 frames. ], tot_loss[loss=0.3024, simple_loss=0.336, pruned_loss=0.1344, over 682995.34 frames. ], batch size: 32, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:24:20,393 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5990.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:24:42,067 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6006.0, num_to_drop=1, layers_to_drop={3} +2023-03-25 23:24:48,586 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.968e+02 2.365e+02 2.842e+02 7.361e+02, threshold=4.731e+02, percent-clipped=2.0 +2023-03-25 23:24:57,926 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4819, 1.2279, 1.2594, 1.1628, 1.5844, 1.5941, 1.4075, 1.1992], + device='cuda:4'), covar=tensor([0.0277, 0.0392, 0.0534, 0.0406, 0.0276, 0.0312, 0.0306, 0.0433], + device='cuda:4'), in_proj_covar=tensor([0.0083, 0.0111, 0.0130, 0.0110, 0.0103, 0.0098, 0.0087, 0.0108], + device='cuda:4'), out_proj_covar=tensor([6.4782e-05, 8.7662e-05, 1.0536e-04, 8.7398e-05, 8.1900e-05, 7.3264e-05, + 6.7211e-05, 8.4981e-05], device='cuda:4') +2023-03-25 23:25:00,267 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.2948, 1.4702, 1.6033, 0.9033, 1.2866, 1.6427, 1.6623, 1.4828], + device='cuda:4'), covar=tensor([0.1114, 0.0574, 0.0440, 0.0723, 0.0538, 0.0576, 0.0400, 0.0655], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0146, 0.0113, 0.0124, 0.0124, 0.0112, 0.0139, 0.0140], + device='cuda:4'), out_proj_covar=tensor([9.2145e-05, 1.0929e-04, 8.2775e-05, 9.0850e-05, 8.9937e-05, 8.2886e-05, + 1.0390e-04, 1.0425e-04], device='cuda:4') +2023-03-25 23:25:01,947 INFO [finetune.py:976] (4/7) Epoch 2, batch 300, loss[loss=0.3763, simple_loss=0.3884, pruned_loss=0.1821, over 4911.00 frames. ], tot_loss[loss=0.3086, simple_loss=0.3423, pruned_loss=0.1374, over 744559.19 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:25:29,066 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6043.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:25:38,834 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6051.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:26:11,334 INFO [finetune.py:976] (4/7) Epoch 2, batch 350, loss[loss=0.2575, simple_loss=0.2836, pruned_loss=0.1157, over 4324.00 frames. ], tot_loss[loss=0.3107, simple_loss=0.3446, pruned_loss=0.1384, over 789650.31 frames. ], batch size: 19, lr: 4.00e-03, grad_scale: 32.0 +2023-03-25 23:26:35,002 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6104.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:26:39,149 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6111.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:26:41,480 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.275e+02 2.097e+02 2.536e+02 2.955e+02 5.135e+02, threshold=5.071e+02, percent-clipped=1.0 +2023-03-25 23:26:59,638 INFO [finetune.py:976] (4/7) Epoch 2, batch 400, loss[loss=0.3261, simple_loss=0.3624, pruned_loss=0.1449, over 4812.00 frames. ], tot_loss[loss=0.3118, simple_loss=0.3459, pruned_loss=0.1388, over 824718.71 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:27:09,953 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6135.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:27:31,138 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-03-25 23:27:49,727 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6170.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:27:59,427 INFO [finetune.py:976] (4/7) Epoch 2, batch 450, loss[loss=0.2768, simple_loss=0.3254, pruned_loss=0.1141, over 4886.00 frames. ], tot_loss[loss=0.3083, simple_loss=0.3428, pruned_loss=0.1369, over 854131.94 frames. ], batch size: 35, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:28:07,490 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5323, 1.4571, 1.1205, 1.5091, 1.7072, 1.3002, 1.9870, 1.4875], + device='cuda:4'), covar=tensor([0.3492, 0.7227, 0.7207, 0.7103, 0.4383, 0.3305, 0.5418, 0.5090], + device='cuda:4'), in_proj_covar=tensor([0.0157, 0.0186, 0.0228, 0.0240, 0.0201, 0.0173, 0.0188, 0.0180], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:28:14,473 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6196.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:28:17,367 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.66 vs. limit=2.0 +2023-03-25 23:28:25,704 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6203.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:28:26,774 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-25 23:28:33,022 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6211.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:28:34,828 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6214.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:28:35,341 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.140e+02 1.932e+02 2.244e+02 2.718e+02 3.817e+02, threshold=4.487e+02, percent-clipped=0.0 +2023-03-25 23:28:45,040 INFO [finetune.py:976] (4/7) Epoch 2, batch 500, loss[loss=0.2756, simple_loss=0.3154, pruned_loss=0.1179, over 4845.00 frames. ], tot_loss[loss=0.3044, simple_loss=0.3395, pruned_loss=0.1347, over 877847.95 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:28:52,708 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6231.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:29:19,934 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1089, 2.2574, 1.8754, 1.4486, 2.4795, 2.2882, 2.1356, 1.9564], + device='cuda:4'), covar=tensor([0.0724, 0.0597, 0.0888, 0.1097, 0.0363, 0.0788, 0.0843, 0.1054], + device='cuda:4'), in_proj_covar=tensor([0.0136, 0.0134, 0.0139, 0.0128, 0.0109, 0.0138, 0.0144, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:29:25,577 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6264.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:29:26,793 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6266.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:29:35,120 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6272.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:29:38,741 INFO [finetune.py:976] (4/7) Epoch 2, batch 550, loss[loss=0.2766, simple_loss=0.3181, pruned_loss=0.1176, over 4757.00 frames. ], tot_loss[loss=0.3005, simple_loss=0.3356, pruned_loss=0.1327, over 895141.31 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:29:48,756 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5817, 1.3536, 1.1212, 1.2372, 1.3117, 1.2776, 1.2759, 2.2229], + device='cuda:4'), covar=tensor([2.8988, 2.7941, 2.3560, 3.8383, 2.2194, 1.6408, 2.8087, 0.7440], + device='cuda:4'), in_proj_covar=tensor([0.0230, 0.0217, 0.0198, 0.0252, 0.0210, 0.0179, 0.0215, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-25 23:30:17,724 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6306.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 23:30:18,834 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.4480, 1.5048, 1.6070, 0.8515, 1.3746, 1.7529, 1.6587, 1.4428], + device='cuda:4'), covar=tensor([0.0997, 0.0610, 0.0347, 0.0696, 0.0437, 0.0472, 0.0344, 0.0551], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0146, 0.0112, 0.0124, 0.0123, 0.0112, 0.0138, 0.0139], + device='cuda:4'), out_proj_covar=tensor([9.2348e-05, 1.0918e-04, 8.2350e-05, 9.0821e-05, 8.9646e-05, 8.2816e-05, + 1.0366e-04, 1.0342e-04], device='cuda:4') +2023-03-25 23:30:29,438 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.346e+02 1.947e+02 2.353e+02 2.718e+02 5.175e+02, threshold=4.705e+02, percent-clipped=1.0 +2023-03-25 23:30:30,202 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4075, 1.1015, 1.1266, 1.0923, 0.9743, 1.0382, 1.0886, 1.2031], + device='cuda:4'), covar=tensor([2.9702, 5.0178, 3.4971, 4.6733, 4.9860, 3.2478, 6.5623, 3.3865], + device='cuda:4'), in_proj_covar=tensor([0.0215, 0.0248, 0.0232, 0.0260, 0.0239, 0.0211, 0.0269, 0.0206], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-25 23:30:47,513 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6327.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:30:48,004 INFO [finetune.py:976] (4/7) Epoch 2, batch 600, loss[loss=0.2963, simple_loss=0.3605, pruned_loss=0.1161, over 4921.00 frames. ], tot_loss[loss=0.3033, simple_loss=0.3377, pruned_loss=0.1344, over 904569.62 frames. ], batch size: 42, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:31:07,302 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6346.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:31:13,142 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6354.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 23:31:28,048 INFO [finetune.py:976] (4/7) Epoch 2, batch 650, loss[loss=0.3688, simple_loss=0.4083, pruned_loss=0.1646, over 4801.00 frames. ], tot_loss[loss=0.3056, simple_loss=0.3403, pruned_loss=0.1354, over 913710.23 frames. ], batch size: 45, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:31:42,402 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6399.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:31:51,330 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6411.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:31:53,613 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.435e+02 2.011e+02 2.373e+02 2.999e+02 4.783e+02, threshold=4.746e+02, percent-clipped=1.0 +2023-03-25 23:32:01,502 INFO [finetune.py:976] (4/7) Epoch 2, batch 700, loss[loss=0.3171, simple_loss=0.3559, pruned_loss=0.1392, over 4932.00 frames. ], tot_loss[loss=0.3059, simple_loss=0.3412, pruned_loss=0.1353, over 923334.38 frames. ], batch size: 33, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:32:12,908 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.6495, 4.9003, 5.1736, 5.4928, 5.2952, 5.1233, 5.7058, 2.4790], + device='cuda:4'), covar=tensor([0.0596, 0.0791, 0.0640, 0.0626, 0.1223, 0.1053, 0.0495, 0.4334], + device='cuda:4'), in_proj_covar=tensor([0.0371, 0.0248, 0.0271, 0.0298, 0.0350, 0.0292, 0.0313, 0.0304], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:32:22,199 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6459.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:32:24,545 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6462.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:32:27,974 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6467.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:32:34,494 INFO [finetune.py:976] (4/7) Epoch 2, batch 750, loss[loss=0.3161, simple_loss=0.3338, pruned_loss=0.1492, over 4690.00 frames. ], tot_loss[loss=0.3053, simple_loss=0.341, pruned_loss=0.1348, over 928710.17 frames. ], batch size: 23, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:32:42,500 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6491.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:32:58,113 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-25 23:32:58,322 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6514.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:32:58,815 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.464e+02 1.998e+02 2.267e+02 2.688e+02 5.596e+02, threshold=4.534e+02, percent-clipped=2.0 +2023-03-25 23:33:04,283 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6523.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 23:33:06,050 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6526.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:33:07,184 INFO [finetune.py:976] (4/7) Epoch 2, batch 800, loss[loss=0.2519, simple_loss=0.3152, pruned_loss=0.09426, over 4765.00 frames. ], tot_loss[loss=0.3031, simple_loss=0.3391, pruned_loss=0.1335, over 932718.82 frames. ], batch size: 27, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:33:07,393 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6528.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:33:21,406 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-25 23:33:40,462 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6559.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:33:42,336 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6562.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:33:46,530 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6567.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:33:58,060 INFO [finetune.py:976] (4/7) Epoch 2, batch 850, loss[loss=0.2821, simple_loss=0.3236, pruned_loss=0.1203, over 4789.00 frames. ], tot_loss[loss=0.2991, simple_loss=0.3357, pruned_loss=0.1312, over 938268.42 frames. ], batch size: 29, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:34:37,098 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.216e+02 1.789e+02 2.218e+02 2.697e+02 5.451e+02, threshold=4.436e+02, percent-clipped=1.0 +2023-03-25 23:34:42,623 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6622.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:34:46,287 INFO [finetune.py:976] (4/7) Epoch 2, batch 900, loss[loss=0.2898, simple_loss=0.3183, pruned_loss=0.1307, over 4849.00 frames. ], tot_loss[loss=0.293, simple_loss=0.3303, pruned_loss=0.1278, over 944197.42 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:34:47,130 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6629.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:35:01,647 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4524, 1.2134, 1.2253, 0.9750, 1.6144, 1.2727, 1.4769, 1.3706], + device='cuda:4'), covar=tensor([0.3358, 0.6328, 0.6699, 0.6534, 0.4477, 0.3194, 0.4517, 0.4336], + device='cuda:4'), in_proj_covar=tensor([0.0158, 0.0188, 0.0230, 0.0243, 0.0203, 0.0175, 0.0191, 0.0182], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:35:02,805 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6646.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:35:25,034 INFO [finetune.py:976] (4/7) Epoch 2, batch 950, loss[loss=0.2428, simple_loss=0.2778, pruned_loss=0.1039, over 4195.00 frames. ], tot_loss[loss=0.2901, simple_loss=0.3271, pruned_loss=0.1266, over 945208.22 frames. ], batch size: 18, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:35:25,165 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1995, 1.7580, 1.4244, 1.2855, 2.1037, 2.5755, 2.0792, 1.7148], + device='cuda:4'), covar=tensor([0.0219, 0.0372, 0.0478, 0.0461, 0.0253, 0.0218, 0.0234, 0.0357], + device='cuda:4'), in_proj_covar=tensor([0.0082, 0.0112, 0.0131, 0.0111, 0.0103, 0.0098, 0.0088, 0.0108], + device='cuda:4'), out_proj_covar=tensor([6.4526e-05, 8.8240e-05, 1.0577e-04, 8.8144e-05, 8.1760e-05, 7.3266e-05, + 6.7708e-05, 8.5001e-05], device='cuda:4') +2023-03-25 23:35:32,461 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6690.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:35:34,864 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6694.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:35:37,925 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6699.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:35:37,977 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8400, 1.4219, 1.6334, 1.5439, 1.3549, 1.4052, 1.4903, 1.6195], + device='cuda:4'), covar=tensor([2.6435, 5.0614, 3.3269, 4.4524, 4.8611, 3.1170, 6.0759, 3.1167], + device='cuda:4'), in_proj_covar=tensor([0.0217, 0.0250, 0.0235, 0.0263, 0.0241, 0.0213, 0.0271, 0.0209], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-25 23:35:48,888 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.729e+02 2.077e+02 2.706e+02 4.933e+02, threshold=4.155e+02, percent-clipped=1.0 +2023-03-25 23:36:03,625 INFO [finetune.py:976] (4/7) Epoch 2, batch 1000, loss[loss=0.377, simple_loss=0.397, pruned_loss=0.1785, over 4846.00 frames. ], tot_loss[loss=0.2917, simple_loss=0.3289, pruned_loss=0.1273, over 946492.52 frames. ], batch size: 44, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:36:28,552 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6747.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:37:01,238 INFO [finetune.py:976] (4/7) Epoch 2, batch 1050, loss[loss=0.3048, simple_loss=0.3465, pruned_loss=0.1316, over 4739.00 frames. ], tot_loss[loss=0.2955, simple_loss=0.3329, pruned_loss=0.129, over 949148.06 frames. ], batch size: 59, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:37:09,221 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6791.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:37:19,919 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1287, 1.5901, 1.0459, 1.9193, 2.2366, 1.5000, 1.8728, 1.9206], + device='cuda:4'), covar=tensor([0.1651, 0.2018, 0.2216, 0.1198, 0.2219, 0.2159, 0.1346, 0.2219], + device='cuda:4'), in_proj_covar=tensor([0.0092, 0.0096, 0.0115, 0.0092, 0.0123, 0.0096, 0.0098, 0.0093], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-25 23:37:29,665 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.307e+02 2.050e+02 2.547e+02 2.958e+02 5.414e+02, threshold=5.095e+02, percent-clipped=8.0 +2023-03-25 23:37:37,329 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6818.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 23:37:40,272 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6823.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:37:47,837 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6826.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:37:48,961 INFO [finetune.py:976] (4/7) Epoch 2, batch 1100, loss[loss=0.3515, simple_loss=0.3708, pruned_loss=0.1662, over 4800.00 frames. ], tot_loss[loss=0.2968, simple_loss=0.3347, pruned_loss=0.1294, over 952077.11 frames. ], batch size: 51, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:37:56,678 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6839.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:38:18,885 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6859.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:38:23,766 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6867.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:38:28,921 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6874.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:38:31,313 INFO [finetune.py:976] (4/7) Epoch 2, batch 1150, loss[loss=0.3063, simple_loss=0.3579, pruned_loss=0.1273, over 4808.00 frames. ], tot_loss[loss=0.2982, simple_loss=0.3365, pruned_loss=0.13, over 954451.61 frames. ], batch size: 39, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:38:31,434 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6878.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 23:38:46,250 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-25 23:38:52,178 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6907.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:38:53,391 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6909.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:38:56,972 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.223e+02 2.014e+02 2.348e+02 2.865e+02 5.036e+02, threshold=4.696e+02, percent-clipped=0.0 +2023-03-25 23:38:57,046 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6915.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:39:07,708 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6922.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:39:07,747 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5425, 1.3826, 1.2252, 1.1581, 1.7190, 1.7046, 1.3966, 1.1349], + device='cuda:4'), covar=tensor([0.0281, 0.0391, 0.0591, 0.0390, 0.0266, 0.0301, 0.0368, 0.0435], + device='cuda:4'), in_proj_covar=tensor([0.0082, 0.0112, 0.0131, 0.0112, 0.0103, 0.0098, 0.0088, 0.0109], + device='cuda:4'), out_proj_covar=tensor([6.4705e-05, 8.8486e-05, 1.0632e-04, 8.8722e-05, 8.2211e-05, 7.3208e-05, + 6.8192e-05, 8.5234e-05], device='cuda:4') +2023-03-25 23:39:17,352 INFO [finetune.py:976] (4/7) Epoch 2, batch 1200, loss[loss=0.2487, simple_loss=0.3115, pruned_loss=0.09297, over 4768.00 frames. ], tot_loss[loss=0.2965, simple_loss=0.3347, pruned_loss=0.1291, over 954493.90 frames. ], batch size: 28, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:39:17,473 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8750, 2.0710, 1.9101, 1.3341, 2.3389, 2.2062, 1.9960, 1.9241], + device='cuda:4'), covar=tensor([0.0923, 0.0691, 0.0858, 0.1201, 0.0401, 0.0764, 0.0869, 0.1116], + device='cuda:4'), in_proj_covar=tensor([0.0136, 0.0132, 0.0139, 0.0128, 0.0108, 0.0138, 0.0144, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:39:17,848 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-25 23:39:31,058 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6939.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 23:39:49,902 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=6970.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:39:49,979 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6970.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:39:56,186 INFO [finetune.py:976] (4/7) Epoch 2, batch 1250, loss[loss=0.2679, simple_loss=0.3122, pruned_loss=0.1119, over 4859.00 frames. ], tot_loss[loss=0.2905, simple_loss=0.3294, pruned_loss=0.1258, over 953887.55 frames. ], batch size: 47, lr: 4.00e-03, grad_scale: 16.0 +2023-03-25 23:40:00,603 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6985.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:40:23,991 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.288e+02 1.918e+02 2.412e+02 2.798e+02 4.765e+02, threshold=4.825e+02, percent-clipped=1.0 +2023-03-25 23:40:30,942 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8278, 1.1317, 0.7602, 1.6834, 2.1422, 1.2268, 1.4176, 1.7368], + device='cuda:4'), covar=tensor([0.1707, 0.2368, 0.2443, 0.1292, 0.2216, 0.2112, 0.1429, 0.1971], + device='cuda:4'), in_proj_covar=tensor([0.0092, 0.0096, 0.0115, 0.0092, 0.0123, 0.0095, 0.0098, 0.0093], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-25 23:40:39,102 INFO [finetune.py:976] (4/7) Epoch 2, batch 1300, loss[loss=0.2918, simple_loss=0.3058, pruned_loss=0.1389, over 4160.00 frames. ], tot_loss[loss=0.2888, simple_loss=0.3271, pruned_loss=0.1252, over 953388.80 frames. ], batch size: 18, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:40:52,831 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7048.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 23:41:17,563 INFO [finetune.py:976] (4/7) Epoch 2, batch 1350, loss[loss=0.2809, simple_loss=0.3227, pruned_loss=0.1195, over 4901.00 frames. ], tot_loss[loss=0.2904, simple_loss=0.3286, pruned_loss=0.1262, over 954954.80 frames. ], batch size: 36, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:41:41,164 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7109.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 23:41:47,303 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.456e+02 1.876e+02 2.208e+02 2.586e+02 5.614e+02, threshold=4.416e+02, percent-clipped=5.0 +2023-03-25 23:41:49,187 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7118.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 23:41:52,294 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7123.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:41:55,447 INFO [finetune.py:976] (4/7) Epoch 2, batch 1400, loss[loss=0.2588, simple_loss=0.3128, pruned_loss=0.1024, over 4762.00 frames. ], tot_loss[loss=0.2941, simple_loss=0.3327, pruned_loss=0.1277, over 956890.66 frames. ], batch size: 27, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:42:02,445 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=2.04 vs. limit=2.0 +2023-03-25 23:42:28,988 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.62 vs. limit=2.0 +2023-03-25 23:42:31,262 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7163.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:42:33,029 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7166.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:42:36,038 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7171.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:42:40,186 INFO [finetune.py:976] (4/7) Epoch 2, batch 1450, loss[loss=0.2904, simple_loss=0.3299, pruned_loss=0.1254, over 4818.00 frames. ], tot_loss[loss=0.2965, simple_loss=0.3356, pruned_loss=0.1287, over 957790.49 frames. ], batch size: 39, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:42:51,380 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=2.01 vs. limit=2.0 +2023-03-25 23:43:29,961 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.933e+02 2.198e+02 2.731e+02 4.077e+02, threshold=4.395e+02, percent-clipped=0.0 +2023-03-25 23:43:40,414 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7224.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 23:43:42,757 INFO [finetune.py:976] (4/7) Epoch 2, batch 1500, loss[loss=0.3244, simple_loss=0.3646, pruned_loss=0.1421, over 4806.00 frames. ], tot_loss[loss=0.2985, simple_loss=0.3376, pruned_loss=0.1296, over 957727.95 frames. ], batch size: 40, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:43:51,987 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7234.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 23:44:11,438 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7200, 1.0227, 0.8605, 1.5408, 1.9427, 1.4289, 1.1927, 1.5323], + device='cuda:4'), covar=tensor([0.1733, 0.2550, 0.2383, 0.1345, 0.2375, 0.2061, 0.1698, 0.2100], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0098, 0.0116, 0.0093, 0.0125, 0.0097, 0.0099, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-25 23:44:27,653 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-25 23:44:36,526 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7265.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:44:47,940 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7275.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:44:49,673 INFO [finetune.py:976] (4/7) Epoch 2, batch 1550, loss[loss=0.2374, simple_loss=0.2988, pruned_loss=0.08803, over 4810.00 frames. ], tot_loss[loss=0.2983, simple_loss=0.3373, pruned_loss=0.1297, over 954329.15 frames. ], batch size: 40, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:44:57,965 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7285.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:45:13,037 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-25 23:45:31,203 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.380e+02 1.989e+02 2.300e+02 2.720e+02 4.709e+02, threshold=4.600e+02, percent-clipped=4.0 +2023-03-25 23:45:39,141 INFO [finetune.py:976] (4/7) Epoch 2, batch 1600, loss[loss=0.2236, simple_loss=0.2793, pruned_loss=0.08397, over 4769.00 frames. ], tot_loss[loss=0.2947, simple_loss=0.3336, pruned_loss=0.1279, over 954466.63 frames. ], batch size: 28, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:45:42,236 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7333.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:45:44,174 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7336.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:46:28,861 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6124, 1.5016, 1.9592, 1.3784, 1.6372, 1.8697, 1.7070, 2.0430], + device='cuda:4'), covar=tensor([0.1757, 0.2093, 0.1237, 0.1561, 0.1076, 0.1470, 0.2285, 0.0922], + device='cuda:4'), in_proj_covar=tensor([0.0202, 0.0203, 0.0201, 0.0193, 0.0175, 0.0221, 0.0212, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:46:35,386 INFO [finetune.py:976] (4/7) Epoch 2, batch 1650, loss[loss=0.304, simple_loss=0.3456, pruned_loss=0.1312, over 4806.00 frames. ], tot_loss[loss=0.2913, simple_loss=0.3298, pruned_loss=0.1264, over 953993.91 frames. ], batch size: 29, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:46:52,962 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4774, 2.1294, 1.9985, 2.6355, 2.7127, 2.1793, 3.0871, 2.3472], + device='cuda:4'), covar=tensor([0.2854, 0.7399, 0.5985, 0.5604, 0.3332, 0.2781, 0.3878, 0.3912], + device='cuda:4'), in_proj_covar=tensor([0.0161, 0.0192, 0.0235, 0.0247, 0.0208, 0.0179, 0.0197, 0.0186], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:46:53,655 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-25 23:46:59,571 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7404.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 23:47:01,932 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7407.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:47:07,159 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.344e+02 1.965e+02 2.306e+02 2.769e+02 7.650e+02, threshold=4.611e+02, percent-clipped=3.0 +2023-03-25 23:47:15,140 INFO [finetune.py:976] (4/7) Epoch 2, batch 1700, loss[loss=0.3072, simple_loss=0.3451, pruned_loss=0.1347, over 4835.00 frames. ], tot_loss[loss=0.2883, simple_loss=0.3266, pruned_loss=0.1249, over 954352.25 frames. ], batch size: 39, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:47:22,032 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0177, 2.1469, 2.0168, 1.3544, 2.4595, 2.2966, 2.1639, 2.0213], + device='cuda:4'), covar=tensor([0.0831, 0.0679, 0.0842, 0.1174, 0.0366, 0.0793, 0.0736, 0.1079], + device='cuda:4'), in_proj_covar=tensor([0.0136, 0.0132, 0.0140, 0.0127, 0.0108, 0.0137, 0.0144, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:47:34,994 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-25 23:47:48,738 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7468.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:48:00,185 INFO [finetune.py:976] (4/7) Epoch 2, batch 1750, loss[loss=0.2489, simple_loss=0.2962, pruned_loss=0.1008, over 4820.00 frames. ], tot_loss[loss=0.2886, simple_loss=0.3273, pruned_loss=0.1249, over 954546.59 frames. ], batch size: 25, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:48:35,406 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2320, 1.9695, 1.6927, 0.7608, 1.8530, 1.7328, 1.4623, 1.9558], + device='cuda:4'), covar=tensor([0.1008, 0.0862, 0.1675, 0.2333, 0.1288, 0.2597, 0.2539, 0.1006], + device='cuda:4'), in_proj_covar=tensor([0.0162, 0.0185, 0.0198, 0.0181, 0.0206, 0.0204, 0.0208, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:48:46,574 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.280e+02 2.019e+02 2.351e+02 2.794e+02 5.482e+02, threshold=4.701e+02, percent-clipped=1.0 +2023-03-25 23:48:53,182 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7519.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 23:49:03,492 INFO [finetune.py:976] (4/7) Epoch 2, batch 1800, loss[loss=0.3094, simple_loss=0.354, pruned_loss=0.1325, over 4902.00 frames. ], tot_loss[loss=0.2919, simple_loss=0.3316, pruned_loss=0.1261, over 955975.22 frames. ], batch size: 43, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:49:12,369 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7534.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 23:49:17,293 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-25 23:49:44,260 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7565.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:49:57,759 INFO [finetune.py:976] (4/7) Epoch 2, batch 1850, loss[loss=0.3674, simple_loss=0.3689, pruned_loss=0.1829, over 4881.00 frames. ], tot_loss[loss=0.2942, simple_loss=0.3339, pruned_loss=0.1273, over 956452.55 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:49:57,855 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4982, 1.4384, 1.9736, 2.8610, 2.0586, 2.1229, 1.0079, 2.3808], + device='cuda:4'), covar=tensor([0.1835, 0.1558, 0.1185, 0.0624, 0.0836, 0.1439, 0.1830, 0.0614], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0120, 0.0139, 0.0161, 0.0105, 0.0145, 0.0130, 0.0108], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-25 23:50:05,100 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7582.0, num_to_drop=1, layers_to_drop={0} +2023-03-25 23:50:47,093 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7613.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:50:48,710 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.368e+02 2.140e+02 2.552e+02 3.133e+02 4.516e+02, threshold=5.105e+02, percent-clipped=0.0 +2023-03-25 23:51:00,615 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7624.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:51:02,927 INFO [finetune.py:976] (4/7) Epoch 2, batch 1900, loss[loss=0.3312, simple_loss=0.362, pruned_loss=0.1502, over 4915.00 frames. ], tot_loss[loss=0.2946, simple_loss=0.3347, pruned_loss=0.1272, over 957283.28 frames. ], batch size: 46, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:51:09,870 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7631.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:51:40,753 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1836, 1.8960, 2.5806, 1.6114, 2.3451, 2.3086, 1.9841, 2.4506], + device='cuda:4'), covar=tensor([0.2303, 0.2669, 0.2020, 0.3289, 0.1334, 0.2480, 0.2969, 0.1470], + device='cuda:4'), in_proj_covar=tensor([0.0203, 0.0205, 0.0203, 0.0194, 0.0176, 0.0222, 0.0212, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:51:41,824 INFO [finetune.py:976] (4/7) Epoch 2, batch 1950, loss[loss=0.3348, simple_loss=0.3592, pruned_loss=0.1552, over 4830.00 frames. ], tot_loss[loss=0.2908, simple_loss=0.3315, pruned_loss=0.1251, over 958419.32 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:51:47,809 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7685.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:51:59,349 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7704.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 23:52:07,432 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.355e+02 1.951e+02 2.306e+02 2.786e+02 5.176e+02, threshold=4.611e+02, percent-clipped=1.0 +2023-03-25 23:52:08,728 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7765, 1.5474, 1.2762, 1.6116, 1.7913, 1.4313, 2.1159, 1.6173], + device='cuda:4'), covar=tensor([0.2973, 0.6232, 0.6366, 0.6202, 0.4417, 0.3176, 0.7296, 0.4418], + device='cuda:4'), in_proj_covar=tensor([0.0161, 0.0193, 0.0236, 0.0247, 0.0209, 0.0179, 0.0198, 0.0186], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:52:17,329 INFO [finetune.py:976] (4/7) Epoch 2, batch 2000, loss[loss=0.2483, simple_loss=0.2893, pruned_loss=0.1036, over 4797.00 frames. ], tot_loss[loss=0.2866, simple_loss=0.3272, pruned_loss=0.123, over 956131.45 frames. ], batch size: 25, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:52:26,515 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7743.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:52:31,891 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7752.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 23:52:34,932 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9499, 1.6203, 2.1851, 1.4842, 2.0804, 2.1512, 1.7079, 2.2833], + device='cuda:4'), covar=tensor([0.1660, 0.2211, 0.1593, 0.2303, 0.1019, 0.1659, 0.2506, 0.1108], + device='cuda:4'), in_proj_covar=tensor([0.0203, 0.0205, 0.0203, 0.0194, 0.0176, 0.0221, 0.0212, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:52:37,224 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6952, 1.6625, 1.6478, 1.6784, 1.1754, 2.7471, 1.2081, 1.7575], + device='cuda:4'), covar=tensor([0.2719, 0.1978, 0.1708, 0.1827, 0.1739, 0.0268, 0.2690, 0.1205], + device='cuda:4'), in_proj_covar=tensor([0.0125, 0.0107, 0.0113, 0.0115, 0.0110, 0.0093, 0.0097, 0.0093], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-25 23:52:39,939 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7763.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:52:49,967 INFO [finetune.py:976] (4/7) Epoch 2, batch 2050, loss[loss=0.2586, simple_loss=0.3015, pruned_loss=0.1078, over 4919.00 frames. ], tot_loss[loss=0.2806, simple_loss=0.3211, pruned_loss=0.1201, over 954401.40 frames. ], batch size: 37, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:53:05,258 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7802.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:53:06,504 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7804.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:53:14,628 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.310e+02 1.893e+02 2.255e+02 2.795e+02 6.508e+02, threshold=4.510e+02, percent-clipped=3.0 +2023-03-25 23:53:17,164 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7819.0, num_to_drop=1, layers_to_drop={1} +2023-03-25 23:53:19,418 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3101, 1.2953, 1.5833, 1.1518, 1.2035, 1.5055, 1.3932, 1.6395], + device='cuda:4'), covar=tensor([0.1235, 0.2028, 0.1222, 0.1289, 0.1053, 0.1047, 0.2311, 0.0873], + device='cuda:4'), in_proj_covar=tensor([0.0203, 0.0205, 0.0202, 0.0194, 0.0176, 0.0221, 0.0212, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:53:26,208 INFO [finetune.py:976] (4/7) Epoch 2, batch 2100, loss[loss=0.3205, simple_loss=0.3557, pruned_loss=0.1426, over 4924.00 frames. ], tot_loss[loss=0.2799, simple_loss=0.3206, pruned_loss=0.1196, over 955702.59 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:53:47,053 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1525, 1.8418, 2.6309, 3.9547, 2.8670, 2.6771, 0.6606, 3.1832], + device='cuda:4'), covar=tensor([0.1875, 0.1612, 0.1380, 0.0584, 0.0846, 0.1545, 0.2405, 0.0600], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0119, 0.0139, 0.0162, 0.0105, 0.0145, 0.0131, 0.0108], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-25 23:53:56,807 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7863.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:54:00,168 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7867.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:54:08,359 INFO [finetune.py:976] (4/7) Epoch 2, batch 2150, loss[loss=0.3047, simple_loss=0.3474, pruned_loss=0.131, over 4908.00 frames. ], tot_loss[loss=0.2842, simple_loss=0.3254, pruned_loss=0.1215, over 956294.34 frames. ], batch size: 36, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:54:49,701 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.034e+02 1.983e+02 2.392e+02 2.856e+02 4.131e+02, threshold=4.785e+02, percent-clipped=0.0 +2023-03-25 23:55:10,156 INFO [finetune.py:976] (4/7) Epoch 2, batch 2200, loss[loss=0.2891, simple_loss=0.3232, pruned_loss=0.1276, over 4851.00 frames. ], tot_loss[loss=0.2877, simple_loss=0.3296, pruned_loss=0.1229, over 954745.74 frames. ], batch size: 44, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:55:16,957 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7931.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:55:37,550 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2668, 1.2838, 1.5116, 1.0515, 1.2077, 1.4170, 1.3477, 1.5465], + device='cuda:4'), covar=tensor([0.1600, 0.2430, 0.1474, 0.1564, 0.1179, 0.1324, 0.2765, 0.1086], + device='cuda:4'), in_proj_covar=tensor([0.0202, 0.0204, 0.0202, 0.0193, 0.0175, 0.0221, 0.0212, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:56:10,099 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4853, 2.1175, 1.9833, 0.9491, 2.1497, 1.9755, 1.7174, 2.1011], + device='cuda:4'), covar=tensor([0.0693, 0.1045, 0.1704, 0.2296, 0.1334, 0.1970, 0.2190, 0.1202], + device='cuda:4'), in_proj_covar=tensor([0.0164, 0.0188, 0.0200, 0.0184, 0.0209, 0.0206, 0.0210, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:56:12,551 INFO [finetune.py:976] (4/7) Epoch 2, batch 2250, loss[loss=0.3448, simple_loss=0.3815, pruned_loss=0.154, over 4811.00 frames. ], tot_loss[loss=0.2887, simple_loss=0.3312, pruned_loss=0.1231, over 956480.51 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:56:13,682 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=7979.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:56:19,275 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7980.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:56:40,583 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.2523, 1.2262, 1.3766, 0.7581, 1.0382, 1.4551, 1.5421, 1.3649], + device='cuda:4'), covar=tensor([0.0811, 0.0518, 0.0348, 0.0507, 0.0413, 0.0402, 0.0254, 0.0419], + device='cuda:4'), in_proj_covar=tensor([0.0124, 0.0150, 0.0114, 0.0127, 0.0126, 0.0114, 0.0141, 0.0139], + device='cuda:4'), out_proj_covar=tensor([9.3986e-05, 1.1156e-04, 8.3292e-05, 9.3419e-05, 9.1790e-05, 8.4161e-05, + 1.0521e-04, 1.0332e-04], device='cuda:4') +2023-03-25 23:57:05,131 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.321e+02 1.899e+02 2.269e+02 2.653e+02 4.132e+02, threshold=4.538e+02, percent-clipped=0.0 +2023-03-25 23:57:24,778 INFO [finetune.py:976] (4/7) Epoch 2, batch 2300, loss[loss=0.2897, simple_loss=0.3279, pruned_loss=0.1257, over 4024.00 frames. ], tot_loss[loss=0.289, simple_loss=0.3314, pruned_loss=0.1233, over 955767.49 frames. ], batch size: 65, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:57:34,083 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5779, 1.3092, 1.1098, 1.1586, 1.2932, 1.2528, 1.2033, 2.1230], + device='cuda:4'), covar=tensor([2.4876, 2.6344, 2.0656, 2.8839, 2.0560, 1.3865, 2.5853, 0.7207], + device='cuda:4'), in_proj_covar=tensor([0.0247, 0.0231, 0.0209, 0.0266, 0.0222, 0.0189, 0.0228, 0.0170], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-25 23:57:53,415 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8063.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:57:55,306 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7621, 1.7668, 1.7805, 1.1820, 2.0627, 1.8274, 1.7231, 1.6089], + device='cuda:4'), covar=tensor([0.0761, 0.0690, 0.0742, 0.1021, 0.0509, 0.0860, 0.0836, 0.1221], + device='cuda:4'), in_proj_covar=tensor([0.0137, 0.0132, 0.0140, 0.0129, 0.0108, 0.0138, 0.0145, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:58:02,341 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5505, 1.6049, 1.2326, 1.4041, 1.7433, 1.8674, 1.4745, 1.2299], + device='cuda:4'), covar=tensor([0.0336, 0.0303, 0.0586, 0.0338, 0.0239, 0.0282, 0.0372, 0.0408], + device='cuda:4'), in_proj_covar=tensor([0.0081, 0.0112, 0.0132, 0.0112, 0.0103, 0.0097, 0.0088, 0.0108], + device='cuda:4'), out_proj_covar=tensor([6.3878e-05, 8.8487e-05, 1.0682e-04, 8.8850e-05, 8.1678e-05, 7.2749e-05, + 6.7918e-05, 8.4815e-05], device='cuda:4') +2023-03-25 23:58:05,650 INFO [finetune.py:976] (4/7) Epoch 2, batch 2350, loss[loss=0.2974, simple_loss=0.3291, pruned_loss=0.1328, over 4896.00 frames. ], tot_loss[loss=0.2854, simple_loss=0.3275, pruned_loss=0.1217, over 955943.48 frames. ], batch size: 35, lr: 3.99e-03, grad_scale: 16.0 +2023-03-25 23:58:08,419 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5153, 0.9777, 1.2601, 1.2437, 1.1496, 1.2232, 1.1865, 1.3661], + device='cuda:4'), covar=tensor([1.9343, 3.7051, 2.5748, 3.2070, 3.2239, 2.3155, 4.2808, 2.3037], + device='cuda:4'), in_proj_covar=tensor([0.0222, 0.0256, 0.0242, 0.0268, 0.0245, 0.0218, 0.0277, 0.0213], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-25 23:58:20,621 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=8099.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:58:39,058 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8111.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:58:41,410 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.369e+02 1.949e+02 2.211e+02 2.649e+02 5.737e+02, threshold=4.422e+02, percent-clipped=2.0 +2023-03-25 23:59:01,111 INFO [finetune.py:976] (4/7) Epoch 2, batch 2400, loss[loss=0.261, simple_loss=0.2903, pruned_loss=0.1159, over 4740.00 frames. ], tot_loss[loss=0.2822, simple_loss=0.3238, pruned_loss=0.1203, over 956863.28 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 32.0 +2023-03-25 23:59:03,938 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.3164, 2.9594, 3.0127, 3.2444, 3.0680, 2.9039, 3.3834, 1.0263], + device='cuda:4'), covar=tensor([0.1075, 0.0997, 0.1016, 0.1140, 0.1661, 0.1496, 0.1092, 0.5079], + device='cuda:4'), in_proj_covar=tensor([0.0371, 0.0248, 0.0275, 0.0298, 0.0348, 0.0290, 0.0316, 0.0306], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-25 23:59:27,570 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=8158.0, num_to_drop=0, layers_to_drop=set() +2023-03-25 23:59:39,736 INFO [finetune.py:976] (4/7) Epoch 2, batch 2450, loss[loss=0.2652, simple_loss=0.3057, pruned_loss=0.1123, over 4830.00 frames. ], tot_loss[loss=0.2801, simple_loss=0.3211, pruned_loss=0.1196, over 955616.53 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:00:02,529 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8248, 1.2102, 0.9202, 1.6786, 2.0776, 1.3867, 1.4429, 1.7509], + device='cuda:4'), covar=tensor([0.1506, 0.2242, 0.2255, 0.1290, 0.2190, 0.2062, 0.1400, 0.1900], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0098, 0.0117, 0.0093, 0.0125, 0.0097, 0.0099, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 00:00:10,508 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1113, 2.6765, 1.8682, 1.5866, 2.9018, 2.6227, 2.2203, 2.3106], + device='cuda:4'), covar=tensor([0.1011, 0.0596, 0.1232, 0.1388, 0.0504, 0.0939, 0.0990, 0.1030], + device='cuda:4'), in_proj_covar=tensor([0.0136, 0.0131, 0.0140, 0.0128, 0.0107, 0.0137, 0.0144, 0.0159], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:00:20,950 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.358e+02 1.915e+02 2.165e+02 2.652e+02 4.234e+02, threshold=4.330e+02, percent-clipped=0.0 +2023-03-26 00:00:22,900 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0210, 1.6405, 1.8720, 0.8061, 2.0718, 2.5087, 1.6661, 1.8489], + device='cuda:4'), covar=tensor([0.1593, 0.1330, 0.0950, 0.1115, 0.1045, 0.0494, 0.0814, 0.0751], + device='cuda:4'), in_proj_covar=tensor([0.0124, 0.0150, 0.0113, 0.0127, 0.0126, 0.0114, 0.0140, 0.0139], + device='cuda:4'), out_proj_covar=tensor([9.3555e-05, 1.1158e-04, 8.3277e-05, 9.3325e-05, 9.1409e-05, 8.3967e-05, + 1.0488e-04, 1.0324e-04], device='cuda:4') +2023-03-26 00:00:31,376 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 00:00:33,536 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6953, 1.4691, 1.0798, 0.3424, 1.2980, 1.4742, 1.2731, 1.4830], + device='cuda:4'), covar=tensor([0.0835, 0.0838, 0.1479, 0.2125, 0.1308, 0.2609, 0.2448, 0.0906], + device='cuda:4'), in_proj_covar=tensor([0.0161, 0.0186, 0.0197, 0.0181, 0.0207, 0.0203, 0.0207, 0.0193], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:00:34,018 INFO [finetune.py:976] (4/7) Epoch 2, batch 2500, loss[loss=0.3673, simple_loss=0.3926, pruned_loss=0.171, over 4822.00 frames. ], tot_loss[loss=0.2803, simple_loss=0.3214, pruned_loss=0.1195, over 956953.09 frames. ], batch size: 40, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:01:11,699 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8265.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:01:30,123 INFO [finetune.py:976] (4/7) Epoch 2, batch 2550, loss[loss=0.3086, simple_loss=0.3548, pruned_loss=0.1312, over 4798.00 frames. ], tot_loss[loss=0.2838, simple_loss=0.3261, pruned_loss=0.1208, over 955830.09 frames. ], batch size: 45, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:01:31,480 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8280.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:02:02,895 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.611e+02 2.134e+02 2.551e+02 3.097e+02 6.135e+02, threshold=5.102e+02, percent-clipped=4.0 +2023-03-26 00:02:09,751 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=8326.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:02:10,867 INFO [finetune.py:976] (4/7) Epoch 2, batch 2600, loss[loss=0.3144, simple_loss=0.3175, pruned_loss=0.1557, over 4022.00 frames. ], tot_loss[loss=0.2857, simple_loss=0.3277, pruned_loss=0.1219, over 953488.63 frames. ], batch size: 17, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:02:10,929 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8328.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:02:14,794 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1589, 2.0222, 2.7098, 1.6287, 2.3746, 2.2666, 1.9363, 2.5837], + device='cuda:4'), covar=tensor([0.2161, 0.2308, 0.1879, 0.2943, 0.1263, 0.2432, 0.2611, 0.1525], + device='cuda:4'), in_proj_covar=tensor([0.0203, 0.0205, 0.0202, 0.0193, 0.0176, 0.0221, 0.0211, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:02:32,990 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6055, 1.3802, 1.3571, 1.2862, 1.7307, 1.7907, 1.4964, 1.1649], + device='cuda:4'), covar=tensor([0.0232, 0.0339, 0.0513, 0.0385, 0.0250, 0.0257, 0.0278, 0.0434], + device='cuda:4'), in_proj_covar=tensor([0.0082, 0.0113, 0.0132, 0.0113, 0.0103, 0.0098, 0.0088, 0.0109], + device='cuda:4'), out_proj_covar=tensor([6.4251e-05, 8.9270e-05, 1.0713e-04, 8.9335e-05, 8.2196e-05, 7.3230e-05, + 6.7837e-05, 8.5217e-05], device='cuda:4') +2023-03-26 00:03:02,205 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6096, 1.5143, 1.5395, 1.7168, 2.0882, 1.6305, 1.1576, 1.3721], + device='cuda:4'), covar=tensor([0.2631, 0.2674, 0.2108, 0.2031, 0.2225, 0.1464, 0.3426, 0.2128], + device='cuda:4'), in_proj_covar=tensor([0.0220, 0.0202, 0.0189, 0.0175, 0.0224, 0.0167, 0.0205, 0.0177], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:03:09,431 INFO [finetune.py:976] (4/7) Epoch 2, batch 2650, loss[loss=0.2726, simple_loss=0.3264, pruned_loss=0.1094, over 4806.00 frames. ], tot_loss[loss=0.2876, simple_loss=0.3299, pruned_loss=0.1227, over 954469.35 frames. ], batch size: 45, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:03:28,494 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8399.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:03:45,674 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.253e+02 1.975e+02 2.270e+02 2.796e+02 5.066e+02, threshold=4.539e+02, percent-clipped=0.0 +2023-03-26 00:03:50,040 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9654, 2.0344, 1.7718, 2.1187, 1.1455, 4.5391, 1.7026, 2.2634], + device='cuda:4'), covar=tensor([0.3167, 0.2250, 0.1931, 0.2006, 0.1909, 0.0115, 0.2524, 0.1385], + device='cuda:4'), in_proj_covar=tensor([0.0125, 0.0107, 0.0113, 0.0115, 0.0111, 0.0093, 0.0097, 0.0093], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 00:03:57,444 INFO [finetune.py:976] (4/7) Epoch 2, batch 2700, loss[loss=0.3365, simple_loss=0.3595, pruned_loss=0.1568, over 4843.00 frames. ], tot_loss[loss=0.2846, simple_loss=0.3276, pruned_loss=0.1208, over 952926.07 frames. ], batch size: 44, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:04:21,449 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8447.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:04:28,898 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8458.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:04:31,115 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9804, 1.8659, 1.6254, 1.4492, 2.0238, 2.3788, 2.0313, 1.7269], + device='cuda:4'), covar=tensor([0.0375, 0.0439, 0.0566, 0.0494, 0.0417, 0.0364, 0.0376, 0.0493], + device='cuda:4'), in_proj_covar=tensor([0.0082, 0.0113, 0.0133, 0.0113, 0.0104, 0.0098, 0.0088, 0.0109], + device='cuda:4'), out_proj_covar=tensor([6.4286e-05, 8.9359e-05, 1.0722e-04, 8.9406e-05, 8.2251e-05, 7.2988e-05, + 6.8055e-05, 8.5395e-05], device='cuda:4') +2023-03-26 00:04:47,920 INFO [finetune.py:976] (4/7) Epoch 2, batch 2750, loss[loss=0.284, simple_loss=0.3183, pruned_loss=0.1248, over 4870.00 frames. ], tot_loss[loss=0.2796, simple_loss=0.3226, pruned_loss=0.1183, over 954329.50 frames. ], batch size: 31, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:05:15,012 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8506.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:05:25,226 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.170e+02 1.975e+02 2.311e+02 2.901e+02 5.278e+02, threshold=4.621e+02, percent-clipped=1.0 +2023-03-26 00:05:38,378 INFO [finetune.py:976] (4/7) Epoch 2, batch 2800, loss[loss=0.2633, simple_loss=0.3067, pruned_loss=0.11, over 4904.00 frames. ], tot_loss[loss=0.2762, simple_loss=0.3188, pruned_loss=0.1168, over 954665.71 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:06:44,221 INFO [finetune.py:976] (4/7) Epoch 2, batch 2850, loss[loss=0.3451, simple_loss=0.3752, pruned_loss=0.1575, over 4906.00 frames. ], tot_loss[loss=0.2744, simple_loss=0.317, pruned_loss=0.1159, over 955487.52 frames. ], batch size: 43, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:07:08,736 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.223e+02 1.882e+02 2.289e+02 2.777e+02 4.779e+02, threshold=4.578e+02, percent-clipped=1.0 +2023-03-26 00:07:10,625 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-26 00:07:13,831 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=8621.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:07:14,483 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7723, 1.4277, 2.1909, 1.4139, 1.8384, 1.8175, 1.3324, 2.0003], + device='cuda:4'), covar=tensor([0.2073, 0.2542, 0.1826, 0.2257, 0.1469, 0.2013, 0.3103, 0.1434], + device='cuda:4'), in_proj_covar=tensor([0.0205, 0.0207, 0.0205, 0.0195, 0.0178, 0.0223, 0.0214, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:07:15,127 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2476, 2.1454, 2.5280, 1.3212, 2.4233, 2.7881, 2.1645, 2.2269], + device='cuda:4'), covar=tensor([0.1133, 0.0894, 0.0420, 0.0894, 0.0653, 0.0595, 0.0482, 0.0678], + device='cuda:4'), in_proj_covar=tensor([0.0124, 0.0149, 0.0114, 0.0127, 0.0126, 0.0114, 0.0140, 0.0139], + device='cuda:4'), out_proj_covar=tensor([9.3576e-05, 1.1141e-04, 8.3275e-05, 9.3542e-05, 9.1014e-05, 8.3809e-05, + 1.0493e-04, 1.0313e-04], device='cuda:4') +2023-03-26 00:07:18,066 INFO [finetune.py:976] (4/7) Epoch 2, batch 2900, loss[loss=0.3327, simple_loss=0.3655, pruned_loss=0.15, over 4858.00 frames. ], tot_loss[loss=0.2751, simple_loss=0.3179, pruned_loss=0.1162, over 953003.24 frames. ], batch size: 44, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:07:51,240 INFO [finetune.py:976] (4/7) Epoch 2, batch 2950, loss[loss=0.2594, simple_loss=0.3154, pruned_loss=0.1017, over 4758.00 frames. ], tot_loss[loss=0.2789, simple_loss=0.322, pruned_loss=0.1179, over 953273.78 frames. ], batch size: 27, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:08:12,826 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7400, 1.6834, 1.2704, 1.8435, 1.8711, 1.4050, 2.3933, 1.6920], + device='cuda:4'), covar=tensor([0.2997, 0.6935, 0.6596, 0.6372, 0.4392, 0.3153, 0.5492, 0.4146], + device='cuda:4'), in_proj_covar=tensor([0.0163, 0.0195, 0.0237, 0.0251, 0.0213, 0.0181, 0.0202, 0.0187], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:08:17,805 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.163e+02 1.871e+02 2.334e+02 2.881e+02 5.890e+02, threshold=4.669e+02, percent-clipped=2.0 +2023-03-26 00:08:27,818 INFO [finetune.py:976] (4/7) Epoch 2, batch 3000, loss[loss=0.3432, simple_loss=0.3627, pruned_loss=0.1618, over 4243.00 frames. ], tot_loss[loss=0.2805, simple_loss=0.3234, pruned_loss=0.1188, over 950857.78 frames. ], batch size: 65, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:08:27,818 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 00:08:34,591 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9240, 1.4136, 1.1595, 1.7779, 2.1004, 1.3882, 1.6047, 1.7982], + device='cuda:4'), covar=tensor([0.1216, 0.1656, 0.1667, 0.0934, 0.1778, 0.1855, 0.1004, 0.1574], + device='cuda:4'), in_proj_covar=tensor([0.0092, 0.0097, 0.0115, 0.0092, 0.0124, 0.0096, 0.0098, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 00:08:35,183 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1894, 1.9245, 1.5107, 0.6093, 1.6069, 1.8959, 1.7016, 1.8325], + device='cuda:4'), covar=tensor([0.1011, 0.0978, 0.1826, 0.2671, 0.1812, 0.2777, 0.2494, 0.1047], + device='cuda:4'), in_proj_covar=tensor([0.0163, 0.0188, 0.0199, 0.0183, 0.0209, 0.0204, 0.0209, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:08:43,569 INFO [finetune.py:1010] (4/7) Epoch 2, validation: loss=0.1956, simple_loss=0.2636, pruned_loss=0.06384, over 2265189.00 frames. +2023-03-26 00:08:43,569 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6301MB +2023-03-26 00:09:09,168 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8760.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:09:26,207 INFO [finetune.py:976] (4/7) Epoch 2, batch 3050, loss[loss=0.2763, simple_loss=0.3214, pruned_loss=0.1156, over 4882.00 frames. ], tot_loss[loss=0.2796, simple_loss=0.3237, pruned_loss=0.1177, over 952782.23 frames. ], batch size: 35, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:09:26,904 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2083, 3.6002, 3.7853, 4.0657, 3.9607, 3.7332, 4.3055, 1.5225], + device='cuda:4'), covar=tensor([0.0660, 0.0708, 0.0721, 0.0783, 0.1029, 0.1191, 0.0619, 0.4416], + device='cuda:4'), in_proj_covar=tensor([0.0369, 0.0245, 0.0273, 0.0297, 0.0345, 0.0288, 0.0313, 0.0303], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:09:40,524 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6026, 1.5033, 1.5508, 0.8799, 1.5686, 1.8803, 1.6482, 1.4686], + device='cuda:4'), covar=tensor([0.1065, 0.0727, 0.0453, 0.0731, 0.0449, 0.0439, 0.0406, 0.0592], + device='cuda:4'), in_proj_covar=tensor([0.0125, 0.0151, 0.0114, 0.0129, 0.0126, 0.0115, 0.0141, 0.0140], + device='cuda:4'), out_proj_covar=tensor([9.4688e-05, 1.1240e-04, 8.3767e-05, 9.4443e-05, 9.1627e-05, 8.4606e-05, + 1.0560e-04, 1.0361e-04], device='cuda:4') +2023-03-26 00:10:07,396 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.149e+02 1.935e+02 2.296e+02 2.584e+02 4.666e+02, threshold=4.592e+02, percent-clipped=0.0 +2023-03-26 00:10:11,741 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=8821.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:10:16,913 INFO [finetune.py:976] (4/7) Epoch 2, batch 3100, loss[loss=0.267, simple_loss=0.3145, pruned_loss=0.1098, over 4795.00 frames. ], tot_loss[loss=0.2764, simple_loss=0.3208, pruned_loss=0.116, over 952397.43 frames. ], batch size: 45, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:10:27,626 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4821, 0.7373, 1.3221, 1.1539, 1.1417, 1.1248, 1.0503, 1.1762], + device='cuda:4'), covar=tensor([1.5054, 2.9483, 2.0137, 2.4416, 2.6049, 1.8029, 3.2071, 1.9190], + device='cuda:4'), in_proj_covar=tensor([0.0223, 0.0255, 0.0244, 0.0268, 0.0245, 0.0218, 0.0277, 0.0215], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 00:10:55,626 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1077, 2.6630, 2.0383, 1.6543, 3.1032, 2.7562, 2.4128, 2.2741], + device='cuda:4'), covar=tensor([0.0910, 0.0626, 0.0931, 0.1083, 0.0280, 0.0789, 0.0924, 0.1047], + device='cuda:4'), in_proj_covar=tensor([0.0136, 0.0131, 0.0141, 0.0128, 0.0107, 0.0138, 0.0145, 0.0159], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:10:59,136 INFO [finetune.py:976] (4/7) Epoch 2, batch 3150, loss[loss=0.2735, simple_loss=0.3143, pruned_loss=0.1164, over 4772.00 frames. ], tot_loss[loss=0.2747, simple_loss=0.3183, pruned_loss=0.1156, over 953801.45 frames. ], batch size: 28, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:11:03,383 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4959, 1.2231, 1.2089, 1.2308, 1.5443, 1.5771, 1.3810, 1.1399], + device='cuda:4'), covar=tensor([0.0242, 0.0370, 0.0536, 0.0348, 0.0276, 0.0319, 0.0332, 0.0434], + device='cuda:4'), in_proj_covar=tensor([0.0082, 0.0113, 0.0133, 0.0112, 0.0103, 0.0097, 0.0088, 0.0108], + device='cuda:4'), out_proj_covar=tensor([6.4334e-05, 8.9150e-05, 1.0719e-04, 8.9266e-05, 8.2058e-05, 7.2557e-05, + 6.7780e-05, 8.4587e-05], device='cuda:4') +2023-03-26 00:11:28,134 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.188e+02 1.838e+02 2.165e+02 2.835e+02 5.909e+02, threshold=4.329e+02, percent-clipped=1.0 +2023-03-26 00:11:33,406 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8921.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:11:43,241 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8927.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:11:43,729 INFO [finetune.py:976] (4/7) Epoch 2, batch 3200, loss[loss=0.3175, simple_loss=0.3464, pruned_loss=0.1443, over 4935.00 frames. ], tot_loss[loss=0.2702, simple_loss=0.3133, pruned_loss=0.1135, over 955290.76 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:12:25,336 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=8969.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:12:30,758 INFO [finetune.py:976] (4/7) Epoch 2, batch 3250, loss[loss=0.2784, simple_loss=0.3213, pruned_loss=0.1178, over 4765.00 frames. ], tot_loss[loss=0.2708, simple_loss=0.3138, pruned_loss=0.114, over 954674.33 frames. ], batch size: 27, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:12:36,304 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8986.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:12:37,527 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=8988.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:12:44,510 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.71 vs. limit=2.0 +2023-03-26 00:13:02,651 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9013.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:13:03,778 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.184e+02 1.894e+02 2.281e+02 2.922e+02 4.541e+02, threshold=4.561e+02, percent-clipped=3.0 +2023-03-26 00:13:09,828 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=5.05 vs. limit=5.0 +2023-03-26 00:13:11,710 INFO [finetune.py:976] (4/7) Epoch 2, batch 3300, loss[loss=0.2919, simple_loss=0.3358, pruned_loss=0.124, over 4906.00 frames. ], tot_loss[loss=0.2755, simple_loss=0.3185, pruned_loss=0.1162, over 953576.28 frames. ], batch size: 37, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:13:19,243 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-03-26 00:13:25,345 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9047.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:13:42,667 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9074.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:13:44,937 INFO [finetune.py:976] (4/7) Epoch 2, batch 3350, loss[loss=0.2951, simple_loss=0.3348, pruned_loss=0.1278, over 4857.00 frames. ], tot_loss[loss=0.2774, simple_loss=0.3206, pruned_loss=0.1171, over 953432.14 frames. ], batch size: 34, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:14:16,991 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=2.01 vs. limit=2.0 +2023-03-26 00:14:20,187 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.430e+02 1.942e+02 2.307e+02 2.996e+02 6.023e+02, threshold=4.614e+02, percent-clipped=2.0 +2023-03-26 00:14:20,877 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9116.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:14:26,093 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 00:14:28,092 INFO [finetune.py:976] (4/7) Epoch 2, batch 3400, loss[loss=0.28, simple_loss=0.3248, pruned_loss=0.1176, over 4886.00 frames. ], tot_loss[loss=0.2799, simple_loss=0.3228, pruned_loss=0.1185, over 952833.48 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:14:52,346 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 00:15:03,027 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-26 00:15:23,465 INFO [finetune.py:976] (4/7) Epoch 2, batch 3450, loss[loss=0.2889, simple_loss=0.3298, pruned_loss=0.124, over 4891.00 frames. ], tot_loss[loss=0.2802, simple_loss=0.3231, pruned_loss=0.1186, over 953732.77 frames. ], batch size: 37, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:15:59,482 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.266e+02 1.993e+02 2.349e+02 2.850e+02 4.291e+02, threshold=4.698e+02, percent-clipped=0.0 +2023-03-26 00:16:12,593 INFO [finetune.py:976] (4/7) Epoch 2, batch 3500, loss[loss=0.1868, simple_loss=0.243, pruned_loss=0.06533, over 4895.00 frames. ], tot_loss[loss=0.2756, simple_loss=0.3187, pruned_loss=0.1162, over 955356.43 frames. ], batch size: 43, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:16:46,211 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 00:17:13,578 INFO [finetune.py:976] (4/7) Epoch 2, batch 3550, loss[loss=0.2066, simple_loss=0.2648, pruned_loss=0.07421, over 4763.00 frames. ], tot_loss[loss=0.2707, simple_loss=0.3144, pruned_loss=0.1135, over 955289.26 frames. ], batch size: 28, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:17:16,752 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9283.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:17:53,394 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.215e+02 1.769e+02 2.188e+02 2.766e+02 5.069e+02, threshold=4.376e+02, percent-clipped=2.0 +2023-03-26 00:18:09,362 INFO [finetune.py:976] (4/7) Epoch 2, batch 3600, loss[loss=0.2489, simple_loss=0.2973, pruned_loss=0.1002, over 4863.00 frames. ], tot_loss[loss=0.2675, simple_loss=0.3111, pruned_loss=0.1119, over 956798.23 frames. ], batch size: 31, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:18:21,815 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.2731, 2.5734, 2.3813, 1.3170, 2.5968, 2.3024, 2.0452, 2.3266], + device='cuda:4'), covar=tensor([0.0736, 0.1206, 0.2400, 0.2875, 0.2030, 0.2356, 0.2457, 0.1648], + device='cuda:4'), in_proj_covar=tensor([0.0162, 0.0189, 0.0199, 0.0184, 0.0208, 0.0204, 0.0210, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:18:22,988 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9342.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:18:45,804 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9369.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:18:51,740 INFO [finetune.py:976] (4/7) Epoch 2, batch 3650, loss[loss=0.2476, simple_loss=0.2982, pruned_loss=0.09846, over 4768.00 frames. ], tot_loss[loss=0.2677, simple_loss=0.3117, pruned_loss=0.1118, over 955425.16 frames. ], batch size: 28, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:19:24,491 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9414.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:19:24,982 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.270e+02 1.902e+02 2.269e+02 2.850e+02 5.426e+02, threshold=4.539e+02, percent-clipped=4.0 +2023-03-26 00:19:31,918 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9416.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:19:45,283 INFO [finetune.py:976] (4/7) Epoch 2, batch 3700, loss[loss=0.2808, simple_loss=0.3251, pruned_loss=0.1182, over 4817.00 frames. ], tot_loss[loss=0.2708, simple_loss=0.3153, pruned_loss=0.1132, over 954585.15 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:20:05,111 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9276, 1.7614, 1.5103, 1.4050, 1.9246, 2.1554, 1.7754, 1.3856], + device='cuda:4'), covar=tensor([0.0205, 0.0304, 0.0546, 0.0373, 0.0247, 0.0250, 0.0271, 0.0407], + device='cuda:4'), in_proj_covar=tensor([0.0082, 0.0112, 0.0132, 0.0112, 0.0103, 0.0097, 0.0086, 0.0108], + device='cuda:4'), out_proj_covar=tensor([6.3882e-05, 8.8702e-05, 1.0699e-04, 8.8793e-05, 8.1464e-05, 7.2068e-05, + 6.6660e-05, 8.4443e-05], device='cuda:4') +2023-03-26 00:20:15,847 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=9464.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:20:20,474 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8926, 1.3852, 1.0708, 1.6788, 2.1682, 1.4596, 1.5178, 1.8246], + device='cuda:4'), covar=tensor([0.1643, 0.2191, 0.2125, 0.1359, 0.2266, 0.2169, 0.1458, 0.2092], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0098, 0.0118, 0.0094, 0.0126, 0.0097, 0.0100, 0.0096], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 00:20:23,957 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9475.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:20:26,040 INFO [finetune.py:976] (4/7) Epoch 2, batch 3750, loss[loss=0.2672, simple_loss=0.3209, pruned_loss=0.1067, over 4739.00 frames. ], tot_loss[loss=0.2718, simple_loss=0.3166, pruned_loss=0.1135, over 954103.96 frames. ], batch size: 54, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:20:55,391 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.357e+02 1.893e+02 2.405e+02 2.686e+02 6.929e+02, threshold=4.810e+02, percent-clipped=1.0 +2023-03-26 00:21:01,427 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0690, 2.1935, 2.0216, 1.3375, 2.3541, 2.2686, 2.1041, 1.9091], + device='cuda:4'), covar=tensor([0.0820, 0.0696, 0.0909, 0.1245, 0.0436, 0.0824, 0.0922, 0.1088], + device='cuda:4'), in_proj_covar=tensor([0.0138, 0.0133, 0.0143, 0.0130, 0.0109, 0.0140, 0.0147, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:21:10,671 INFO [finetune.py:976] (4/7) Epoch 2, batch 3800, loss[loss=0.2748, simple_loss=0.333, pruned_loss=0.1083, over 4922.00 frames. ], tot_loss[loss=0.276, simple_loss=0.3206, pruned_loss=0.1157, over 953561.41 frames. ], batch size: 42, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:21:44,958 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-26 00:22:03,072 INFO [finetune.py:976] (4/7) Epoch 2, batch 3850, loss[loss=0.1961, simple_loss=0.254, pruned_loss=0.06916, over 4701.00 frames. ], tot_loss[loss=0.2746, simple_loss=0.3193, pruned_loss=0.115, over 952823.70 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:22:07,250 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6767, 1.6145, 1.5485, 1.6935, 1.1722, 3.4932, 1.3591, 1.9417], + device='cuda:4'), covar=tensor([0.3483, 0.2463, 0.2002, 0.2268, 0.1994, 0.0181, 0.2929, 0.1490], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0108, 0.0114, 0.0116, 0.0112, 0.0094, 0.0098, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 00:22:07,253 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9583.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:22:07,849 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9584.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:22:07,874 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4968, 1.1767, 1.2787, 1.1202, 1.6136, 1.6694, 1.4266, 1.1325], + device='cuda:4'), covar=tensor([0.0237, 0.0381, 0.0534, 0.0363, 0.0261, 0.0326, 0.0275, 0.0465], + device='cuda:4'), in_proj_covar=tensor([0.0082, 0.0112, 0.0132, 0.0112, 0.0102, 0.0097, 0.0086, 0.0108], + device='cuda:4'), out_proj_covar=tensor([6.4058e-05, 8.8607e-05, 1.0689e-04, 8.8806e-05, 8.1414e-05, 7.2361e-05, + 6.6649e-05, 8.4316e-05], device='cuda:4') +2023-03-26 00:22:09,659 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7626, 1.5079, 2.0195, 3.3658, 2.4122, 2.4289, 1.1625, 2.5890], + device='cuda:4'), covar=tensor([0.1988, 0.1866, 0.1638, 0.0652, 0.0882, 0.1586, 0.2031, 0.0848], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0120, 0.0138, 0.0163, 0.0105, 0.0146, 0.0131, 0.0108], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 00:22:33,609 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.288e+02 1.890e+02 2.331e+02 2.867e+02 5.576e+02, threshold=4.662e+02, percent-clipped=3.0 +2023-03-26 00:22:48,505 INFO [finetune.py:976] (4/7) Epoch 2, batch 3900, loss[loss=0.3309, simple_loss=0.3394, pruned_loss=0.1612, over 4247.00 frames. ], tot_loss[loss=0.2726, simple_loss=0.3165, pruned_loss=0.1143, over 952612.36 frames. ], batch size: 65, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:22:55,984 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=9631.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:23:08,231 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9642.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:23:10,117 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9645.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:23:28,457 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9658.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:23:40,494 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9669.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:23:46,465 INFO [finetune.py:976] (4/7) Epoch 2, batch 3950, loss[loss=0.2766, simple_loss=0.31, pruned_loss=0.1216, over 4747.00 frames. ], tot_loss[loss=0.2669, simple_loss=0.3114, pruned_loss=0.1112, over 955275.91 frames. ], batch size: 54, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:24:00,118 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=9690.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:24:28,617 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.967e+02 2.302e+02 2.784e+02 7.100e+02, threshold=4.604e+02, percent-clipped=2.0 +2023-03-26 00:24:29,311 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=9717.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:24:30,626 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9719.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:24:33,521 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6032, 1.4057, 1.9763, 2.7860, 2.0096, 2.0658, 1.1221, 2.2498], + device='cuda:4'), covar=tensor([0.1612, 0.1501, 0.1084, 0.0612, 0.0780, 0.1711, 0.1607, 0.0653], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0120, 0.0138, 0.0163, 0.0105, 0.0146, 0.0131, 0.0108], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 00:24:36,433 INFO [finetune.py:976] (4/7) Epoch 2, batch 4000, loss[loss=0.2437, simple_loss=0.2794, pruned_loss=0.104, over 4223.00 frames. ], tot_loss[loss=0.2662, simple_loss=0.3102, pruned_loss=0.1111, over 955190.95 frames. ], batch size: 18, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:24:38,636 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.88 vs. limit=5.0 +2023-03-26 00:24:47,533 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 00:24:48,458 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9743.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:25:00,840 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-03-26 00:25:10,941 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9770.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:25:16,247 INFO [finetune.py:976] (4/7) Epoch 2, batch 4050, loss[loss=0.2957, simple_loss=0.3442, pruned_loss=0.1236, over 4777.00 frames. ], tot_loss[loss=0.269, simple_loss=0.3132, pruned_loss=0.1124, over 953992.98 frames. ], batch size: 54, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:25:37,137 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9804.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:25:44,425 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.258e+02 1.934e+02 2.184e+02 2.585e+02 5.351e+02, threshold=4.368e+02, percent-clipped=2.0 +2023-03-26 00:25:54,159 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 00:25:57,344 INFO [finetune.py:976] (4/7) Epoch 2, batch 4100, loss[loss=0.2713, simple_loss=0.3261, pruned_loss=0.1082, over 4905.00 frames. ], tot_loss[loss=0.2729, simple_loss=0.3173, pruned_loss=0.1142, over 952061.18 frames. ], batch size: 37, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:26:34,082 INFO [finetune.py:976] (4/7) Epoch 2, batch 4150, loss[loss=0.3455, simple_loss=0.363, pruned_loss=0.164, over 4185.00 frames. ], tot_loss[loss=0.2767, simple_loss=0.3205, pruned_loss=0.1165, over 947695.43 frames. ], batch size: 66, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:26:37,265 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6032, 1.4655, 1.1800, 1.1972, 1.3692, 1.3165, 1.3385, 2.1641], + device='cuda:4'), covar=tensor([1.7797, 1.6799, 1.3650, 1.8994, 1.3364, 0.9319, 1.6316, 0.4718], + device='cuda:4'), in_proj_covar=tensor([0.0255, 0.0235, 0.0212, 0.0272, 0.0227, 0.0191, 0.0231, 0.0174], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 00:26:46,910 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.8361, 3.3533, 3.4546, 3.7159, 3.5471, 3.3575, 3.9171, 1.3542], + device='cuda:4'), covar=tensor([0.0790, 0.0716, 0.0847, 0.0888, 0.1273, 0.1367, 0.0800, 0.4514], + device='cuda:4'), in_proj_covar=tensor([0.0370, 0.0245, 0.0275, 0.0297, 0.0344, 0.0288, 0.0313, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:27:05,051 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.284e+02 1.789e+02 2.151e+02 2.605e+02 5.306e+02, threshold=4.302e+02, percent-clipped=2.0 +2023-03-26 00:27:17,334 INFO [finetune.py:976] (4/7) Epoch 2, batch 4200, loss[loss=0.2536, simple_loss=0.3028, pruned_loss=0.1022, over 4895.00 frames. ], tot_loss[loss=0.2751, simple_loss=0.3198, pruned_loss=0.1152, over 947961.52 frames. ], batch size: 35, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:27:25,691 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9940.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:27:57,509 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.89 vs. limit=5.0 +2023-03-26 00:28:05,867 INFO [finetune.py:976] (4/7) Epoch 2, batch 4250, loss[loss=0.2659, simple_loss=0.3134, pruned_loss=0.1091, over 4792.00 frames. ], tot_loss[loss=0.2718, simple_loss=0.3167, pruned_loss=0.1135, over 948779.53 frames. ], batch size: 51, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:28:16,777 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5395, 2.1721, 1.8743, 0.8117, 1.9979, 2.0381, 1.7556, 1.9576], + device='cuda:4'), covar=tensor([0.0801, 0.1052, 0.1634, 0.2396, 0.1484, 0.2264, 0.2182, 0.1153], + device='cuda:4'), in_proj_covar=tensor([0.0163, 0.0191, 0.0200, 0.0184, 0.0210, 0.0205, 0.0211, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:28:33,592 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4126, 1.3000, 1.7629, 2.3705, 1.7048, 2.0805, 1.0036, 1.9793], + device='cuda:4'), covar=tensor([0.1692, 0.1582, 0.1022, 0.0729, 0.0870, 0.1366, 0.1522, 0.0762], + device='cuda:4'), in_proj_covar=tensor([0.0104, 0.0121, 0.0139, 0.0165, 0.0106, 0.0147, 0.0132, 0.0109], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 00:28:37,760 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0847, 1.5122, 1.5216, 1.6290, 1.4566, 1.5632, 1.6370, 1.7553], + device='cuda:4'), covar=tensor([2.0406, 3.7004, 2.6940, 3.4548, 3.6954, 2.4577, 4.3200, 2.2198], + device='cuda:4'), in_proj_covar=tensor([0.0225, 0.0257, 0.0247, 0.0270, 0.0246, 0.0219, 0.0279, 0.0216], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 00:28:41,979 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10014.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:28:47,605 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.000e+02 1.798e+02 2.175e+02 2.768e+02 4.624e+02, threshold=4.351e+02, percent-clipped=3.0 +2023-03-26 00:29:00,126 INFO [finetune.py:976] (4/7) Epoch 2, batch 4300, loss[loss=0.2761, simple_loss=0.3128, pruned_loss=0.1197, over 4811.00 frames. ], tot_loss[loss=0.2694, simple_loss=0.3137, pruned_loss=0.1125, over 950086.74 frames. ], batch size: 51, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:29:12,675 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10039.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:29:28,980 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6725, 1.4983, 1.2118, 1.2964, 1.3821, 1.3796, 1.3409, 2.2564], + device='cuda:4'), covar=tensor([1.7923, 1.7284, 1.3650, 1.9173, 1.4367, 0.9978, 1.7368, 0.5138], + device='cuda:4'), in_proj_covar=tensor([0.0255, 0.0236, 0.0213, 0.0273, 0.0227, 0.0192, 0.0232, 0.0175], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 00:29:36,680 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10062.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 00:29:46,281 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10070.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:29:52,919 INFO [finetune.py:976] (4/7) Epoch 2, batch 4350, loss[loss=0.2799, simple_loss=0.317, pruned_loss=0.1214, over 4746.00 frames. ], tot_loss[loss=0.2659, simple_loss=0.3102, pruned_loss=0.1108, over 951650.01 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:30:06,665 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10099.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:30:07,340 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10100.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 00:30:16,971 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7891, 1.7939, 1.2335, 2.0785, 2.0891, 1.5258, 2.5222, 1.7946], + device='cuda:4'), covar=tensor([0.2664, 0.5754, 0.6133, 0.5545, 0.3668, 0.2774, 0.5088, 0.3955], + device='cuda:4'), in_proj_covar=tensor([0.0163, 0.0196, 0.0238, 0.0253, 0.0215, 0.0182, 0.0205, 0.0188], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:30:25,353 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.331e+02 1.930e+02 2.275e+02 2.717e+02 4.483e+02, threshold=4.550e+02, percent-clipped=1.0 +2023-03-26 00:30:27,187 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10118.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:30:34,448 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10123.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 00:30:37,372 INFO [finetune.py:976] (4/7) Epoch 2, batch 4400, loss[loss=0.2992, simple_loss=0.3443, pruned_loss=0.127, over 4817.00 frames. ], tot_loss[loss=0.267, simple_loss=0.3113, pruned_loss=0.1113, over 952605.47 frames. ], batch size: 39, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:31:22,685 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=5.15 vs. limit=5.0 +2023-03-26 00:31:35,878 INFO [finetune.py:976] (4/7) Epoch 2, batch 4450, loss[loss=0.2364, simple_loss=0.2822, pruned_loss=0.09529, over 4417.00 frames. ], tot_loss[loss=0.2683, simple_loss=0.3135, pruned_loss=0.1115, over 951971.95 frames. ], batch size: 19, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:32:18,680 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.270e+02 1.960e+02 2.424e+02 3.015e+02 7.276e+02, threshold=4.848e+02, percent-clipped=5.0 +2023-03-26 00:32:36,845 INFO [finetune.py:976] (4/7) Epoch 2, batch 4500, loss[loss=0.2341, simple_loss=0.2681, pruned_loss=0.1001, over 4312.00 frames. ], tot_loss[loss=0.2691, simple_loss=0.3147, pruned_loss=0.1117, over 950588.55 frames. ], batch size: 18, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:32:44,230 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10240.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:32:45,195 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 00:33:08,144 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 00:33:19,721 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6712, 1.5877, 1.5368, 1.6708, 1.1134, 3.4046, 1.3232, 1.9595], + device='cuda:4'), covar=tensor([0.3430, 0.2368, 0.2090, 0.2300, 0.2045, 0.0190, 0.3075, 0.1444], + device='cuda:4'), in_proj_covar=tensor([0.0128, 0.0109, 0.0115, 0.0117, 0.0113, 0.0095, 0.0099, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 00:33:20,855 INFO [finetune.py:976] (4/7) Epoch 2, batch 4550, loss[loss=0.2591, simple_loss=0.3141, pruned_loss=0.102, over 4766.00 frames. ], tot_loss[loss=0.2701, simple_loss=0.316, pruned_loss=0.1121, over 951656.50 frames. ], batch size: 51, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:33:32,453 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10288.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:33:59,121 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10314.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:34:00,222 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.209e+02 1.803e+02 2.127e+02 2.576e+02 5.771e+02, threshold=4.255e+02, percent-clipped=1.0 +2023-03-26 00:34:14,432 INFO [finetune.py:976] (4/7) Epoch 2, batch 4600, loss[loss=0.2801, simple_loss=0.3164, pruned_loss=0.1219, over 4929.00 frames. ], tot_loss[loss=0.2682, simple_loss=0.3144, pruned_loss=0.1111, over 949361.49 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:34:25,894 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0395, 2.3648, 1.8517, 1.4751, 2.7748, 2.5881, 2.2327, 1.8059], + device='cuda:4'), covar=tensor([0.1016, 0.0657, 0.1072, 0.1282, 0.0608, 0.0855, 0.1097, 0.1583], + device='cuda:4'), in_proj_covar=tensor([0.0137, 0.0130, 0.0142, 0.0128, 0.0107, 0.0139, 0.0145, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:34:50,216 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10362.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:34:50,292 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5363, 1.3512, 1.4141, 1.6642, 1.8293, 1.5427, 0.9422, 1.3115], + device='cuda:4'), covar=tensor([0.2898, 0.2937, 0.2336, 0.2072, 0.2147, 0.1472, 0.3443, 0.2346], + device='cuda:4'), in_proj_covar=tensor([0.0224, 0.0205, 0.0192, 0.0177, 0.0227, 0.0170, 0.0208, 0.0181], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:35:01,346 INFO [finetune.py:976] (4/7) Epoch 2, batch 4650, loss[loss=0.2473, simple_loss=0.296, pruned_loss=0.09924, over 4786.00 frames. ], tot_loss[loss=0.2664, simple_loss=0.3127, pruned_loss=0.11, over 953082.95 frames. ], batch size: 29, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:35:12,310 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10395.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 00:35:12,397 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0622, 1.4684, 1.6652, 1.6699, 1.5010, 1.4816, 1.6173, 1.6189], + device='cuda:4'), covar=tensor([1.4437, 2.5774, 1.8873, 2.3333, 2.5090, 1.7923, 3.0979, 1.7304], + device='cuda:4'), in_proj_covar=tensor([0.0226, 0.0257, 0.0248, 0.0270, 0.0246, 0.0219, 0.0279, 0.0217], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 00:35:14,753 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10399.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:35:21,430 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10409.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:35:25,534 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.369e+02 1.834e+02 2.117e+02 2.478e+02 4.313e+02, threshold=4.233e+02, percent-clipped=1.0 +2023-03-26 00:35:27,260 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10418.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 00:35:34,606 INFO [finetune.py:976] (4/7) Epoch 2, batch 4700, loss[loss=0.2573, simple_loss=0.2903, pruned_loss=0.1122, over 3115.00 frames. ], tot_loss[loss=0.2644, simple_loss=0.3097, pruned_loss=0.1095, over 950686.50 frames. ], batch size: 13, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:35:46,697 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10447.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:35:50,976 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10454.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:35:56,337 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 00:36:01,579 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10470.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:36:07,253 INFO [finetune.py:976] (4/7) Epoch 2, batch 4750, loss[loss=0.2464, simple_loss=0.293, pruned_loss=0.09992, over 4813.00 frames. ], tot_loss[loss=0.2614, simple_loss=0.3067, pruned_loss=0.108, over 952690.98 frames. ], batch size: 51, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:36:19,144 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6891, 1.6357, 1.4130, 1.3919, 1.9322, 2.0469, 1.6746, 1.3825], + device='cuda:4'), covar=tensor([0.0283, 0.0318, 0.0526, 0.0353, 0.0207, 0.0303, 0.0313, 0.0345], + device='cuda:4'), in_proj_covar=tensor([0.0082, 0.0112, 0.0133, 0.0112, 0.0102, 0.0097, 0.0088, 0.0107], + device='cuda:4'), out_proj_covar=tensor([6.3787e-05, 8.8678e-05, 1.0751e-04, 8.9148e-05, 8.1123e-05, 7.2801e-05, + 6.7568e-05, 8.3805e-05], device='cuda:4') +2023-03-26 00:36:20,607 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 00:36:41,890 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10515.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 00:36:42,340 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.257e+02 1.733e+02 2.215e+02 2.656e+02 7.843e+02, threshold=4.429e+02, percent-clipped=2.0 +2023-03-26 00:36:51,176 INFO [finetune.py:976] (4/7) Epoch 2, batch 4800, loss[loss=0.2277, simple_loss=0.2872, pruned_loss=0.08413, over 4832.00 frames. ], tot_loss[loss=0.2645, simple_loss=0.3099, pruned_loss=0.1095, over 953339.82 frames. ], batch size: 30, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:37:15,260 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1828, 3.5516, 3.7188, 4.0486, 3.9075, 3.7181, 4.2727, 1.3441], + device='cuda:4'), covar=tensor([0.0743, 0.0810, 0.0730, 0.0913, 0.1248, 0.1317, 0.0655, 0.5286], + device='cuda:4'), in_proj_covar=tensor([0.0372, 0.0247, 0.0277, 0.0299, 0.0344, 0.0290, 0.0313, 0.0303], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:37:24,886 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4289, 1.2313, 1.1879, 1.3899, 1.5791, 1.3906, 0.8347, 1.1797], + device='cuda:4'), covar=tensor([0.3007, 0.3043, 0.2559, 0.2308, 0.2189, 0.1482, 0.3415, 0.2334], + device='cuda:4'), in_proj_covar=tensor([0.0224, 0.0205, 0.0192, 0.0178, 0.0227, 0.0170, 0.0208, 0.0181], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:37:43,276 INFO [finetune.py:976] (4/7) Epoch 2, batch 4850, loss[loss=0.2574, simple_loss=0.3123, pruned_loss=0.1012, over 4927.00 frames. ], tot_loss[loss=0.2672, simple_loss=0.3136, pruned_loss=0.1104, over 953154.00 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:37:56,925 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0751, 1.9832, 1.7712, 1.5416, 2.2944, 2.6386, 2.0722, 1.8536], + device='cuda:4'), covar=tensor([0.0332, 0.0335, 0.0541, 0.0410, 0.0355, 0.0291, 0.0346, 0.0381], + device='cuda:4'), in_proj_covar=tensor([0.0081, 0.0111, 0.0132, 0.0111, 0.0101, 0.0097, 0.0087, 0.0106], + device='cuda:4'), out_proj_covar=tensor([6.3109e-05, 8.7534e-05, 1.0636e-04, 8.7974e-05, 8.0310e-05, 7.2461e-05, + 6.6893e-05, 8.2847e-05], device='cuda:4') +2023-03-26 00:38:22,412 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.244e+02 1.945e+02 2.280e+02 2.839e+02 5.226e+02, threshold=4.560e+02, percent-clipped=1.0 +2023-03-26 00:38:33,264 INFO [finetune.py:976] (4/7) Epoch 2, batch 4900, loss[loss=0.2797, simple_loss=0.3283, pruned_loss=0.1155, over 4815.00 frames. ], tot_loss[loss=0.2679, simple_loss=0.3143, pruned_loss=0.1107, over 953250.02 frames. ], batch size: 40, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:38:40,470 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0010, 1.7837, 1.7099, 1.6778, 2.0962, 2.2315, 1.8938, 1.5758], + device='cuda:4'), covar=tensor([0.0233, 0.0355, 0.0546, 0.0338, 0.0218, 0.0299, 0.0287, 0.0407], + device='cuda:4'), in_proj_covar=tensor([0.0081, 0.0111, 0.0132, 0.0111, 0.0101, 0.0097, 0.0087, 0.0106], + device='cuda:4'), out_proj_covar=tensor([6.3206e-05, 8.7866e-05, 1.0667e-04, 8.8175e-05, 8.0588e-05, 7.2436e-05, + 6.7018e-05, 8.3114e-05], device='cuda:4') +2023-03-26 00:39:05,543 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 00:39:15,448 INFO [finetune.py:976] (4/7) Epoch 2, batch 4950, loss[loss=0.288, simple_loss=0.3328, pruned_loss=0.1216, over 4889.00 frames. ], tot_loss[loss=0.2678, simple_loss=0.3151, pruned_loss=0.1103, over 953859.52 frames. ], batch size: 35, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:39:27,276 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10695.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:39:40,468 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.382e+02 1.832e+02 2.144e+02 2.563e+02 4.788e+02, threshold=4.289e+02, percent-clipped=1.0 +2023-03-26 00:39:42,230 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10718.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 00:39:48,192 INFO [finetune.py:976] (4/7) Epoch 2, batch 5000, loss[loss=0.219, simple_loss=0.2682, pruned_loss=0.0849, over 4877.00 frames. ], tot_loss[loss=0.2649, simple_loss=0.3124, pruned_loss=0.1087, over 953857.75 frames. ], batch size: 31, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:39:59,862 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10743.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:40:00,534 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10744.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:40:01,215 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 00:40:13,903 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10765.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:40:14,499 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=10766.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 00:40:19,852 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10774.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:40:25,991 INFO [finetune.py:976] (4/7) Epoch 2, batch 5050, loss[loss=0.2667, simple_loss=0.3092, pruned_loss=0.1121, over 4855.00 frames. ], tot_loss[loss=0.2638, simple_loss=0.3107, pruned_loss=0.1085, over 955389.78 frames. ], batch size: 44, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:40:48,588 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10805.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:40:52,130 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10810.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 00:40:53,402 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5631, 1.3032, 1.2098, 0.9463, 1.2988, 1.2834, 1.2301, 1.9641], + device='cuda:4'), covar=tensor([1.6698, 1.3600, 1.1773, 1.5126, 1.1769, 0.8658, 1.4039, 0.4883], + device='cuda:4'), in_proj_covar=tensor([0.0259, 0.0238, 0.0214, 0.0274, 0.0230, 0.0192, 0.0234, 0.0177], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 00:40:55,203 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5199, 2.1816, 1.9139, 0.9409, 2.0299, 1.9303, 1.6866, 1.9534], + device='cuda:4'), covar=tensor([0.0875, 0.1033, 0.2008, 0.2728, 0.1738, 0.2347, 0.2323, 0.1265], + device='cuda:4'), in_proj_covar=tensor([0.0164, 0.0193, 0.0202, 0.0186, 0.0212, 0.0207, 0.0214, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:40:55,640 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.711e+02 2.025e+02 2.497e+02 4.623e+02, threshold=4.049e+02, percent-clipped=1.0 +2023-03-26 00:41:03,400 INFO [finetune.py:976] (4/7) Epoch 2, batch 5100, loss[loss=0.2357, simple_loss=0.2854, pruned_loss=0.09301, over 4776.00 frames. ], tot_loss[loss=0.2593, simple_loss=0.3068, pruned_loss=0.1059, over 955305.33 frames. ], batch size: 26, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:41:07,762 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10835.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:41:47,384 INFO [finetune.py:976] (4/7) Epoch 2, batch 5150, loss[loss=0.2479, simple_loss=0.2922, pruned_loss=0.1018, over 4693.00 frames. ], tot_loss[loss=0.2587, simple_loss=0.3062, pruned_loss=0.1056, over 955947.58 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:41:54,269 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10885.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 00:42:18,067 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4635, 1.3160, 1.3891, 1.3961, 0.6573, 2.3122, 0.7457, 1.2893], + device='cuda:4'), covar=tensor([0.3616, 0.2490, 0.2156, 0.2442, 0.2395, 0.0409, 0.2706, 0.1545], + device='cuda:4'), in_proj_covar=tensor([0.0128, 0.0109, 0.0114, 0.0117, 0.0114, 0.0096, 0.0099, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 00:42:25,067 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.204e+02 1.747e+02 2.089e+02 2.521e+02 5.475e+02, threshold=4.178e+02, percent-clipped=1.0 +2023-03-26 00:42:37,754 INFO [finetune.py:976] (4/7) Epoch 2, batch 5200, loss[loss=0.2527, simple_loss=0.319, pruned_loss=0.09321, over 4814.00 frames. ], tot_loss[loss=0.2628, simple_loss=0.3107, pruned_loss=0.1075, over 955040.16 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:42:57,278 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0636, 0.9671, 1.0638, 0.2778, 0.7980, 1.1421, 1.2277, 1.0833], + device='cuda:4'), covar=tensor([0.0996, 0.0580, 0.0440, 0.0737, 0.0490, 0.0544, 0.0365, 0.0596], + device='cuda:4'), in_proj_covar=tensor([0.0128, 0.0155, 0.0117, 0.0134, 0.0130, 0.0117, 0.0145, 0.0142], + device='cuda:4'), out_proj_covar=tensor([9.6618e-05, 1.1535e-04, 8.5296e-05, 9.8867e-05, 9.4545e-05, 8.6506e-05, + 1.0824e-04, 1.0556e-04], device='cuda:4') +2023-03-26 00:43:00,770 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10946.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 00:43:29,732 INFO [finetune.py:976] (4/7) Epoch 2, batch 5250, loss[loss=0.2074, simple_loss=0.2598, pruned_loss=0.07748, over 4793.00 frames. ], tot_loss[loss=0.2621, simple_loss=0.3101, pruned_loss=0.107, over 954724.16 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:43:31,747 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=2.18 vs. limit=2.0 +2023-03-26 00:43:45,327 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0577, 1.7626, 1.7570, 0.9716, 1.9114, 2.1468, 1.8297, 1.7951], + device='cuda:4'), covar=tensor([0.0782, 0.0628, 0.0613, 0.0784, 0.0475, 0.0418, 0.0451, 0.0492], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0155, 0.0117, 0.0135, 0.0131, 0.0118, 0.0145, 0.0143], + device='cuda:4'), out_proj_covar=tensor([9.6925e-05, 1.1553e-04, 8.5569e-05, 9.8943e-05, 9.4944e-05, 8.6698e-05, + 1.0841e-04, 1.0586e-04], device='cuda:4') +2023-03-26 00:43:54,653 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 00:44:03,160 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.206e+02 2.014e+02 2.376e+02 2.957e+02 8.531e+02, threshold=4.753e+02, percent-clipped=3.0 +2023-03-26 00:44:10,959 INFO [finetune.py:976] (4/7) Epoch 2, batch 5300, loss[loss=0.273, simple_loss=0.3285, pruned_loss=0.1088, over 4820.00 frames. ], tot_loss[loss=0.2636, simple_loss=0.3117, pruned_loss=0.1078, over 953008.25 frames. ], batch size: 39, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:44:44,465 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11065.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:44:46,281 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1137, 3.5562, 3.7327, 3.9973, 3.8325, 3.5837, 4.1864, 1.3940], + device='cuda:4'), covar=tensor([0.0695, 0.0743, 0.0781, 0.0842, 0.1178, 0.1454, 0.0713, 0.4782], + device='cuda:4'), in_proj_covar=tensor([0.0370, 0.0246, 0.0276, 0.0298, 0.0344, 0.0290, 0.0312, 0.0301], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:44:58,486 INFO [finetune.py:976] (4/7) Epoch 2, batch 5350, loss[loss=0.222, simple_loss=0.2781, pruned_loss=0.08293, over 4789.00 frames. ], tot_loss[loss=0.2641, simple_loss=0.3127, pruned_loss=0.1078, over 953282.75 frames. ], batch size: 51, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:45:16,021 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11100.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:45:21,392 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 00:45:23,530 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11110.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 00:45:25,461 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11113.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:45:26,733 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.4733, 1.4187, 1.4821, 1.0088, 1.4071, 1.7734, 1.7175, 1.4588], + device='cuda:4'), covar=tensor([0.1048, 0.0661, 0.0530, 0.0685, 0.0410, 0.0438, 0.0299, 0.0583], + device='cuda:4'), in_proj_covar=tensor([0.0128, 0.0155, 0.0117, 0.0134, 0.0130, 0.0117, 0.0145, 0.0142], + device='cuda:4'), out_proj_covar=tensor([9.6767e-05, 1.1531e-04, 8.5677e-05, 9.8603e-05, 9.4617e-05, 8.6583e-05, + 1.0796e-04, 1.0539e-04], device='cuda:4') +2023-03-26 00:45:27,191 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.217e+02 1.830e+02 2.178e+02 2.684e+02 7.389e+02, threshold=4.357e+02, percent-clipped=1.0 +2023-03-26 00:45:27,495 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.25 vs. limit=5.0 +2023-03-26 00:45:39,915 INFO [finetune.py:976] (4/7) Epoch 2, batch 5400, loss[loss=0.2404, simple_loss=0.2917, pruned_loss=0.09457, over 4900.00 frames. ], tot_loss[loss=0.2621, simple_loss=0.3101, pruned_loss=0.1071, over 954352.99 frames. ], batch size: 35, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:45:41,699 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11130.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:46:05,249 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11158.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:46:22,021 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11174.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:46:24,382 INFO [finetune.py:976] (4/7) Epoch 2, batch 5450, loss[loss=0.2722, simple_loss=0.3239, pruned_loss=0.1103, over 4940.00 frames. ], tot_loss[loss=0.2581, simple_loss=0.306, pruned_loss=0.1051, over 955238.94 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:46:55,234 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.192e+02 1.719e+02 1.995e+02 2.396e+02 4.116e+02, threshold=3.991e+02, percent-clipped=0.0 +2023-03-26 00:47:00,293 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6557, 1.0177, 1.3799, 1.3585, 1.2313, 1.2869, 1.2763, 1.3651], + device='cuda:4'), covar=tensor([1.0650, 1.8651, 1.5964, 1.5407, 1.8313, 1.2263, 2.1589, 1.2793], + device='cuda:4'), in_proj_covar=tensor([0.0228, 0.0259, 0.0252, 0.0271, 0.0247, 0.0220, 0.0281, 0.0219], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 00:47:11,562 INFO [finetune.py:976] (4/7) Epoch 2, batch 5500, loss[loss=0.239, simple_loss=0.2909, pruned_loss=0.09356, over 4768.00 frames. ], tot_loss[loss=0.2551, simple_loss=0.3029, pruned_loss=0.1036, over 955549.16 frames. ], batch size: 28, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:47:21,374 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11235.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:47:30,524 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11241.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 00:47:35,449 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2060, 2.7951, 2.0201, 1.6314, 2.8667, 2.7541, 2.4101, 2.3544], + device='cuda:4'), covar=tensor([0.0948, 0.0624, 0.1142, 0.1279, 0.0557, 0.0857, 0.1000, 0.1014], + device='cuda:4'), in_proj_covar=tensor([0.0139, 0.0133, 0.0144, 0.0130, 0.0110, 0.0141, 0.0147, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 00:48:00,496 INFO [finetune.py:976] (4/7) Epoch 2, batch 5550, loss[loss=0.1956, simple_loss=0.2415, pruned_loss=0.0748, over 4735.00 frames. ], tot_loss[loss=0.2581, simple_loss=0.3057, pruned_loss=0.1053, over 954796.48 frames. ], batch size: 23, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:48:10,882 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6435, 1.5464, 2.1278, 3.4204, 2.4397, 2.3122, 1.1643, 2.5883], + device='cuda:4'), covar=tensor([0.2040, 0.1658, 0.1453, 0.0576, 0.0942, 0.1630, 0.2033, 0.0752], + device='cuda:4'), in_proj_covar=tensor([0.0105, 0.0122, 0.0141, 0.0167, 0.0107, 0.0148, 0.0132, 0.0110], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 00:48:30,779 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.394e+02 1.988e+02 2.263e+02 2.636e+02 5.646e+02, threshold=4.525e+02, percent-clipped=4.0 +2023-03-26 00:48:38,278 INFO [finetune.py:976] (4/7) Epoch 2, batch 5600, loss[loss=0.2966, simple_loss=0.3404, pruned_loss=0.1264, over 4931.00 frames. ], tot_loss[loss=0.263, simple_loss=0.3111, pruned_loss=0.1074, over 956131.17 frames. ], batch size: 42, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:49:19,623 INFO [finetune.py:976] (4/7) Epoch 2, batch 5650, loss[loss=0.2516, simple_loss=0.3093, pruned_loss=0.09697, over 4381.00 frames. ], tot_loss[loss=0.2649, simple_loss=0.3138, pruned_loss=0.108, over 955361.22 frames. ], batch size: 65, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:49:44,844 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11400.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:49:54,702 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.185e+02 1.700e+02 2.043e+02 2.333e+02 3.537e+02, threshold=4.085e+02, percent-clipped=0.0 +2023-03-26 00:50:01,883 INFO [finetune.py:976] (4/7) Epoch 2, batch 5700, loss[loss=0.2399, simple_loss=0.2755, pruned_loss=0.1021, over 4212.00 frames. ], tot_loss[loss=0.2607, simple_loss=0.308, pruned_loss=0.1067, over 936391.00 frames. ], batch size: 18, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:50:03,178 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4301, 1.5732, 1.4413, 1.7114, 1.6332, 2.9880, 1.4254, 1.7004], + device='cuda:4'), covar=tensor([0.1066, 0.1684, 0.1074, 0.0970, 0.1603, 0.0315, 0.1376, 0.1632], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0081, 0.0078, 0.0080, 0.0093, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 00:50:03,187 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11430.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:50:13,812 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11448.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:50:33,841 INFO [finetune.py:976] (4/7) Epoch 3, batch 0, loss[loss=0.3477, simple_loss=0.3817, pruned_loss=0.1568, over 4750.00 frames. ], tot_loss[loss=0.3477, simple_loss=0.3817, pruned_loss=0.1568, over 4750.00 frames. ], batch size: 54, lr: 3.99e-03, grad_scale: 16.0 +2023-03-26 00:50:33,841 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 00:50:42,516 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6248, 0.7931, 1.4681, 1.3457, 1.3166, 1.2735, 1.1596, 1.3706], + device='cuda:4'), covar=tensor([1.2360, 2.1492, 1.7036, 1.8488, 2.0748, 1.4346, 2.3934, 1.5203], + device='cuda:4'), in_proj_covar=tensor([0.0227, 0.0257, 0.0250, 0.0269, 0.0244, 0.0218, 0.0279, 0.0217], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 00:50:46,684 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6005, 1.4964, 1.8960, 2.8489, 2.1809, 2.1146, 0.9967, 2.3145], + device='cuda:4'), covar=tensor([0.1842, 0.1653, 0.1357, 0.0612, 0.0784, 0.1297, 0.1925, 0.0705], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0120, 0.0139, 0.0164, 0.0105, 0.0146, 0.0130, 0.0108], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 00:50:55,329 INFO [finetune.py:1010] (4/7) Epoch 3, validation: loss=0.1864, simple_loss=0.2566, pruned_loss=0.05807, over 2265189.00 frames. +2023-03-26 00:50:55,329 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6301MB +2023-03-26 00:51:20,330 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11478.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:51:29,402 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3547, 1.6684, 1.6665, 1.7847, 1.6127, 3.4584, 1.3143, 1.6264], + device='cuda:4'), covar=tensor([0.1163, 0.1616, 0.1283, 0.1127, 0.1717, 0.0246, 0.1566, 0.1764], + device='cuda:4'), in_proj_covar=tensor([0.0077, 0.0080, 0.0078, 0.0079, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 00:51:38,815 INFO [finetune.py:976] (4/7) Epoch 3, batch 50, loss[loss=0.2851, simple_loss=0.3182, pruned_loss=0.126, over 4912.00 frames. ], tot_loss[loss=0.2657, simple_loss=0.3136, pruned_loss=0.1089, over 215513.72 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:51:45,832 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.161e+02 1.779e+02 2.075e+02 2.495e+02 4.593e+02, threshold=4.151e+02, percent-clipped=1.0 +2023-03-26 00:51:54,921 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11530.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:51:56,200 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11532.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:52:01,995 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11541.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 00:52:03,919 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-26 00:52:12,048 INFO [finetune.py:976] (4/7) Epoch 3, batch 100, loss[loss=0.2572, simple_loss=0.3074, pruned_loss=0.1036, over 4900.00 frames. ], tot_loss[loss=0.2592, simple_loss=0.3062, pruned_loss=0.1061, over 378130.61 frames. ], batch size: 36, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:52:25,136 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.66 vs. limit=5.0 +2023-03-26 00:52:33,516 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11589.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 00:52:36,422 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11593.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:52:39,393 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1879, 1.4819, 0.6757, 1.8207, 2.4067, 1.7259, 1.7338, 2.1180], + device='cuda:4'), covar=tensor([0.2006, 0.2858, 0.3318, 0.1689, 0.2538, 0.2801, 0.1933, 0.2614], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0099, 0.0118, 0.0095, 0.0125, 0.0098, 0.0100, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 00:52:47,409 INFO [finetune.py:976] (4/7) Epoch 3, batch 150, loss[loss=0.2067, simple_loss=0.2666, pruned_loss=0.07336, over 4871.00 frames. ], tot_loss[loss=0.2518, simple_loss=0.2995, pruned_loss=0.102, over 506572.44 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:53:00,294 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.308e+02 1.853e+02 2.230e+02 2.582e+02 4.758e+02, threshold=4.459e+02, percent-clipped=3.0 +2023-03-26 00:53:24,154 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11635.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:53:48,108 INFO [finetune.py:976] (4/7) Epoch 3, batch 200, loss[loss=0.3193, simple_loss=0.3593, pruned_loss=0.1397, over 4703.00 frames. ], tot_loss[loss=0.2504, simple_loss=0.2983, pruned_loss=0.1012, over 608797.96 frames. ], batch size: 59, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:54:15,073 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-03-26 00:54:23,965 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=2.02 vs. limit=2.0 +2023-03-26 00:54:35,492 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11696.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:54:37,337 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11699.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:54:44,849 INFO [finetune.py:976] (4/7) Epoch 3, batch 250, loss[loss=0.2786, simple_loss=0.3304, pruned_loss=0.1134, over 4840.00 frames. ], tot_loss[loss=0.2531, simple_loss=0.3015, pruned_loss=0.1024, over 687651.09 frames. ], batch size: 49, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:55:02,878 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.255e+02 1.783e+02 2.124e+02 2.629e+02 6.988e+02, threshold=4.248e+02, percent-clipped=1.0 +2023-03-26 00:55:32,488 INFO [finetune.py:976] (4/7) Epoch 3, batch 300, loss[loss=0.2512, simple_loss=0.3103, pruned_loss=0.09603, over 4930.00 frames. ], tot_loss[loss=0.2598, simple_loss=0.3085, pruned_loss=0.1056, over 749353.06 frames. ], batch size: 42, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:55:40,986 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11760.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:56:13,665 INFO [finetune.py:976] (4/7) Epoch 3, batch 350, loss[loss=0.3536, simple_loss=0.3938, pruned_loss=0.1567, over 4808.00 frames. ], tot_loss[loss=0.2626, simple_loss=0.3114, pruned_loss=0.107, over 795763.12 frames. ], batch size: 39, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:56:20,323 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.891e+02 2.265e+02 2.576e+02 3.939e+02, threshold=4.529e+02, percent-clipped=0.0 +2023-03-26 00:56:20,466 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11816.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:56:30,466 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11830.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:56:45,042 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11846.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:56:51,647 INFO [finetune.py:976] (4/7) Epoch 3, batch 400, loss[loss=0.2843, simple_loss=0.3258, pruned_loss=0.1213, over 4889.00 frames. ], tot_loss[loss=0.2633, simple_loss=0.3121, pruned_loss=0.1072, over 831910.24 frames. ], batch size: 32, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:57:20,622 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11877.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:57:21,182 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=11878.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:57:33,128 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11888.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:57:55,603 INFO [finetune.py:976] (4/7) Epoch 3, batch 450, loss[loss=0.21, simple_loss=0.2604, pruned_loss=0.07979, over 4800.00 frames. ], tot_loss[loss=0.261, simple_loss=0.3104, pruned_loss=0.1058, over 860989.58 frames. ], batch size: 25, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:58:01,507 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11907.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:58:12,548 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.239e+02 1.814e+02 2.317e+02 2.793e+02 4.030e+02, threshold=4.633e+02, percent-clipped=0.0 +2023-03-26 00:58:15,059 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11920.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:58:39,269 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11948.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:58:50,440 INFO [finetune.py:976] (4/7) Epoch 3, batch 500, loss[loss=0.2368, simple_loss=0.2833, pruned_loss=0.09513, over 4933.00 frames. ], tot_loss[loss=0.2587, simple_loss=0.3073, pruned_loss=0.105, over 882167.03 frames. ], batch size: 38, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:58:58,245 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8600, 1.6641, 1.3121, 1.6774, 1.5734, 1.4714, 1.4689, 2.5083], + device='cuda:4'), covar=tensor([1.4062, 1.3598, 1.1536, 1.4815, 1.2024, 0.7766, 1.4463, 0.4093], + device='cuda:4'), in_proj_covar=tensor([0.0262, 0.0241, 0.0216, 0.0277, 0.0231, 0.0193, 0.0235, 0.0179], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 00:59:19,996 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11981.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:59:27,105 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11984.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:59:31,810 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11991.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:59:46,860 INFO [finetune.py:976] (4/7) Epoch 3, batch 550, loss[loss=0.2511, simple_loss=0.2977, pruned_loss=0.1022, over 4827.00 frames. ], tot_loss[loss=0.2547, simple_loss=0.3029, pruned_loss=0.1032, over 898828.56 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 00:59:49,293 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12009.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 00:59:53,393 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.776e+02 2.066e+02 2.700e+02 4.009e+02, threshold=4.133e+02, percent-clipped=0.0 +2023-03-26 01:00:02,292 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5963, 1.6022, 1.8663, 2.0124, 1.7500, 3.1255, 1.4375, 1.7767], + device='cuda:4'), covar=tensor([0.0927, 0.1359, 0.1246, 0.0838, 0.1274, 0.0277, 0.1179, 0.1344], + device='cuda:4'), in_proj_covar=tensor([0.0077, 0.0080, 0.0077, 0.0079, 0.0092, 0.0082, 0.0084, 0.0078], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 01:00:10,588 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7858, 1.7286, 1.6559, 1.8741, 1.3069, 4.4054, 1.6460, 2.2483], + device='cuda:4'), covar=tensor([0.3398, 0.2381, 0.2039, 0.2229, 0.1946, 0.0082, 0.2749, 0.1432], + device='cuda:4'), in_proj_covar=tensor([0.0127, 0.0109, 0.0114, 0.0117, 0.0114, 0.0096, 0.0099, 0.0096], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 01:00:13,577 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12045.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:00:21,763 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12055.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:00:22,312 INFO [finetune.py:976] (4/7) Epoch 3, batch 600, loss[loss=0.2266, simple_loss=0.2897, pruned_loss=0.08176, over 4799.00 frames. ], tot_loss[loss=0.2541, simple_loss=0.3024, pruned_loss=0.1029, over 912059.61 frames. ], batch size: 51, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 01:00:41,419 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=12076.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:00:47,398 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0266, 0.9612, 1.0457, 0.3592, 0.7087, 1.1510, 1.2159, 1.1008], + device='cuda:4'), covar=tensor([0.1385, 0.0674, 0.0601, 0.0897, 0.0672, 0.0639, 0.0492, 0.0809], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0156, 0.0117, 0.0135, 0.0132, 0.0118, 0.0146, 0.0143], + device='cuda:4'), out_proj_covar=tensor([9.7573e-05, 1.1579e-04, 8.5631e-05, 9.9222e-05, 9.5680e-05, 8.7229e-05, + 1.0896e-04, 1.0614e-04], device='cuda:4') +2023-03-26 01:00:48,054 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 01:01:08,113 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-26 01:01:09,713 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-26 01:01:16,813 INFO [finetune.py:976] (4/7) Epoch 3, batch 650, loss[loss=0.2941, simple_loss=0.3367, pruned_loss=0.1258, over 4931.00 frames. ], tot_loss[loss=0.2578, simple_loss=0.3068, pruned_loss=0.1044, over 921571.78 frames. ], batch size: 33, lr: 3.99e-03, grad_scale: 32.0 +2023-03-26 01:01:23,450 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.190e+02 1.901e+02 2.250e+02 2.651e+02 5.885e+02, threshold=4.501e+02, percent-clipped=2.0 +2023-03-26 01:01:50,153 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12137.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:02:02,047 INFO [finetune.py:976] (4/7) Epoch 3, batch 700, loss[loss=0.2799, simple_loss=0.312, pruned_loss=0.1239, over 4160.00 frames. ], tot_loss[loss=0.2588, simple_loss=0.3076, pruned_loss=0.105, over 926349.08 frames. ], batch size: 65, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:02:12,358 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12172.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:02:24,516 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12188.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:02:34,484 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7867, 1.5660, 2.0957, 3.4405, 2.5243, 2.4928, 1.1632, 2.6892], + device='cuda:4'), covar=tensor([0.1804, 0.1606, 0.1484, 0.0537, 0.0784, 0.1468, 0.1909, 0.0692], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0119, 0.0138, 0.0163, 0.0105, 0.0145, 0.0130, 0.0107], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 01:02:39,280 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12202.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:02:43,053 INFO [finetune.py:976] (4/7) Epoch 3, batch 750, loss[loss=0.2785, simple_loss=0.3386, pruned_loss=0.1092, over 4796.00 frames. ], tot_loss[loss=0.2599, simple_loss=0.3093, pruned_loss=0.1053, over 933822.31 frames. ], batch size: 45, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:02:54,946 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.207e+02 1.858e+02 2.308e+02 2.738e+02 5.308e+02, threshold=4.616e+02, percent-clipped=1.0 +2023-03-26 01:03:12,811 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.93 vs. limit=5.0 +2023-03-26 01:03:14,109 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12236.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:03:20,996 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.75 vs. limit=5.0 +2023-03-26 01:03:30,124 INFO [finetune.py:976] (4/7) Epoch 3, batch 800, loss[loss=0.2194, simple_loss=0.2723, pruned_loss=0.08322, over 4749.00 frames. ], tot_loss[loss=0.2585, simple_loss=0.3082, pruned_loss=0.1044, over 936496.74 frames. ], batch size: 59, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:03:40,508 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5150, 1.4745, 1.4716, 1.4948, 1.0394, 2.9555, 1.0919, 1.5330], + device='cuda:4'), covar=tensor([0.3539, 0.2418, 0.2145, 0.2380, 0.2157, 0.0247, 0.2848, 0.1518], + device='cuda:4'), in_proj_covar=tensor([0.0128, 0.0110, 0.0114, 0.0118, 0.0114, 0.0096, 0.0099, 0.0096], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 01:03:42,862 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12276.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:04:03,350 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12291.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:04:11,112 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12304.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:04:12,239 INFO [finetune.py:976] (4/7) Epoch 3, batch 850, loss[loss=0.2828, simple_loss=0.3181, pruned_loss=0.1238, over 4752.00 frames. ], tot_loss[loss=0.2565, simple_loss=0.3058, pruned_loss=0.1036, over 940468.51 frames. ], batch size: 59, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:04:19,496 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.224e+02 1.828e+02 2.109e+02 2.576e+02 5.946e+02, threshold=4.217e+02, percent-clipped=1.0 +2023-03-26 01:04:46,215 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12339.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:04:46,839 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12340.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:05:06,399 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12355.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:05:06,919 INFO [finetune.py:976] (4/7) Epoch 3, batch 900, loss[loss=0.2411, simple_loss=0.2924, pruned_loss=0.09488, over 4663.00 frames. ], tot_loss[loss=0.2531, simple_loss=0.3024, pruned_loss=0.1019, over 944656.70 frames. ], batch size: 23, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:05:31,954 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.2971, 4.5279, 4.7322, 5.0927, 4.9077, 4.6881, 5.4208, 1.6980], + device='cuda:4'), covar=tensor([0.0784, 0.0703, 0.0726, 0.0954, 0.1473, 0.1311, 0.0491, 0.5465], + device='cuda:4'), in_proj_covar=tensor([0.0368, 0.0245, 0.0274, 0.0298, 0.0342, 0.0286, 0.0311, 0.0301], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:05:43,765 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12403.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:05:45,596 INFO [finetune.py:976] (4/7) Epoch 3, batch 950, loss[loss=0.2404, simple_loss=0.2944, pruned_loss=0.09317, over 4779.00 frames. ], tot_loss[loss=0.2522, simple_loss=0.3011, pruned_loss=0.1016, over 947056.84 frames. ], batch size: 29, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:05:57,751 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.215e+02 1.779e+02 2.123e+02 2.532e+02 4.452e+02, threshold=4.246e+02, percent-clipped=1.0 +2023-03-26 01:06:12,420 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12432.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:06:22,003 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=12446.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:06:28,915 INFO [finetune.py:976] (4/7) Epoch 3, batch 1000, loss[loss=0.3252, simple_loss=0.3693, pruned_loss=0.1406, over 4818.00 frames. ], tot_loss[loss=0.2549, simple_loss=0.3043, pruned_loss=0.1028, over 950592.63 frames. ], batch size: 40, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:06:32,229 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-03-26 01:06:39,167 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12472.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:06:51,371 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7092, 1.5827, 1.2208, 1.5131, 1.4431, 1.3641, 1.3732, 2.2449], + device='cuda:4'), covar=tensor([1.4101, 1.3526, 1.0549, 1.4954, 1.1758, 0.7696, 1.4768, 0.4220], + device='cuda:4'), in_proj_covar=tensor([0.0263, 0.0241, 0.0215, 0.0276, 0.0230, 0.0192, 0.0234, 0.0178], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 01:07:06,797 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4153, 0.5567, 1.2556, 1.1448, 1.1280, 1.1039, 0.9679, 1.2034], + device='cuda:4'), covar=tensor([0.9468, 1.6623, 1.3411, 1.3974, 1.4896, 1.0705, 1.7938, 1.1434], + device='cuda:4'), in_proj_covar=tensor([0.0228, 0.0257, 0.0253, 0.0270, 0.0246, 0.0219, 0.0280, 0.0219], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 01:07:17,559 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12502.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:07:19,894 INFO [finetune.py:976] (4/7) Epoch 3, batch 1050, loss[loss=0.295, simple_loss=0.3438, pruned_loss=0.1231, over 4917.00 frames. ], tot_loss[loss=0.2577, simple_loss=0.3078, pruned_loss=0.1038, over 951010.71 frames. ], batch size: 38, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:07:20,619 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12507.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:07:31,082 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.261e+02 2.008e+02 2.380e+02 2.733e+02 7.204e+02, threshold=4.759e+02, percent-clipped=3.0 +2023-03-26 01:07:38,121 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12520.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:07:38,856 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 01:08:10,400 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12550.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:08:17,959 INFO [finetune.py:976] (4/7) Epoch 3, batch 1100, loss[loss=0.2655, simple_loss=0.3187, pruned_loss=0.1062, over 4898.00 frames. ], tot_loss[loss=0.2588, simple_loss=0.3087, pruned_loss=0.1045, over 951817.63 frames. ], batch size: 37, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:08:35,452 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12576.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:08:58,294 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12604.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:08:59,395 INFO [finetune.py:976] (4/7) Epoch 3, batch 1150, loss[loss=0.2874, simple_loss=0.3315, pruned_loss=0.1216, over 4844.00 frames. ], tot_loss[loss=0.2582, simple_loss=0.3086, pruned_loss=0.1039, over 953558.52 frames. ], batch size: 44, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:09:11,543 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.144e+02 1.712e+02 1.953e+02 2.432e+02 5.551e+02, threshold=3.906e+02, percent-clipped=1.0 +2023-03-26 01:09:18,675 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6635, 1.2407, 1.3408, 1.4142, 1.7416, 1.8105, 1.5432, 1.2598], + device='cuda:4'), covar=tensor([0.0238, 0.0380, 0.0553, 0.0322, 0.0243, 0.0311, 0.0314, 0.0415], + device='cuda:4'), in_proj_covar=tensor([0.0083, 0.0113, 0.0136, 0.0116, 0.0104, 0.0098, 0.0089, 0.0109], + device='cuda:4'), out_proj_covar=tensor([6.4788e-05, 8.9466e-05, 1.0980e-04, 9.1638e-05, 8.2694e-05, 7.3238e-05, + 6.8267e-05, 8.5235e-05], device='cuda:4') +2023-03-26 01:09:21,030 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12624.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:09:42,256 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12640.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:09:55,582 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12652.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:10:03,871 INFO [finetune.py:976] (4/7) Epoch 3, batch 1200, loss[loss=0.2396, simple_loss=0.2952, pruned_loss=0.09206, over 4716.00 frames. ], tot_loss[loss=0.256, simple_loss=0.3068, pruned_loss=0.1026, over 954976.94 frames. ], batch size: 23, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:10:31,613 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12688.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:10:43,571 INFO [finetune.py:976] (4/7) Epoch 3, batch 1250, loss[loss=0.2606, simple_loss=0.3039, pruned_loss=0.1087, over 4929.00 frames. ], tot_loss[loss=0.254, simple_loss=0.3041, pruned_loss=0.102, over 954725.25 frames. ], batch size: 38, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:10:46,734 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0104, 1.7866, 1.4479, 1.8067, 1.6970, 1.6773, 1.6400, 2.5814], + device='cuda:4'), covar=tensor([1.2422, 1.2832, 1.0129, 1.3140, 1.0414, 0.7075, 1.2434, 0.3836], + device='cuda:4'), in_proj_covar=tensor([0.0265, 0.0242, 0.0216, 0.0278, 0.0231, 0.0193, 0.0235, 0.0180], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 01:10:51,325 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.791e+02 2.152e+02 2.550e+02 3.946e+02, threshold=4.304e+02, percent-clipped=1.0 +2023-03-26 01:11:04,166 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1235, 0.9807, 1.0177, 0.4041, 0.7413, 1.1605, 1.2555, 1.0738], + device='cuda:4'), covar=tensor([0.0987, 0.0546, 0.0430, 0.0642, 0.0486, 0.0462, 0.0341, 0.0545], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0155, 0.0116, 0.0134, 0.0131, 0.0118, 0.0145, 0.0142], + device='cuda:4'), out_proj_covar=tensor([9.7039e-05, 1.1552e-04, 8.5187e-05, 9.8084e-05, 9.4892e-05, 8.7152e-05, + 1.0859e-04, 1.0569e-04], device='cuda:4') +2023-03-26 01:11:06,471 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12732.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:11:13,783 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1698, 1.2738, 1.3632, 0.7176, 1.0647, 1.4875, 1.5553, 1.3292], + device='cuda:4'), covar=tensor([0.0871, 0.0483, 0.0414, 0.0538, 0.0388, 0.0440, 0.0261, 0.0460], + device='cuda:4'), in_proj_covar=tensor([0.0128, 0.0154, 0.0116, 0.0133, 0.0130, 0.0118, 0.0145, 0.0142], + device='cuda:4'), out_proj_covar=tensor([9.6600e-05, 1.1498e-04, 8.4865e-05, 9.7698e-05, 9.4540e-05, 8.6750e-05, + 1.0812e-04, 1.0526e-04], device='cuda:4') +2023-03-26 01:11:25,781 INFO [finetune.py:976] (4/7) Epoch 3, batch 1300, loss[loss=0.2274, simple_loss=0.275, pruned_loss=0.08985, over 4926.00 frames. ], tot_loss[loss=0.2506, simple_loss=0.3005, pruned_loss=0.1003, over 956662.53 frames. ], batch size: 37, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:11:26,266 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 01:11:28,847 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8253, 1.2300, 0.9471, 1.7835, 2.0069, 1.5885, 1.4730, 1.9086], + device='cuda:4'), covar=tensor([0.1319, 0.2014, 0.2149, 0.1077, 0.2093, 0.2074, 0.1306, 0.1563], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0099, 0.0117, 0.0094, 0.0125, 0.0098, 0.0100, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 01:11:31,194 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=12763.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:11:48,952 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=12780.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:12:19,505 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12802.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:12:22,426 INFO [finetune.py:976] (4/7) Epoch 3, batch 1350, loss[loss=0.2408, simple_loss=0.2981, pruned_loss=0.09175, over 4795.00 frames. ], tot_loss[loss=0.2508, simple_loss=0.3008, pruned_loss=0.1004, over 958804.99 frames. ], batch size: 29, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:12:40,681 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.832e+02 2.182e+02 2.674e+02 3.468e+02, threshold=4.364e+02, percent-clipped=0.0 +2023-03-26 01:12:42,325 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-26 01:12:50,691 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12824.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 01:12:53,694 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7756, 4.3492, 4.3282, 2.1875, 4.5038, 3.3262, 0.7536, 3.0691], + device='cuda:4'), covar=tensor([0.2485, 0.1495, 0.1104, 0.2840, 0.0720, 0.0803, 0.4629, 0.1276], + device='cuda:4'), in_proj_covar=tensor([0.0156, 0.0167, 0.0165, 0.0128, 0.0156, 0.0120, 0.0146, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 01:13:21,549 INFO [finetune.py:976] (4/7) Epoch 3, batch 1400, loss[loss=0.2759, simple_loss=0.3236, pruned_loss=0.1141, over 4829.00 frames. ], tot_loss[loss=0.2534, simple_loss=0.3041, pruned_loss=0.1014, over 958420.67 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:13:50,005 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.38 vs. limit=5.0 +2023-03-26 01:14:18,649 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9517, 1.7138, 1.4518, 1.7702, 1.9805, 1.6300, 2.2851, 1.7688], + device='cuda:4'), covar=tensor([0.2289, 0.4507, 0.5360, 0.4793, 0.3444, 0.2498, 0.4094, 0.3109], + device='cuda:4'), in_proj_covar=tensor([0.0164, 0.0198, 0.0239, 0.0255, 0.0220, 0.0184, 0.0209, 0.0189], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:14:19,109 INFO [finetune.py:976] (4/7) Epoch 3, batch 1450, loss[loss=0.2388, simple_loss=0.2949, pruned_loss=0.09135, over 4751.00 frames. ], tot_loss[loss=0.2545, simple_loss=0.306, pruned_loss=0.1015, over 957708.46 frames. ], batch size: 26, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:14:38,006 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.344e+02 1.849e+02 2.229e+02 2.789e+02 4.972e+02, threshold=4.459e+02, percent-clipped=1.0 +2023-03-26 01:15:13,774 INFO [finetune.py:976] (4/7) Epoch 3, batch 1500, loss[loss=0.2168, simple_loss=0.2838, pruned_loss=0.07491, over 4922.00 frames. ], tot_loss[loss=0.2554, simple_loss=0.3069, pruned_loss=0.102, over 957855.65 frames. ], batch size: 42, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:15:39,516 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0258, 2.2226, 1.9339, 1.4268, 2.4292, 2.2757, 2.1971, 1.9791], + device='cuda:4'), covar=tensor([0.0930, 0.0709, 0.1081, 0.1299, 0.0450, 0.0943, 0.0837, 0.1097], + device='cuda:4'), in_proj_covar=tensor([0.0142, 0.0135, 0.0146, 0.0132, 0.0112, 0.0145, 0.0151, 0.0166], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:15:49,885 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3914, 1.3528, 1.4321, 0.7785, 1.6337, 1.4338, 1.4060, 1.3323], + device='cuda:4'), covar=tensor([0.0788, 0.0860, 0.0851, 0.1245, 0.0645, 0.0902, 0.0881, 0.1337], + device='cuda:4'), in_proj_covar=tensor([0.0142, 0.0135, 0.0146, 0.0132, 0.0112, 0.0145, 0.0151, 0.0166], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:15:58,824 INFO [finetune.py:976] (4/7) Epoch 3, batch 1550, loss[loss=0.2274, simple_loss=0.2828, pruned_loss=0.08606, over 4755.00 frames. ], tot_loss[loss=0.2539, simple_loss=0.3058, pruned_loss=0.1011, over 958873.34 frames. ], batch size: 28, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:16:09,622 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.917e+01 1.840e+02 2.219e+02 2.788e+02 8.539e+02, threshold=4.437e+02, percent-clipped=2.0 +2023-03-26 01:16:32,113 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.56 vs. limit=5.0 +2023-03-26 01:16:44,396 INFO [finetune.py:976] (4/7) Epoch 3, batch 1600, loss[loss=0.2459, simple_loss=0.2928, pruned_loss=0.09957, over 4757.00 frames. ], tot_loss[loss=0.2515, simple_loss=0.3027, pruned_loss=0.1002, over 956356.19 frames. ], batch size: 54, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:17:12,649 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9771, 1.7505, 1.3846, 1.8777, 1.9426, 1.6330, 2.3522, 1.8525], + device='cuda:4'), covar=tensor([0.2511, 0.5065, 0.5489, 0.4896, 0.3648, 0.2559, 0.4469, 0.3179], + device='cuda:4'), in_proj_covar=tensor([0.0165, 0.0198, 0.0240, 0.0256, 0.0221, 0.0185, 0.0209, 0.0190], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:17:41,255 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=13102.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:17:43,644 INFO [finetune.py:976] (4/7) Epoch 3, batch 1650, loss[loss=0.2051, simple_loss=0.2637, pruned_loss=0.07327, over 4802.00 frames. ], tot_loss[loss=0.2486, simple_loss=0.2991, pruned_loss=0.09901, over 956272.90 frames. ], batch size: 51, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:17:55,149 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.264e+02 1.841e+02 2.109e+02 2.450e+02 4.226e+02, threshold=4.217e+02, percent-clipped=0.0 +2023-03-26 01:17:55,250 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7475, 3.9833, 3.9244, 2.0739, 4.0375, 3.0801, 0.9745, 2.9760], + device='cuda:4'), covar=tensor([0.2499, 0.1655, 0.1475, 0.3330, 0.0973, 0.0977, 0.4767, 0.1544], + device='cuda:4'), in_proj_covar=tensor([0.0157, 0.0169, 0.0167, 0.0130, 0.0157, 0.0122, 0.0148, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 01:17:56,997 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=13119.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 01:18:08,373 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7048, 1.5296, 1.9671, 3.1090, 2.2259, 2.2918, 0.8987, 2.4815], + device='cuda:4'), covar=tensor([0.1681, 0.1563, 0.1302, 0.0587, 0.0800, 0.1330, 0.1926, 0.0660], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0120, 0.0137, 0.0164, 0.0105, 0.0144, 0.0130, 0.0107], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 01:18:32,151 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=13150.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:18:41,625 INFO [finetune.py:976] (4/7) Epoch 3, batch 1700, loss[loss=0.2598, simple_loss=0.3097, pruned_loss=0.105, over 4914.00 frames. ], tot_loss[loss=0.2473, simple_loss=0.2975, pruned_loss=0.09852, over 957126.79 frames. ], batch size: 36, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:18:42,549 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 01:19:15,670 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4494, 1.4927, 1.1433, 1.3427, 1.7035, 1.8055, 1.5791, 1.2845], + device='cuda:4'), covar=tensor([0.0341, 0.0407, 0.0727, 0.0398, 0.0312, 0.0403, 0.0283, 0.0446], + device='cuda:4'), in_proj_covar=tensor([0.0083, 0.0114, 0.0136, 0.0116, 0.0105, 0.0098, 0.0089, 0.0109], + device='cuda:4'), out_proj_covar=tensor([6.5258e-05, 9.0123e-05, 1.0990e-04, 9.2348e-05, 8.3065e-05, 7.3436e-05, + 6.8502e-05, 8.5601e-05], device='cuda:4') +2023-03-26 01:19:22,356 INFO [finetune.py:976] (4/7) Epoch 3, batch 1750, loss[loss=0.3078, simple_loss=0.3685, pruned_loss=0.1235, over 4852.00 frames. ], tot_loss[loss=0.2508, simple_loss=0.3007, pruned_loss=0.1005, over 956987.31 frames. ], batch size: 49, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:19:31,217 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.198e+02 1.973e+02 2.314e+02 2.693e+02 6.749e+02, threshold=4.629e+02, percent-clipped=3.0 +2023-03-26 01:19:41,587 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5936, 1.4194, 1.3023, 1.3333, 1.6846, 1.4003, 1.7202, 1.5299], + device='cuda:4'), covar=tensor([0.2241, 0.4211, 0.4952, 0.4132, 0.3457, 0.2520, 0.3486, 0.2984], + device='cuda:4'), in_proj_covar=tensor([0.0164, 0.0197, 0.0239, 0.0255, 0.0219, 0.0184, 0.0209, 0.0189], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:20:18,402 INFO [finetune.py:976] (4/7) Epoch 3, batch 1800, loss[loss=0.2766, simple_loss=0.3236, pruned_loss=0.1148, over 4925.00 frames. ], tot_loss[loss=0.254, simple_loss=0.3048, pruned_loss=0.1016, over 957321.23 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:20:45,148 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6204, 1.5432, 1.5005, 1.6483, 1.0792, 3.3189, 1.3276, 1.8329], + device='cuda:4'), covar=tensor([0.3389, 0.2477, 0.2028, 0.2269, 0.2058, 0.0218, 0.2816, 0.1341], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0110, 0.0115, 0.0117, 0.0115, 0.0096, 0.0100, 0.0096], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 01:20:59,917 INFO [finetune.py:976] (4/7) Epoch 3, batch 1850, loss[loss=0.3346, simple_loss=0.3552, pruned_loss=0.157, over 4867.00 frames. ], tot_loss[loss=0.255, simple_loss=0.3057, pruned_loss=0.1022, over 958087.42 frames. ], batch size: 34, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:21:08,031 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.104e+02 1.691e+02 1.934e+02 2.488e+02 4.482e+02, threshold=3.868e+02, percent-clipped=0.0 +2023-03-26 01:21:16,547 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6777, 1.8205, 1.7146, 1.0757, 2.0716, 1.9165, 1.8404, 1.6527], + device='cuda:4'), covar=tensor([0.0789, 0.0652, 0.0885, 0.1144, 0.0488, 0.0770, 0.0842, 0.1160], + device='cuda:4'), in_proj_covar=tensor([0.0142, 0.0136, 0.0148, 0.0133, 0.0112, 0.0146, 0.0152, 0.0166], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:21:50,069 INFO [finetune.py:976] (4/7) Epoch 3, batch 1900, loss[loss=0.2635, simple_loss=0.3223, pruned_loss=0.1024, over 4776.00 frames. ], tot_loss[loss=0.2558, simple_loss=0.3067, pruned_loss=0.1024, over 957932.60 frames. ], batch size: 28, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:22:40,355 INFO [finetune.py:976] (4/7) Epoch 3, batch 1950, loss[loss=0.2172, simple_loss=0.2687, pruned_loss=0.08289, over 4900.00 frames. ], tot_loss[loss=0.2528, simple_loss=0.3037, pruned_loss=0.1009, over 956011.52 frames. ], batch size: 32, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:22:46,997 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.849e+02 2.191e+02 2.472e+02 6.030e+02, threshold=4.381e+02, percent-clipped=4.0 +2023-03-26 01:22:48,823 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=13419.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:23:25,638 INFO [finetune.py:976] (4/7) Epoch 3, batch 2000, loss[loss=0.1818, simple_loss=0.2477, pruned_loss=0.05797, over 4846.00 frames. ], tot_loss[loss=0.251, simple_loss=0.3017, pruned_loss=0.1002, over 954663.45 frames. ], batch size: 47, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:23:36,608 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=13467.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:24:18,862 INFO [finetune.py:976] (4/7) Epoch 3, batch 2050, loss[loss=0.2492, simple_loss=0.2807, pruned_loss=0.1088, over 4739.00 frames. ], tot_loss[loss=0.2487, simple_loss=0.2988, pruned_loss=0.09932, over 954551.43 frames. ], batch size: 54, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:24:23,471 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.8542, 3.3895, 3.4812, 3.7161, 3.6015, 3.3359, 3.9376, 1.3515], + device='cuda:4'), covar=tensor([0.0873, 0.0829, 0.0921, 0.1063, 0.1358, 0.1494, 0.0781, 0.4818], + device='cuda:4'), in_proj_covar=tensor([0.0370, 0.0246, 0.0277, 0.0296, 0.0343, 0.0288, 0.0312, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:24:33,951 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.806e+02 2.206e+02 2.674e+02 5.377e+02, threshold=4.412e+02, percent-clipped=2.0 +2023-03-26 01:24:43,043 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3355, 1.4673, 1.3982, 1.6335, 1.6340, 3.0337, 1.3073, 1.6337], + device='cuda:4'), covar=tensor([0.1158, 0.1746, 0.1222, 0.1062, 0.1711, 0.0322, 0.1586, 0.1771], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0081, 0.0078, 0.0079, 0.0093, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 01:25:00,756 INFO [finetune.py:976] (4/7) Epoch 3, batch 2100, loss[loss=0.2916, simple_loss=0.3383, pruned_loss=0.1224, over 4802.00 frames. ], tot_loss[loss=0.2485, simple_loss=0.2987, pruned_loss=0.09921, over 954875.62 frames. ], batch size: 51, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:25:02,002 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3750, 1.2656, 1.2536, 0.6778, 1.0806, 1.3645, 1.4111, 1.1866], + device='cuda:4'), covar=tensor([0.0824, 0.0448, 0.0454, 0.0540, 0.0472, 0.0437, 0.0277, 0.0451], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0157, 0.0118, 0.0136, 0.0132, 0.0119, 0.0146, 0.0144], + device='cuda:4'), out_proj_covar=tensor([9.7899e-05, 1.1698e-04, 8.6579e-05, 9.9517e-05, 9.5834e-05, 8.8209e-05, + 1.0914e-04, 1.0695e-04], device='cuda:4') +2023-03-26 01:25:10,529 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5627, 1.4181, 1.9090, 1.3054, 1.5999, 1.6244, 1.4431, 1.9254], + device='cuda:4'), covar=tensor([0.1333, 0.1998, 0.1287, 0.1765, 0.0932, 0.1509, 0.2452, 0.0854], + device='cuda:4'), in_proj_covar=tensor([0.0206, 0.0203, 0.0204, 0.0197, 0.0180, 0.0227, 0.0214, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:25:45,307 INFO [finetune.py:976] (4/7) Epoch 3, batch 2150, loss[loss=0.3157, simple_loss=0.3522, pruned_loss=0.1396, over 4748.00 frames. ], tot_loss[loss=0.2519, simple_loss=0.3023, pruned_loss=0.1007, over 954847.22 frames. ], batch size: 54, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:26:01,369 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.918e+02 2.256e+02 2.684e+02 5.304e+02, threshold=4.512e+02, percent-clipped=2.0 +2023-03-26 01:26:24,733 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3491, 2.0169, 1.6098, 0.7895, 1.8636, 1.8571, 1.5480, 1.8194], + device='cuda:4'), covar=tensor([0.0991, 0.1148, 0.1818, 0.2678, 0.1576, 0.2469, 0.2626, 0.1305], + device='cuda:4'), in_proj_covar=tensor([0.0165, 0.0196, 0.0202, 0.0188, 0.0214, 0.0209, 0.0215, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:26:27,570 INFO [finetune.py:976] (4/7) Epoch 3, batch 2200, loss[loss=0.2949, simple_loss=0.3429, pruned_loss=0.1234, over 4821.00 frames. ], tot_loss[loss=0.2544, simple_loss=0.3052, pruned_loss=0.1018, over 954260.71 frames. ], batch size: 38, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:26:30,573 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6596, 3.7562, 3.6863, 1.9928, 3.9090, 2.8023, 0.8940, 2.6507], + device='cuda:4'), covar=tensor([0.2752, 0.1695, 0.1512, 0.3164, 0.0983, 0.1036, 0.4626, 0.1556], + device='cuda:4'), in_proj_covar=tensor([0.0156, 0.0168, 0.0166, 0.0129, 0.0156, 0.0121, 0.0147, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 01:27:00,298 INFO [finetune.py:976] (4/7) Epoch 3, batch 2250, loss[loss=0.2221, simple_loss=0.2834, pruned_loss=0.08047, over 4870.00 frames. ], tot_loss[loss=0.255, simple_loss=0.3065, pruned_loss=0.1018, over 956091.77 frames. ], batch size: 31, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:27:08,388 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.215e+02 1.927e+02 2.166e+02 2.564e+02 5.587e+02, threshold=4.333e+02, percent-clipped=2.0 +2023-03-26 01:27:24,309 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=13737.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:27:41,073 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2139, 1.8260, 1.3469, 0.5849, 1.5600, 1.8484, 1.4442, 1.6333], + device='cuda:4'), covar=tensor([0.0716, 0.0916, 0.1516, 0.2067, 0.1281, 0.2099, 0.2425, 0.1044], + device='cuda:4'), in_proj_covar=tensor([0.0164, 0.0196, 0.0202, 0.0187, 0.0214, 0.0208, 0.0215, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:27:42,152 INFO [finetune.py:976] (4/7) Epoch 3, batch 2300, loss[loss=0.2623, simple_loss=0.3145, pruned_loss=0.105, over 4775.00 frames. ], tot_loss[loss=0.2529, simple_loss=0.3052, pruned_loss=0.1003, over 955084.59 frames. ], batch size: 51, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:28:25,844 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=13798.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 01:28:36,680 INFO [finetune.py:976] (4/7) Epoch 3, batch 2350, loss[loss=0.2524, simple_loss=0.303, pruned_loss=0.1009, over 4868.00 frames. ], tot_loss[loss=0.2512, simple_loss=0.3024, pruned_loss=0.1, over 955485.65 frames. ], batch size: 31, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:28:50,252 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.589e+01 1.791e+02 2.182e+02 2.579e+02 6.380e+02, threshold=4.365e+02, percent-clipped=2.0 +2023-03-26 01:29:03,739 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6171, 1.5767, 1.5643, 1.6133, 1.0143, 3.1915, 1.2085, 1.7378], + device='cuda:4'), covar=tensor([0.3199, 0.2127, 0.1888, 0.2182, 0.1981, 0.0192, 0.2737, 0.1331], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0110, 0.0115, 0.0118, 0.0115, 0.0096, 0.0100, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 01:29:31,309 INFO [finetune.py:976] (4/7) Epoch 3, batch 2400, loss[loss=0.2559, simple_loss=0.3017, pruned_loss=0.1051, over 4844.00 frames. ], tot_loss[loss=0.2468, simple_loss=0.2977, pruned_loss=0.09799, over 954240.63 frames. ], batch size: 47, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:29:34,026 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7329, 1.5529, 1.4186, 1.5722, 1.8624, 1.4670, 2.0632, 1.6891], + device='cuda:4'), covar=tensor([0.2189, 0.4114, 0.4715, 0.4061, 0.3155, 0.2311, 0.3879, 0.2890], + device='cuda:4'), in_proj_covar=tensor([0.0164, 0.0196, 0.0238, 0.0253, 0.0219, 0.0183, 0.0207, 0.0187], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:30:15,688 INFO [finetune.py:976] (4/7) Epoch 3, batch 2450, loss[loss=0.226, simple_loss=0.2748, pruned_loss=0.08863, over 4708.00 frames. ], tot_loss[loss=0.245, simple_loss=0.2955, pruned_loss=0.09724, over 956518.22 frames. ], batch size: 23, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:30:29,105 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.953e+01 1.906e+02 2.150e+02 2.578e+02 4.181e+02, threshold=4.299e+02, percent-clipped=0.0 +2023-03-26 01:30:58,548 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6770, 2.3827, 1.9968, 1.0374, 2.1513, 2.1445, 1.8233, 2.0342], + device='cuda:4'), covar=tensor([0.0867, 0.0872, 0.1764, 0.2396, 0.1467, 0.2152, 0.2085, 0.1109], + device='cuda:4'), in_proj_covar=tensor([0.0165, 0.0197, 0.0204, 0.0189, 0.0215, 0.0209, 0.0216, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:30:59,052 INFO [finetune.py:976] (4/7) Epoch 3, batch 2500, loss[loss=0.2083, simple_loss=0.2648, pruned_loss=0.07594, over 4757.00 frames. ], tot_loss[loss=0.2457, simple_loss=0.2967, pruned_loss=0.09738, over 956453.50 frames. ], batch size: 26, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:31:53,429 INFO [finetune.py:976] (4/7) Epoch 3, batch 2550, loss[loss=0.1991, simple_loss=0.2536, pruned_loss=0.07232, over 4717.00 frames. ], tot_loss[loss=0.2482, simple_loss=0.3002, pruned_loss=0.0981, over 955130.87 frames. ], batch size: 23, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:32:02,459 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.924e+01 1.822e+02 2.075e+02 2.520e+02 4.375e+02, threshold=4.150e+02, percent-clipped=1.0 +2023-03-26 01:32:05,738 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9997, 1.7445, 1.4370, 1.7180, 1.7244, 1.6243, 1.5935, 2.5691], + device='cuda:4'), covar=tensor([0.9920, 1.0225, 0.8617, 1.1915, 0.9326, 0.5937, 1.0815, 0.3102], + device='cuda:4'), in_proj_covar=tensor([0.0268, 0.0244, 0.0217, 0.0280, 0.0233, 0.0194, 0.0235, 0.0182], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 01:32:21,736 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7064, 1.4941, 1.4572, 1.7632, 1.8734, 1.7368, 1.0209, 1.3790], + device='cuda:4'), covar=tensor([0.2658, 0.2812, 0.2304, 0.2023, 0.2166, 0.1385, 0.3464, 0.2192], + device='cuda:4'), in_proj_covar=tensor([0.0226, 0.0207, 0.0194, 0.0180, 0.0230, 0.0171, 0.0210, 0.0183], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:32:25,996 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5190, 1.3735, 1.3486, 1.6016, 1.9024, 1.5643, 1.0681, 1.2371], + device='cuda:4'), covar=tensor([0.2902, 0.2967, 0.2410, 0.2163, 0.2380, 0.1596, 0.3646, 0.2388], + device='cuda:4'), in_proj_covar=tensor([0.0227, 0.0207, 0.0194, 0.0180, 0.0230, 0.0171, 0.0210, 0.0184], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:32:26,225 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.65 vs. limit=2.0 +2023-03-26 01:32:30,593 INFO [finetune.py:976] (4/7) Epoch 3, batch 2600, loss[loss=0.2129, simple_loss=0.2712, pruned_loss=0.07728, over 4103.00 frames. ], tot_loss[loss=0.2473, simple_loss=0.2999, pruned_loss=0.0973, over 956911.11 frames. ], batch size: 65, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:32:41,360 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14072.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:33:04,944 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14093.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 01:33:16,409 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14103.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:33:18,155 INFO [finetune.py:976] (4/7) Epoch 3, batch 2650, loss[loss=0.2751, simple_loss=0.3201, pruned_loss=0.1151, over 4858.00 frames. ], tot_loss[loss=0.2486, simple_loss=0.3014, pruned_loss=0.09795, over 957048.79 frames. ], batch size: 31, lr: 3.98e-03, grad_scale: 16.0 +2023-03-26 01:33:28,665 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7233, 1.5645, 2.0257, 1.4454, 1.8142, 1.9794, 1.5735, 2.0893], + device='cuda:4'), covar=tensor([0.1597, 0.2226, 0.1463, 0.2020, 0.1129, 0.1592, 0.2610, 0.1061], + device='cuda:4'), in_proj_covar=tensor([0.0206, 0.0204, 0.0204, 0.0196, 0.0181, 0.0226, 0.0215, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:33:34,826 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.159e+02 1.796e+02 2.195e+02 2.771e+02 4.502e+02, threshold=4.390e+02, percent-clipped=2.0 +2023-03-26 01:33:37,256 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14120.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 01:33:56,205 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14133.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:34:18,023 INFO [finetune.py:976] (4/7) Epoch 3, batch 2700, loss[loss=0.2731, simple_loss=0.3021, pruned_loss=0.122, over 4821.00 frames. ], tot_loss[loss=0.2479, simple_loss=0.3008, pruned_loss=0.09753, over 958699.37 frames. ], batch size: 25, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:34:28,757 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14164.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:34:50,953 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14181.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 01:34:55,807 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 01:35:13,613 INFO [finetune.py:976] (4/7) Epoch 3, batch 2750, loss[loss=0.2351, simple_loss=0.2782, pruned_loss=0.09599, over 4728.00 frames. ], tot_loss[loss=0.2458, simple_loss=0.2979, pruned_loss=0.09691, over 956221.50 frames. ], batch size: 23, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:35:20,822 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.010e+02 1.613e+02 1.949e+02 2.418e+02 3.837e+02, threshold=3.898e+02, percent-clipped=0.0 +2023-03-26 01:35:30,363 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4733, 1.6807, 2.0330, 2.0347, 1.8312, 4.1580, 1.4482, 1.9078], + device='cuda:4'), covar=tensor([0.1038, 0.1604, 0.1087, 0.1020, 0.1496, 0.0176, 0.1491, 0.1658], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0081, 0.0078, 0.0080, 0.0093, 0.0083, 0.0086, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 01:35:50,233 INFO [finetune.py:976] (4/7) Epoch 3, batch 2800, loss[loss=0.2597, simple_loss=0.2931, pruned_loss=0.1132, over 4824.00 frames. ], tot_loss[loss=0.2435, simple_loss=0.2946, pruned_loss=0.0962, over 955945.57 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:36:18,459 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6553, 1.5336, 2.0532, 3.2665, 2.2747, 2.3853, 1.0376, 2.6121], + device='cuda:4'), covar=tensor([0.1845, 0.1600, 0.1362, 0.0571, 0.0838, 0.1534, 0.1893, 0.0641], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0120, 0.0139, 0.0166, 0.0105, 0.0146, 0.0131, 0.0108], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 01:36:35,462 INFO [finetune.py:976] (4/7) Epoch 3, batch 2850, loss[loss=0.2961, simple_loss=0.3359, pruned_loss=0.1281, over 4197.00 frames. ], tot_loss[loss=0.243, simple_loss=0.2939, pruned_loss=0.09604, over 954327.62 frames. ], batch size: 65, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:36:48,616 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.210e+02 1.714e+02 2.069e+02 2.397e+02 3.427e+02, threshold=4.138e+02, percent-clipped=0.0 +2023-03-26 01:37:24,349 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4236, 1.2195, 1.2079, 1.4037, 1.5157, 1.4156, 0.8781, 1.1867], + device='cuda:4'), covar=tensor([0.2397, 0.2469, 0.2034, 0.1855, 0.1913, 0.1281, 0.2896, 0.2078], + device='cuda:4'), in_proj_covar=tensor([0.0227, 0.0207, 0.0194, 0.0180, 0.0230, 0.0171, 0.0210, 0.0184], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:37:26,625 INFO [finetune.py:976] (4/7) Epoch 3, batch 2900, loss[loss=0.2375, simple_loss=0.3132, pruned_loss=0.08089, over 4809.00 frames. ], tot_loss[loss=0.2469, simple_loss=0.2978, pruned_loss=0.09805, over 953050.86 frames. ], batch size: 41, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:38:06,358 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14393.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:38:25,221 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.4518, 3.8581, 4.0365, 4.3162, 4.1891, 3.9631, 4.5676, 1.3460], + device='cuda:4'), covar=tensor([0.0691, 0.0755, 0.0782, 0.0840, 0.1064, 0.1292, 0.0568, 0.5262], + device='cuda:4'), in_proj_covar=tensor([0.0366, 0.0243, 0.0273, 0.0294, 0.0338, 0.0285, 0.0310, 0.0300], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:38:25,744 INFO [finetune.py:976] (4/7) Epoch 3, batch 2950, loss[loss=0.2416, simple_loss=0.3012, pruned_loss=0.091, over 4913.00 frames. ], tot_loss[loss=0.2499, simple_loss=0.3012, pruned_loss=0.0993, over 954775.10 frames. ], batch size: 38, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:38:35,062 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8039, 1.5371, 2.3310, 3.6586, 2.5061, 2.5334, 0.9764, 2.8261], + device='cuda:4'), covar=tensor([0.1815, 0.1643, 0.1264, 0.0595, 0.0775, 0.1403, 0.2057, 0.0596], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0119, 0.0138, 0.0165, 0.0104, 0.0144, 0.0130, 0.0107], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 01:38:35,806 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14413.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:38:38,198 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.306e+02 1.859e+02 2.169e+02 2.721e+02 5.785e+02, threshold=4.339e+02, percent-clipped=3.0 +2023-03-26 01:38:50,259 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14428.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:39:02,086 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=14441.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:39:19,196 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0279, 2.1152, 1.8639, 1.4810, 2.3351, 2.2582, 2.1489, 1.8377], + device='cuda:4'), covar=tensor([0.0807, 0.0616, 0.1037, 0.1134, 0.0390, 0.0773, 0.0779, 0.1073], + device='cuda:4'), in_proj_covar=tensor([0.0140, 0.0133, 0.0146, 0.0130, 0.0110, 0.0143, 0.0148, 0.0164], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:39:20,296 INFO [finetune.py:976] (4/7) Epoch 3, batch 3000, loss[loss=0.3012, simple_loss=0.3179, pruned_loss=0.1423, over 4421.00 frames. ], tot_loss[loss=0.2506, simple_loss=0.3019, pruned_loss=0.09966, over 953595.89 frames. ], batch size: 19, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:39:20,296 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 01:39:24,753 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0116, 1.6960, 1.6909, 1.7692, 2.0702, 1.7861, 2.2362, 1.8927], + device='cuda:4'), covar=tensor([0.2447, 0.4682, 0.5437, 0.4176, 0.3454, 0.2302, 0.4055, 0.3070], + device='cuda:4'), in_proj_covar=tensor([0.0164, 0.0195, 0.0239, 0.0254, 0.0220, 0.0184, 0.0208, 0.0188], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:39:28,627 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5412, 1.2762, 1.3593, 1.2969, 1.6906, 1.6332, 1.5023, 1.3395], + device='cuda:4'), covar=tensor([0.0301, 0.0328, 0.0615, 0.0345, 0.0297, 0.0400, 0.0337, 0.0417], + device='cuda:4'), in_proj_covar=tensor([0.0083, 0.0113, 0.0134, 0.0116, 0.0104, 0.0098, 0.0088, 0.0108], + device='cuda:4'), out_proj_covar=tensor([6.4727e-05, 8.9196e-05, 1.0814e-04, 9.1849e-05, 8.2649e-05, 7.2955e-05, + 6.7992e-05, 8.4359e-05], device='cuda:4') +2023-03-26 01:39:40,246 INFO [finetune.py:1010] (4/7) Epoch 3, validation: loss=0.1777, simple_loss=0.2485, pruned_loss=0.05342, over 2265189.00 frames. +2023-03-26 01:39:40,247 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6301MB +2023-03-26 01:39:42,146 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14459.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:40:02,191 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14474.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:40:03,390 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14476.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 01:40:04,231 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-03-26 01:40:15,393 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14487.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:40:36,154 INFO [finetune.py:976] (4/7) Epoch 3, batch 3050, loss[loss=0.258, simple_loss=0.3049, pruned_loss=0.1055, over 4809.00 frames. ], tot_loss[loss=0.2522, simple_loss=0.3036, pruned_loss=0.1004, over 955111.51 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:40:52,884 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.290e+02 1.934e+02 2.277e+02 2.724e+02 4.940e+02, threshold=4.554e+02, percent-clipped=2.0 +2023-03-26 01:40:53,141 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-03-26 01:40:55,453 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4223, 1.4279, 1.4606, 1.7056, 1.6121, 3.1771, 1.3418, 1.6278], + device='cuda:4'), covar=tensor([0.1088, 0.1779, 0.1276, 0.1083, 0.1596, 0.0275, 0.1598, 0.1787], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0081, 0.0078, 0.0079, 0.0093, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 01:41:18,006 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14548.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:41:22,792 INFO [finetune.py:976] (4/7) Epoch 3, batch 3100, loss[loss=0.2733, simple_loss=0.3076, pruned_loss=0.1195, over 4734.00 frames. ], tot_loss[loss=0.2481, simple_loss=0.2993, pruned_loss=0.09844, over 954732.06 frames. ], batch size: 54, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:41:25,226 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5983, 1.4142, 2.0063, 1.9438, 1.6363, 4.0612, 1.2899, 1.7346], + device='cuda:4'), covar=tensor([0.1029, 0.1711, 0.1197, 0.1037, 0.1567, 0.0206, 0.1554, 0.1670], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0081, 0.0078, 0.0079, 0.0093, 0.0083, 0.0085, 0.0078], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 01:41:38,410 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5341, 1.5429, 2.0092, 1.9177, 1.7503, 3.8198, 1.3817, 1.8455], + device='cuda:4'), covar=tensor([0.0953, 0.1435, 0.1186, 0.0939, 0.1326, 0.0194, 0.1330, 0.1446], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0081, 0.0078, 0.0079, 0.0093, 0.0083, 0.0085, 0.0078], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 01:42:02,522 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 01:42:04,510 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-26 01:42:10,659 INFO [finetune.py:976] (4/7) Epoch 3, batch 3150, loss[loss=0.2404, simple_loss=0.2962, pruned_loss=0.09231, over 4841.00 frames. ], tot_loss[loss=0.245, simple_loss=0.2958, pruned_loss=0.09711, over 955827.27 frames. ], batch size: 47, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:42:18,343 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.181e+02 1.775e+02 2.189e+02 2.683e+02 4.981e+02, threshold=4.378e+02, percent-clipped=2.0 +2023-03-26 01:42:36,247 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-26 01:43:00,564 INFO [finetune.py:976] (4/7) Epoch 3, batch 3200, loss[loss=0.2664, simple_loss=0.307, pruned_loss=0.1129, over 4909.00 frames. ], tot_loss[loss=0.2413, simple_loss=0.2923, pruned_loss=0.09518, over 957651.39 frames. ], batch size: 35, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:43:23,564 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7016, 1.5790, 1.5690, 1.9127, 2.2324, 1.7630, 1.4320, 1.4001], + device='cuda:4'), covar=tensor([0.2374, 0.2332, 0.1904, 0.1690, 0.1984, 0.1265, 0.2792, 0.1956], + device='cuda:4'), in_proj_covar=tensor([0.0228, 0.0207, 0.0194, 0.0181, 0.0231, 0.0171, 0.0211, 0.0185], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:43:41,546 INFO [finetune.py:976] (4/7) Epoch 3, batch 3250, loss[loss=0.2622, simple_loss=0.2955, pruned_loss=0.1145, over 4767.00 frames. ], tot_loss[loss=0.2419, simple_loss=0.2926, pruned_loss=0.09563, over 956390.13 frames. ], batch size: 26, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:43:54,504 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.171e+02 1.718e+02 2.102e+02 2.544e+02 5.358e+02, threshold=4.204e+02, percent-clipped=1.0 +2023-03-26 01:44:04,566 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.69 vs. limit=5.0 +2023-03-26 01:44:08,046 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14728.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:44:15,445 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14740.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:44:19,078 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6766, 1.5136, 1.4577, 1.7647, 2.2691, 1.6922, 1.3886, 1.2904], + device='cuda:4'), covar=tensor([0.2447, 0.2472, 0.2104, 0.1910, 0.2015, 0.1349, 0.3076, 0.2023], + device='cuda:4'), in_proj_covar=tensor([0.0227, 0.0207, 0.0194, 0.0180, 0.0230, 0.0170, 0.0211, 0.0184], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:44:31,816 INFO [finetune.py:976] (4/7) Epoch 3, batch 3300, loss[loss=0.3314, simple_loss=0.3734, pruned_loss=0.1447, over 4899.00 frames. ], tot_loss[loss=0.2456, simple_loss=0.2972, pruned_loss=0.09696, over 954350.24 frames. ], batch size: 36, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:44:35,968 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14759.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:44:42,748 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14769.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:44:48,153 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0420, 0.9816, 1.0205, 0.2789, 0.7916, 1.1864, 1.1971, 1.0533], + device='cuda:4'), covar=tensor([0.1120, 0.0778, 0.0470, 0.0760, 0.0560, 0.0606, 0.0454, 0.0633], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0159, 0.0119, 0.0138, 0.0133, 0.0121, 0.0149, 0.0144], + device='cuda:4'), out_proj_covar=tensor([9.9509e-05, 1.1828e-04, 8.6378e-05, 1.0103e-04, 9.6583e-05, 8.9218e-05, + 1.1096e-04, 1.0709e-04], device='cuda:4') +2023-03-26 01:44:48,717 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=14776.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:44:48,744 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14776.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 01:44:56,768 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7973, 1.6428, 1.3736, 1.7061, 1.8213, 1.4673, 2.2160, 1.7115], + device='cuda:4'), covar=tensor([0.2049, 0.4073, 0.4532, 0.3996, 0.3282, 0.2173, 0.3701, 0.2879], + device='cuda:4'), in_proj_covar=tensor([0.0164, 0.0195, 0.0238, 0.0254, 0.0220, 0.0184, 0.0208, 0.0188], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:45:09,029 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14801.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:45:16,878 INFO [finetune.py:976] (4/7) Epoch 3, batch 3350, loss[loss=0.213, simple_loss=0.2811, pruned_loss=0.07246, over 4754.00 frames. ], tot_loss[loss=0.2472, simple_loss=0.2994, pruned_loss=0.09753, over 955185.47 frames. ], batch size: 27, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:45:17,544 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=14807.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:45:17,608 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3219, 2.2264, 2.8724, 1.8338, 2.5643, 2.6229, 2.1021, 2.6772], + device='cuda:4'), covar=tensor([0.1835, 0.1929, 0.1514, 0.2571, 0.1064, 0.1939, 0.2642, 0.1093], + device='cuda:4'), in_proj_covar=tensor([0.0207, 0.0206, 0.0205, 0.0197, 0.0182, 0.0226, 0.0217, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:45:29,676 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.682e+02 2.084e+02 2.593e+02 4.183e+02, threshold=4.169e+02, percent-clipped=0.0 +2023-03-26 01:45:31,586 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6735, 1.5232, 1.4938, 1.7353, 2.2001, 1.7540, 1.3334, 1.3688], + device='cuda:4'), covar=tensor([0.2487, 0.2563, 0.2193, 0.1980, 0.2219, 0.1328, 0.3033, 0.2071], + device='cuda:4'), in_proj_covar=tensor([0.0229, 0.0208, 0.0196, 0.0181, 0.0232, 0.0172, 0.0212, 0.0186], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:45:39,451 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=14824.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 01:45:52,839 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14843.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:46:01,258 INFO [finetune.py:976] (4/7) Epoch 3, batch 3400, loss[loss=0.2694, simple_loss=0.3188, pruned_loss=0.11, over 4848.00 frames. ], tot_loss[loss=0.2478, simple_loss=0.3003, pruned_loss=0.09763, over 954972.03 frames. ], batch size: 31, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:46:41,114 INFO [finetune.py:976] (4/7) Epoch 3, batch 3450, loss[loss=0.2489, simple_loss=0.3052, pruned_loss=0.09635, over 4885.00 frames. ], tot_loss[loss=0.2464, simple_loss=0.2994, pruned_loss=0.09673, over 954101.17 frames. ], batch size: 43, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:46:53,149 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.352e+02 1.935e+02 2.237e+02 2.692e+02 3.962e+02, threshold=4.475e+02, percent-clipped=0.0 +2023-03-26 01:47:27,612 INFO [finetune.py:976] (4/7) Epoch 3, batch 3500, loss[loss=0.2707, simple_loss=0.315, pruned_loss=0.1132, over 4862.00 frames. ], tot_loss[loss=0.2443, simple_loss=0.297, pruned_loss=0.09576, over 955745.87 frames. ], batch size: 44, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:47:35,090 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5368, 1.3991, 1.3569, 1.6759, 1.9626, 1.5201, 1.0887, 1.2951], + device='cuda:4'), covar=tensor([0.2770, 0.2871, 0.2362, 0.2079, 0.2131, 0.1565, 0.3380, 0.2302], + device='cuda:4'), in_proj_covar=tensor([0.0228, 0.0208, 0.0195, 0.0181, 0.0231, 0.0171, 0.0211, 0.0185], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:48:19,807 INFO [finetune.py:976] (4/7) Epoch 3, batch 3550, loss[loss=0.204, simple_loss=0.2733, pruned_loss=0.06734, over 4762.00 frames. ], tot_loss[loss=0.2424, simple_loss=0.295, pruned_loss=0.09489, over 956990.68 frames. ], batch size: 28, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:48:26,974 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.116e+02 1.780e+02 2.199e+02 2.800e+02 5.904e+02, threshold=4.398e+02, percent-clipped=2.0 +2023-03-26 01:48:27,780 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2738, 1.9795, 1.7047, 2.0362, 1.9431, 1.9116, 1.7789, 2.7932], + device='cuda:4'), covar=tensor([1.0800, 1.0833, 0.8456, 1.0655, 0.8915, 0.6014, 1.0979, 0.3338], + device='cuda:4'), in_proj_covar=tensor([0.0272, 0.0249, 0.0219, 0.0284, 0.0235, 0.0196, 0.0239, 0.0185], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 01:48:37,082 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6836, 1.5326, 1.6203, 1.6225, 1.0691, 3.2639, 1.2170, 1.7636], + device='cuda:4'), covar=tensor([0.3317, 0.2349, 0.1884, 0.2310, 0.1976, 0.0182, 0.2762, 0.1368], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0111, 0.0116, 0.0119, 0.0116, 0.0097, 0.0101, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 01:49:03,542 INFO [finetune.py:976] (4/7) Epoch 3, batch 3600, loss[loss=0.2129, simple_loss=0.2738, pruned_loss=0.07598, over 4932.00 frames. ], tot_loss[loss=0.2393, simple_loss=0.2912, pruned_loss=0.09373, over 955145.90 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:49:12,126 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15069.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:49:12,839 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 01:49:15,761 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0321, 2.4942, 1.9524, 1.6421, 2.1544, 2.5113, 2.3747, 2.0322], + device='cuda:4'), covar=tensor([0.0827, 0.0476, 0.0941, 0.1011, 0.1050, 0.0617, 0.0611, 0.0902], + device='cuda:4'), in_proj_covar=tensor([0.0140, 0.0132, 0.0145, 0.0130, 0.0110, 0.0144, 0.0147, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:49:24,722 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.72 vs. limit=5.0 +2023-03-26 01:49:42,970 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=15096.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:49:46,394 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 01:49:49,068 INFO [finetune.py:976] (4/7) Epoch 3, batch 3650, loss[loss=0.3012, simple_loss=0.3419, pruned_loss=0.1302, over 4912.00 frames. ], tot_loss[loss=0.2429, simple_loss=0.2944, pruned_loss=0.09575, over 953667.45 frames. ], batch size: 37, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:49:56,314 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.204e+02 1.927e+02 2.238e+02 2.686e+02 4.916e+02, threshold=4.476e+02, percent-clipped=1.0 +2023-03-26 01:49:56,388 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=15117.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:50:28,403 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15143.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:50:30,784 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8473, 1.5900, 1.3332, 1.5925, 1.5121, 1.5349, 1.4358, 2.3350], + device='cuda:4'), covar=tensor([1.1244, 1.1413, 0.8584, 1.1696, 0.9537, 0.6080, 1.1383, 0.3882], + device='cuda:4'), in_proj_covar=tensor([0.0271, 0.0249, 0.0219, 0.0283, 0.0235, 0.0196, 0.0239, 0.0185], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 01:50:38,163 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7445, 4.3903, 4.2067, 2.5052, 4.4987, 3.3094, 0.9497, 3.0277], + device='cuda:4'), covar=tensor([0.2653, 0.1715, 0.1268, 0.2784, 0.0756, 0.0848, 0.4315, 0.1444], + device='cuda:4'), in_proj_covar=tensor([0.0156, 0.0168, 0.0165, 0.0128, 0.0155, 0.0121, 0.0146, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 01:50:41,688 INFO [finetune.py:976] (4/7) Epoch 3, batch 3700, loss[loss=0.2607, simple_loss=0.3204, pruned_loss=0.1005, over 4810.00 frames. ], tot_loss[loss=0.2463, simple_loss=0.2983, pruned_loss=0.09715, over 954735.82 frames. ], batch size: 45, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:50:49,395 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.6666, 2.7519, 2.4504, 1.4379, 2.8516, 2.4928, 1.9141, 2.2658], + device='cuda:4'), covar=tensor([0.0477, 0.1657, 0.2318, 0.2958, 0.2156, 0.2105, 0.2794, 0.1895], + device='cuda:4'), in_proj_covar=tensor([0.0166, 0.0198, 0.0205, 0.0189, 0.0217, 0.0211, 0.0217, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:51:07,480 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 01:51:16,177 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=15191.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:51:19,078 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-03-26 01:51:19,620 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=15195.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:51:29,140 INFO [finetune.py:976] (4/7) Epoch 3, batch 3750, loss[loss=0.2607, simple_loss=0.3196, pruned_loss=0.101, over 4850.00 frames. ], tot_loss[loss=0.2478, simple_loss=0.3, pruned_loss=0.09778, over 951549.16 frames. ], batch size: 44, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:51:37,941 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 01:51:40,515 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.801e+02 2.153e+02 2.622e+02 6.720e+02, threshold=4.305e+02, percent-clipped=1.0 +2023-03-26 01:52:30,227 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7050, 1.7779, 2.1265, 2.0684, 1.8657, 4.4362, 1.6989, 2.0893], + device='cuda:4'), covar=tensor([0.1062, 0.1665, 0.1170, 0.1068, 0.1629, 0.0223, 0.1379, 0.1677], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0081, 0.0078, 0.0079, 0.0093, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 01:52:33,791 INFO [finetune.py:976] (4/7) Epoch 3, batch 3800, loss[loss=0.253, simple_loss=0.3093, pruned_loss=0.09835, over 4806.00 frames. ], tot_loss[loss=0.2471, simple_loss=0.2999, pruned_loss=0.09718, over 952205.90 frames. ], batch size: 39, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:52:33,935 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=15256.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 01:52:55,296 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4831, 2.0801, 1.5367, 0.7187, 1.8266, 1.9662, 1.6302, 1.8280], + device='cuda:4'), covar=tensor([0.0739, 0.0888, 0.1668, 0.2329, 0.1433, 0.2260, 0.2323, 0.1114], + device='cuda:4'), in_proj_covar=tensor([0.0167, 0.0199, 0.0206, 0.0190, 0.0219, 0.0213, 0.0218, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 01:53:22,427 INFO [finetune.py:976] (4/7) Epoch 3, batch 3850, loss[loss=0.2035, simple_loss=0.2652, pruned_loss=0.07087, over 4844.00 frames. ], tot_loss[loss=0.2455, simple_loss=0.298, pruned_loss=0.09646, over 951579.24 frames. ], batch size: 44, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:53:39,318 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.276e+02 1.876e+02 2.253e+02 2.579e+02 5.032e+02, threshold=4.505e+02, percent-clipped=1.0 +2023-03-26 01:54:26,485 INFO [finetune.py:976] (4/7) Epoch 3, batch 3900, loss[loss=0.245, simple_loss=0.2859, pruned_loss=0.1021, over 4826.00 frames. ], tot_loss[loss=0.2442, simple_loss=0.2962, pruned_loss=0.09612, over 952904.89 frames. ], batch size: 30, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:54:35,553 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7425, 1.2622, 0.9414, 1.5783, 2.0656, 0.9900, 1.4361, 1.6402], + device='cuda:4'), covar=tensor([0.1449, 0.2005, 0.2053, 0.1099, 0.1993, 0.2090, 0.1423, 0.1972], + device='cuda:4'), in_proj_covar=tensor([0.0092, 0.0098, 0.0117, 0.0093, 0.0124, 0.0097, 0.0099, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 01:54:49,189 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6615, 1.5563, 1.5327, 1.8021, 1.1730, 3.5760, 1.3781, 1.9226], + device='cuda:4'), covar=tensor([0.3708, 0.2659, 0.2147, 0.2338, 0.2049, 0.0157, 0.2860, 0.1466], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0111, 0.0116, 0.0120, 0.0116, 0.0097, 0.0101, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 01:54:52,366 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-26 01:55:10,115 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15396.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:55:14,076 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 01:55:14,921 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 01:55:16,821 INFO [finetune.py:976] (4/7) Epoch 3, batch 3950, loss[loss=0.2271, simple_loss=0.2776, pruned_loss=0.08828, over 4723.00 frames. ], tot_loss[loss=0.2384, simple_loss=0.2909, pruned_loss=0.09301, over 954808.61 frames. ], batch size: 23, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:55:25,255 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.128e+02 1.678e+02 2.164e+02 2.472e+02 4.231e+02, threshold=4.328e+02, percent-clipped=0.0 +2023-03-26 01:55:52,198 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=15444.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 01:55:59,830 INFO [finetune.py:976] (4/7) Epoch 3, batch 4000, loss[loss=0.3279, simple_loss=0.358, pruned_loss=0.1488, over 4816.00 frames. ], tot_loss[loss=0.2396, simple_loss=0.2913, pruned_loss=0.09393, over 956555.91 frames. ], batch size: 41, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:57:05,461 INFO [finetune.py:976] (4/7) Epoch 3, batch 4050, loss[loss=0.1828, simple_loss=0.2576, pruned_loss=0.05402, over 4825.00 frames. ], tot_loss[loss=0.2421, simple_loss=0.2947, pruned_loss=0.09476, over 955275.70 frames. ], batch size: 39, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:57:20,269 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.142e+02 1.798e+02 2.110e+02 2.647e+02 5.396e+02, threshold=4.219e+02, percent-clipped=2.0 +2023-03-26 01:57:58,330 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=15551.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 01:58:06,353 INFO [finetune.py:976] (4/7) Epoch 3, batch 4100, loss[loss=0.2709, simple_loss=0.3129, pruned_loss=0.1145, over 4734.00 frames. ], tot_loss[loss=0.2446, simple_loss=0.2975, pruned_loss=0.09585, over 954704.74 frames. ], batch size: 59, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:59:02,794 INFO [finetune.py:976] (4/7) Epoch 3, batch 4150, loss[loss=0.2303, simple_loss=0.2903, pruned_loss=0.08516, over 4802.00 frames. ], tot_loss[loss=0.245, simple_loss=0.2981, pruned_loss=0.09597, over 954206.32 frames. ], batch size: 39, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 01:59:10,586 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.238e+02 1.791e+02 2.157e+02 2.467e+02 4.537e+02, threshold=4.313e+02, percent-clipped=1.0 +2023-03-26 01:59:51,465 INFO [finetune.py:976] (4/7) Epoch 3, batch 4200, loss[loss=0.2432, simple_loss=0.2962, pruned_loss=0.09513, over 4819.00 frames. ], tot_loss[loss=0.2451, simple_loss=0.2987, pruned_loss=0.09581, over 954488.91 frames. ], batch size: 33, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 02:00:53,256 INFO [finetune.py:976] (4/7) Epoch 3, batch 4250, loss[loss=0.2481, simple_loss=0.299, pruned_loss=0.0986, over 4759.00 frames. ], tot_loss[loss=0.2431, simple_loss=0.2964, pruned_loss=0.09487, over 952993.39 frames. ], batch size: 26, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 02:01:00,002 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.228e+02 1.731e+02 2.073e+02 2.469e+02 5.386e+02, threshold=4.147e+02, percent-clipped=2.0 +2023-03-26 02:01:13,998 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 02:01:43,527 INFO [finetune.py:976] (4/7) Epoch 3, batch 4300, loss[loss=0.2353, simple_loss=0.2936, pruned_loss=0.08846, over 4931.00 frames. ], tot_loss[loss=0.2405, simple_loss=0.2934, pruned_loss=0.09375, over 953910.18 frames. ], batch size: 37, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 02:02:02,519 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-03-26 02:02:43,903 INFO [finetune.py:976] (4/7) Epoch 3, batch 4350, loss[loss=0.2276, simple_loss=0.2598, pruned_loss=0.09768, over 4156.00 frames. ], tot_loss[loss=0.2365, simple_loss=0.2892, pruned_loss=0.09191, over 952833.53 frames. ], batch size: 18, lr: 3.98e-03, grad_scale: 32.0 +2023-03-26 02:02:55,477 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.21 vs. limit=5.0 +2023-03-26 02:03:01,714 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.218e+02 1.712e+02 2.007e+02 2.460e+02 4.679e+02, threshold=4.015e+02, percent-clipped=1.0 +2023-03-26 02:03:12,474 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 02:03:33,442 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15851.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:03:35,939 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8160, 1.8829, 1.8913, 1.2656, 2.0571, 2.0622, 2.0267, 1.5977], + device='cuda:4'), covar=tensor([0.0572, 0.0567, 0.0666, 0.1002, 0.0476, 0.0584, 0.0532, 0.1027], + device='cuda:4'), in_proj_covar=tensor([0.0141, 0.0134, 0.0145, 0.0131, 0.0111, 0.0144, 0.0149, 0.0164], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 02:03:36,424 INFO [finetune.py:976] (4/7) Epoch 3, batch 4400, loss[loss=0.3443, simple_loss=0.3689, pruned_loss=0.1598, over 4793.00 frames. ], tot_loss[loss=0.2388, simple_loss=0.2911, pruned_loss=0.09325, over 954273.64 frames. ], batch size: 51, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:04:05,497 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=15883.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:04:20,878 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=15899.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:04:30,835 INFO [finetune.py:976] (4/7) Epoch 3, batch 4450, loss[loss=0.2897, simple_loss=0.3438, pruned_loss=0.1178, over 4840.00 frames. ], tot_loss[loss=0.244, simple_loss=0.2966, pruned_loss=0.09568, over 954162.68 frames. ], batch size: 47, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:04:40,556 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7791, 1.2156, 0.9639, 1.6121, 2.0916, 1.0380, 1.4488, 1.5707], + device='cuda:4'), covar=tensor([0.1545, 0.2275, 0.2160, 0.1326, 0.2075, 0.2214, 0.1603, 0.2233], + device='cuda:4'), in_proj_covar=tensor([0.0092, 0.0098, 0.0117, 0.0094, 0.0125, 0.0098, 0.0100, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 02:04:48,095 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.872e+02 2.224e+02 2.526e+02 5.583e+02, threshold=4.448e+02, percent-clipped=1.0 +2023-03-26 02:04:59,228 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-03-26 02:05:16,020 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=15944.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:05:23,249 INFO [finetune.py:976] (4/7) Epoch 3, batch 4500, loss[loss=0.249, simple_loss=0.2883, pruned_loss=0.1048, over 4787.00 frames. ], tot_loss[loss=0.2432, simple_loss=0.2964, pruned_loss=0.095, over 953564.86 frames. ], batch size: 25, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:06:21,378 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16001.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:06:22,214 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.20 vs. limit=5.0 +2023-03-26 02:06:24,312 INFO [finetune.py:976] (4/7) Epoch 3, batch 4550, loss[loss=0.2432, simple_loss=0.3115, pruned_loss=0.08745, over 4889.00 frames. ], tot_loss[loss=0.2453, simple_loss=0.2983, pruned_loss=0.09611, over 953702.09 frames. ], batch size: 37, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:06:36,869 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.237e+02 1.760e+02 2.085e+02 2.487e+02 3.865e+02, threshold=4.170e+02, percent-clipped=0.0 +2023-03-26 02:07:05,905 INFO [finetune.py:976] (4/7) Epoch 3, batch 4600, loss[loss=0.2515, simple_loss=0.2989, pruned_loss=0.102, over 4710.00 frames. ], tot_loss[loss=0.245, simple_loss=0.298, pruned_loss=0.09599, over 951577.40 frames. ], batch size: 59, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:07:07,957 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 02:07:11,902 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16062.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:07:35,238 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16085.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:07:41,546 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9123, 1.7414, 1.4264, 1.7836, 1.7322, 1.6513, 1.6033, 2.5058], + device='cuda:4'), covar=tensor([1.1381, 1.2261, 0.8925, 1.1924, 0.9344, 0.6196, 1.1713, 0.3499], + device='cuda:4'), in_proj_covar=tensor([0.0272, 0.0248, 0.0217, 0.0282, 0.0234, 0.0195, 0.0237, 0.0185], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 02:07:59,849 INFO [finetune.py:976] (4/7) Epoch 3, batch 4650, loss[loss=0.2747, simple_loss=0.3078, pruned_loss=0.1208, over 4794.00 frames. ], tot_loss[loss=0.2419, simple_loss=0.2948, pruned_loss=0.09453, over 953666.58 frames. ], batch size: 51, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:08:07,263 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.007e+02 1.817e+02 2.201e+02 2.598e+02 3.850e+02, threshold=4.403e+02, percent-clipped=0.0 +2023-03-26 02:08:39,882 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16146.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:08:42,106 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16149.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:08:52,384 INFO [finetune.py:976] (4/7) Epoch 3, batch 4700, loss[loss=0.2817, simple_loss=0.3219, pruned_loss=0.1208, over 4843.00 frames. ], tot_loss[loss=0.2382, simple_loss=0.2913, pruned_loss=0.09257, over 954298.41 frames. ], batch size: 44, lr: 3.97e-03, grad_scale: 64.0 +2023-03-26 02:08:59,686 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-26 02:09:41,915 INFO [finetune.py:976] (4/7) Epoch 3, batch 4750, loss[loss=0.2685, simple_loss=0.31, pruned_loss=0.1135, over 4727.00 frames. ], tot_loss[loss=0.2357, simple_loss=0.2884, pruned_loss=0.09157, over 953847.33 frames. ], batch size: 59, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:09:44,988 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16210.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:09:49,754 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.299e+02 1.759e+02 2.065e+02 2.416e+02 4.123e+02, threshold=4.129e+02, percent-clipped=0.0 +2023-03-26 02:10:03,619 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16239.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:10:13,289 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3163, 1.1643, 1.5197, 2.3950, 1.6671, 1.9951, 0.9020, 1.9635], + device='cuda:4'), covar=tensor([0.1774, 0.1589, 0.1236, 0.0758, 0.0945, 0.1418, 0.1566, 0.0810], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0120, 0.0139, 0.0166, 0.0104, 0.0144, 0.0130, 0.0108], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 02:10:16,139 INFO [finetune.py:976] (4/7) Epoch 3, batch 4800, loss[loss=0.1731, simple_loss=0.2295, pruned_loss=0.0584, over 4199.00 frames. ], tot_loss[loss=0.2377, simple_loss=0.2905, pruned_loss=0.0925, over 954457.63 frames. ], batch size: 18, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:10:24,550 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5900, 1.2827, 2.0671, 3.3228, 2.2907, 2.3878, 1.0652, 2.6010], + device='cuda:4'), covar=tensor([0.1943, 0.1826, 0.1483, 0.0638, 0.0894, 0.1499, 0.2006, 0.0688], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0120, 0.0139, 0.0166, 0.0104, 0.0144, 0.0130, 0.0107], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 02:10:42,987 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.53 vs. limit=5.0 +2023-03-26 02:11:05,959 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3843, 1.3770, 1.4400, 0.7858, 1.5328, 1.4991, 1.4372, 1.3088], + device='cuda:4'), covar=tensor([0.0692, 0.0762, 0.0750, 0.1139, 0.0717, 0.0765, 0.0795, 0.1287], + device='cuda:4'), in_proj_covar=tensor([0.0140, 0.0133, 0.0144, 0.0130, 0.0111, 0.0143, 0.0148, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 02:11:10,320 INFO [finetune.py:976] (4/7) Epoch 3, batch 4850, loss[loss=0.2563, simple_loss=0.3096, pruned_loss=0.1015, over 4903.00 frames. ], tot_loss[loss=0.2409, simple_loss=0.2939, pruned_loss=0.09388, over 954413.78 frames. ], batch size: 37, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:11:18,710 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.278e+02 1.868e+02 2.289e+02 2.628e+02 4.977e+02, threshold=4.577e+02, percent-clipped=4.0 +2023-03-26 02:11:48,070 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-26 02:11:53,006 INFO [finetune.py:976] (4/7) Epoch 3, batch 4900, loss[loss=0.2632, simple_loss=0.3109, pruned_loss=0.1078, over 4917.00 frames. ], tot_loss[loss=0.2435, simple_loss=0.2964, pruned_loss=0.09528, over 952650.61 frames. ], batch size: 38, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:11:53,735 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16357.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:11:57,829 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.09 vs. limit=5.0 +2023-03-26 02:12:04,193 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16371.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:12:35,109 INFO [finetune.py:976] (4/7) Epoch 3, batch 4950, loss[loss=0.2678, simple_loss=0.3219, pruned_loss=0.1068, over 4808.00 frames. ], tot_loss[loss=0.2444, simple_loss=0.2973, pruned_loss=0.09575, over 950832.93 frames. ], batch size: 45, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:12:43,742 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.165e+02 1.793e+02 2.170e+02 2.564e+02 4.726e+02, threshold=4.340e+02, percent-clipped=1.0 +2023-03-26 02:12:53,379 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16432.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:12:56,907 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6345, 2.1927, 3.0229, 2.2184, 2.7735, 3.1042, 2.3717, 3.1811], + device='cuda:4'), covar=tensor([0.1501, 0.2127, 0.1484, 0.1825, 0.0929, 0.1303, 0.2426, 0.0894], + device='cuda:4'), in_proj_covar=tensor([0.0208, 0.0208, 0.0205, 0.0198, 0.0184, 0.0228, 0.0218, 0.0206], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 02:12:59,268 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16441.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:13:11,535 INFO [finetune.py:976] (4/7) Epoch 3, batch 5000, loss[loss=0.2621, simple_loss=0.3069, pruned_loss=0.1086, over 4818.00 frames. ], tot_loss[loss=0.2426, simple_loss=0.2958, pruned_loss=0.09476, over 952822.76 frames. ], batch size: 39, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:13:48,437 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1221, 1.8069, 2.0761, 0.8429, 2.2750, 2.4774, 1.9847, 2.0000], + device='cuda:4'), covar=tensor([0.0952, 0.0730, 0.0545, 0.0842, 0.0477, 0.0406, 0.0499, 0.0523], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0158, 0.0118, 0.0137, 0.0132, 0.0120, 0.0147, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.8568e-05, 1.1790e-04, 8.5616e-05, 1.0072e-04, 9.5764e-05, 8.8962e-05, + 1.0928e-04, 1.0707e-04], device='cuda:4') +2023-03-26 02:14:08,365 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16505.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:14:08,894 INFO [finetune.py:976] (4/7) Epoch 3, batch 5050, loss[loss=0.2209, simple_loss=0.278, pruned_loss=0.08187, over 4780.00 frames. ], tot_loss[loss=0.2382, simple_loss=0.2913, pruned_loss=0.09258, over 952319.64 frames. ], batch size: 29, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:14:27,727 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.054e+02 1.680e+02 2.024e+02 2.446e+02 4.498e+02, threshold=4.048e+02, percent-clipped=1.0 +2023-03-26 02:14:49,256 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16539.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:15:09,827 INFO [finetune.py:976] (4/7) Epoch 3, batch 5100, loss[loss=0.2472, simple_loss=0.2904, pruned_loss=0.102, over 4824.00 frames. ], tot_loss[loss=0.2321, simple_loss=0.2856, pruned_loss=0.08934, over 954127.15 frames. ], batch size: 38, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:15:16,429 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-26 02:15:36,572 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=16587.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:15:52,954 INFO [finetune.py:976] (4/7) Epoch 3, batch 5150, loss[loss=0.2339, simple_loss=0.281, pruned_loss=0.0934, over 4790.00 frames. ], tot_loss[loss=0.2351, simple_loss=0.2879, pruned_loss=0.09122, over 951529.69 frames. ], batch size: 29, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:16:12,054 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.182e+02 1.762e+02 2.113e+02 2.590e+02 4.768e+02, threshold=4.226e+02, percent-clipped=2.0 +2023-03-26 02:16:47,967 INFO [finetune.py:976] (4/7) Epoch 3, batch 5200, loss[loss=0.2192, simple_loss=0.2812, pruned_loss=0.07862, over 4771.00 frames. ], tot_loss[loss=0.2384, simple_loss=0.2922, pruned_loss=0.0923, over 951275.02 frames. ], batch size: 26, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:16:48,670 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16657.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:17:40,755 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=16705.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:17:41,307 INFO [finetune.py:976] (4/7) Epoch 3, batch 5250, loss[loss=0.2503, simple_loss=0.31, pruned_loss=0.09525, over 4889.00 frames. ], tot_loss[loss=0.2412, simple_loss=0.2953, pruned_loss=0.09355, over 951215.17 frames. ], batch size: 37, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:17:53,306 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.211e+02 1.833e+02 2.103e+02 2.647e+02 4.683e+02, threshold=4.205e+02, percent-clipped=1.0 +2023-03-26 02:18:00,563 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16727.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:18:03,072 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.26 vs. limit=5.0 +2023-03-26 02:18:03,600 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4991, 1.4551, 1.9042, 1.8087, 1.7196, 3.5625, 1.3098, 1.5357], + device='cuda:4'), covar=tensor([0.1063, 0.1725, 0.1090, 0.1028, 0.1561, 0.0212, 0.1486, 0.1762], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0081, 0.0078, 0.0079, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 02:18:12,264 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16741.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:18:13,946 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.58 vs. limit=2.0 +2023-03-26 02:18:21,891 INFO [finetune.py:976] (4/7) Epoch 3, batch 5300, loss[loss=0.2217, simple_loss=0.282, pruned_loss=0.08071, over 4886.00 frames. ], tot_loss[loss=0.2447, simple_loss=0.2989, pruned_loss=0.09525, over 954011.45 frames. ], batch size: 32, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:18:29,882 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 02:18:49,197 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=16789.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:19:07,701 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16805.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:19:08,225 INFO [finetune.py:976] (4/7) Epoch 3, batch 5350, loss[loss=0.1824, simple_loss=0.2463, pruned_loss=0.05928, over 4763.00 frames. ], tot_loss[loss=0.2446, simple_loss=0.2988, pruned_loss=0.09522, over 953684.18 frames. ], batch size: 28, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:19:16,760 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1409, 1.7568, 1.4182, 0.5835, 1.6613, 1.7459, 1.5424, 1.8082], + device='cuda:4'), covar=tensor([0.0919, 0.0885, 0.1752, 0.2421, 0.1328, 0.2679, 0.2371, 0.0918], + device='cuda:4'), in_proj_covar=tensor([0.0167, 0.0200, 0.0205, 0.0191, 0.0218, 0.0211, 0.0219, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 02:19:21,972 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.184e+02 1.880e+02 2.227e+02 2.511e+02 3.677e+02, threshold=4.454e+02, percent-clipped=0.0 +2023-03-26 02:19:28,211 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6767, 1.5791, 1.6300, 1.5760, 1.0996, 3.3360, 1.4168, 1.9678], + device='cuda:4'), covar=tensor([0.3204, 0.2226, 0.1975, 0.2239, 0.1954, 0.0186, 0.2705, 0.1307], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0112, 0.0116, 0.0120, 0.0116, 0.0097, 0.0101, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 02:19:46,296 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=16853.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:19:48,088 INFO [finetune.py:976] (4/7) Epoch 3, batch 5400, loss[loss=0.266, simple_loss=0.311, pruned_loss=0.1105, over 4811.00 frames. ], tot_loss[loss=0.2433, simple_loss=0.2969, pruned_loss=0.09483, over 954628.81 frames. ], batch size: 39, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:19:55,561 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16868.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 02:19:55,579 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4881, 1.2460, 1.2151, 1.4484, 1.7232, 1.4894, 0.8154, 1.2701], + device='cuda:4'), covar=tensor([0.2556, 0.2611, 0.2348, 0.2177, 0.1829, 0.1472, 0.3307, 0.2170], + device='cuda:4'), in_proj_covar=tensor([0.0231, 0.0209, 0.0197, 0.0182, 0.0233, 0.0173, 0.0213, 0.0187], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 02:20:02,710 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7814, 1.5495, 1.3306, 1.3370, 1.5285, 1.4727, 1.4468, 2.2970], + device='cuda:4'), covar=tensor([0.8746, 0.9019, 0.6955, 0.8843, 0.7355, 0.4805, 0.8300, 0.2876], + device='cuda:4'), in_proj_covar=tensor([0.0273, 0.0249, 0.0218, 0.0283, 0.0235, 0.0195, 0.0239, 0.0186], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 02:20:10,896 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 02:20:31,378 INFO [finetune.py:976] (4/7) Epoch 3, batch 5450, loss[loss=0.1895, simple_loss=0.2529, pruned_loss=0.06307, over 4773.00 frames. ], tot_loss[loss=0.2389, simple_loss=0.2924, pruned_loss=0.09271, over 956540.11 frames. ], batch size: 28, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:20:31,473 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16906.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:20:32,652 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9092, 1.7800, 1.7544, 1.8566, 1.3002, 3.8131, 1.6573, 2.2480], + device='cuda:4'), covar=tensor([0.3397, 0.2380, 0.2044, 0.2168, 0.1973, 0.0180, 0.2348, 0.1300], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0111, 0.0116, 0.0119, 0.0116, 0.0097, 0.0101, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 02:20:38,655 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.665e+02 2.000e+02 2.450e+02 4.433e+02, threshold=4.000e+02, percent-clipped=0.0 +2023-03-26 02:20:46,032 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16929.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 02:21:14,236 INFO [finetune.py:976] (4/7) Epoch 3, batch 5500, loss[loss=0.2295, simple_loss=0.2868, pruned_loss=0.08608, over 4739.00 frames. ], tot_loss[loss=0.2343, simple_loss=0.2879, pruned_loss=0.09035, over 955476.13 frames. ], batch size: 54, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:21:24,958 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0727, 1.2228, 0.8883, 1.8392, 2.2422, 1.7469, 1.5132, 2.0189], + device='cuda:4'), covar=tensor([0.1397, 0.2085, 0.2395, 0.1179, 0.1956, 0.2261, 0.1421, 0.1714], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0099, 0.0118, 0.0094, 0.0125, 0.0098, 0.0101, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 02:21:26,182 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16967.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 02:21:43,872 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8834, 1.6916, 1.4911, 1.7742, 1.9137, 1.5943, 2.2511, 1.8552], + device='cuda:4'), covar=tensor([0.2081, 0.4046, 0.4831, 0.4046, 0.3212, 0.2309, 0.4052, 0.3019], + device='cuda:4'), in_proj_covar=tensor([0.0165, 0.0195, 0.0239, 0.0254, 0.0221, 0.0186, 0.0209, 0.0189], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 02:22:07,781 INFO [finetune.py:976] (4/7) Epoch 3, batch 5550, loss[loss=0.3266, simple_loss=0.3595, pruned_loss=0.1469, over 4221.00 frames. ], tot_loss[loss=0.2372, simple_loss=0.2898, pruned_loss=0.0923, over 953341.81 frames. ], batch size: 65, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:22:13,273 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2608, 1.6225, 1.6262, 0.9580, 1.7912, 2.0005, 1.7494, 1.6106], + device='cuda:4'), covar=tensor([0.1093, 0.0991, 0.0666, 0.0887, 0.0759, 0.0542, 0.0594, 0.0692], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0157, 0.0117, 0.0137, 0.0132, 0.0121, 0.0146, 0.0144], + device='cuda:4'), out_proj_covar=tensor([9.7934e-05, 1.1711e-04, 8.5342e-05, 1.0049e-04, 9.5880e-05, 8.9726e-05, + 1.0894e-04, 1.0696e-04], device='cuda:4') +2023-03-26 02:22:15,686 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.144e+02 1.712e+02 2.015e+02 2.380e+02 4.122e+02, threshold=4.030e+02, percent-clipped=1.0 +2023-03-26 02:22:21,780 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17027.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:22:22,691 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 02:22:41,563 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.50 vs. limit=5.0 +2023-03-26 02:22:49,844 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-26 02:22:52,958 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.10 vs. limit=5.0 +2023-03-26 02:22:53,359 INFO [finetune.py:976] (4/7) Epoch 3, batch 5600, loss[loss=0.2458, simple_loss=0.2926, pruned_loss=0.09946, over 4766.00 frames. ], tot_loss[loss=0.2395, simple_loss=0.2928, pruned_loss=0.09313, over 949297.91 frames. ], batch size: 26, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:23:10,687 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=17075.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:23:36,458 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-26 02:23:38,819 INFO [finetune.py:976] (4/7) Epoch 3, batch 5650, loss[loss=0.2426, simple_loss=0.3107, pruned_loss=0.08727, over 4928.00 frames. ], tot_loss[loss=0.2424, simple_loss=0.2964, pruned_loss=0.09415, over 950908.88 frames. ], batch size: 38, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:23:45,804 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.238e+02 1.755e+02 2.152e+02 2.681e+02 4.789e+02, threshold=4.305e+02, percent-clipped=1.0 +2023-03-26 02:24:07,225 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3098, 2.1217, 1.8336, 2.3134, 2.3286, 1.9511, 2.5852, 2.2347], + device='cuda:4'), covar=tensor([0.1752, 0.3497, 0.4562, 0.3544, 0.3033, 0.2067, 0.3357, 0.2584], + device='cuda:4'), in_proj_covar=tensor([0.0164, 0.0195, 0.0238, 0.0254, 0.0221, 0.0186, 0.0209, 0.0189], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 02:24:15,375 INFO [finetune.py:976] (4/7) Epoch 3, batch 5700, loss[loss=0.2216, simple_loss=0.2604, pruned_loss=0.09136, over 3962.00 frames. ], tot_loss[loss=0.24, simple_loss=0.2922, pruned_loss=0.09385, over 932671.27 frames. ], batch size: 17, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:24:56,865 INFO [finetune.py:976] (4/7) Epoch 4, batch 0, loss[loss=0.2951, simple_loss=0.3395, pruned_loss=0.1254, over 4812.00 frames. ], tot_loss[loss=0.2951, simple_loss=0.3395, pruned_loss=0.1254, over 4812.00 frames. ], batch size: 39, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:24:56,865 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 02:25:05,676 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.2291, 1.3603, 1.3736, 0.7960, 1.2434, 1.5636, 1.6002, 1.3469], + device='cuda:4'), covar=tensor([0.1108, 0.0827, 0.0556, 0.0584, 0.0501, 0.0566, 0.0366, 0.0737], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0157, 0.0117, 0.0136, 0.0132, 0.0121, 0.0146, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.8181e-05, 1.1719e-04, 8.5409e-05, 1.0017e-04, 9.5355e-05, 9.0035e-05, + 1.0877e-04, 1.0736e-04], device='cuda:4') +2023-03-26 02:25:18,215 INFO [finetune.py:1010] (4/7) Epoch 4, validation: loss=0.1768, simple_loss=0.2473, pruned_loss=0.0532, over 2265189.00 frames. +2023-03-26 02:25:18,215 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6301MB +2023-03-26 02:25:22,712 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17189.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:25:30,434 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-26 02:25:55,806 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.188e+02 1.713e+02 2.128e+02 2.708e+02 4.853e+02, threshold=4.257e+02, percent-clipped=3.0 +2023-03-26 02:26:00,042 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17224.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 02:26:05,894 INFO [finetune.py:976] (4/7) Epoch 4, batch 50, loss[loss=0.246, simple_loss=0.2831, pruned_loss=0.1045, over 4721.00 frames. ], tot_loss[loss=0.2447, simple_loss=0.2985, pruned_loss=0.09545, over 216728.71 frames. ], batch size: 59, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:26:06,643 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7690, 1.6689, 1.6410, 1.6538, 1.0822, 3.4147, 1.4635, 2.0417], + device='cuda:4'), covar=tensor([0.3121, 0.2171, 0.1941, 0.2112, 0.1859, 0.0190, 0.2758, 0.1218], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0111, 0.0116, 0.0120, 0.0116, 0.0097, 0.0101, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 02:26:29,261 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17250.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:26:36,474 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17262.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 02:26:36,518 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3716, 2.0659, 2.3570, 1.3059, 2.5297, 2.7828, 2.2020, 2.2493], + device='cuda:4'), covar=tensor([0.0876, 0.0709, 0.0411, 0.0643, 0.0485, 0.0399, 0.0377, 0.0544], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0157, 0.0117, 0.0137, 0.0132, 0.0121, 0.0146, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.8010e-05, 1.1719e-04, 8.5152e-05, 1.0037e-04, 9.5639e-05, 8.9861e-05, + 1.0876e-04, 1.0725e-04], device='cuda:4') +2023-03-26 02:26:55,371 INFO [finetune.py:976] (4/7) Epoch 4, batch 100, loss[loss=0.2037, simple_loss=0.2605, pruned_loss=0.07344, over 4758.00 frames. ], tot_loss[loss=0.2317, simple_loss=0.2857, pruned_loss=0.0889, over 380128.39 frames. ], batch size: 54, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:27:26,863 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.228e+02 1.697e+02 1.982e+02 2.273e+02 3.827e+02, threshold=3.964e+02, percent-clipped=0.0 +2023-03-26 02:27:30,723 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6213, 1.4352, 1.7883, 2.9094, 2.0047, 2.2837, 1.1894, 2.2955], + device='cuda:4'), covar=tensor([0.1889, 0.1618, 0.1457, 0.0645, 0.0956, 0.1184, 0.1857, 0.0708], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0119, 0.0137, 0.0165, 0.0103, 0.0143, 0.0130, 0.0106], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 02:27:36,642 INFO [finetune.py:976] (4/7) Epoch 4, batch 150, loss[loss=0.2238, simple_loss=0.278, pruned_loss=0.08481, over 4820.00 frames. ], tot_loss[loss=0.2321, simple_loss=0.2845, pruned_loss=0.08991, over 508053.41 frames. ], batch size: 38, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:27:46,580 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8063, 1.5562, 1.3007, 1.3327, 1.5296, 1.5303, 1.4414, 2.3409], + device='cuda:4'), covar=tensor([0.9322, 0.8664, 0.7173, 0.8886, 0.7350, 0.4950, 0.8689, 0.3001], + device='cuda:4'), in_proj_covar=tensor([0.0275, 0.0251, 0.0219, 0.0284, 0.0235, 0.0196, 0.0240, 0.0187], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 02:28:01,478 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.28 vs. limit=5.0 +2023-03-26 02:28:01,916 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17362.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:28:25,904 INFO [finetune.py:976] (4/7) Epoch 4, batch 200, loss[loss=0.1807, simple_loss=0.2463, pruned_loss=0.0575, over 4893.00 frames. ], tot_loss[loss=0.2326, simple_loss=0.2841, pruned_loss=0.09053, over 604507.44 frames. ], batch size: 35, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:28:26,036 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17383.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 02:28:55,774 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.314e+02 1.771e+02 2.098e+02 2.514e+02 4.657e+02, threshold=4.195e+02, percent-clipped=1.0 +2023-03-26 02:29:01,233 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17423.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:29:09,066 INFO [finetune.py:976] (4/7) Epoch 4, batch 250, loss[loss=0.2778, simple_loss=0.3332, pruned_loss=0.1112, over 4803.00 frames. ], tot_loss[loss=0.2367, simple_loss=0.2891, pruned_loss=0.09212, over 681238.28 frames. ], batch size: 41, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:29:17,850 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17444.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 02:29:26,239 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1516, 1.6831, 2.5684, 3.9036, 2.8345, 2.6815, 0.8650, 3.0900], + device='cuda:4'), covar=tensor([0.1765, 0.1538, 0.1332, 0.0501, 0.0748, 0.1517, 0.2171, 0.0573], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0118, 0.0137, 0.0165, 0.0103, 0.0142, 0.0129, 0.0105], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 02:29:37,664 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8495, 1.8512, 1.8130, 1.2544, 2.0349, 1.9997, 1.9098, 1.6226], + device='cuda:4'), covar=tensor([0.0759, 0.0697, 0.0865, 0.1054, 0.0535, 0.0739, 0.0758, 0.1187], + device='cuda:4'), in_proj_covar=tensor([0.0139, 0.0133, 0.0144, 0.0129, 0.0111, 0.0142, 0.0147, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 02:29:49,004 INFO [finetune.py:976] (4/7) Epoch 4, batch 300, loss[loss=0.259, simple_loss=0.3117, pruned_loss=0.1031, over 4820.00 frames. ], tot_loss[loss=0.2413, simple_loss=0.2945, pruned_loss=0.09409, over 742824.39 frames. ], batch size: 49, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:30:35,038 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.302e+02 1.964e+02 2.271e+02 2.699e+02 6.272e+02, threshold=4.542e+02, percent-clipped=2.0 +2023-03-26 02:30:39,305 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17524.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 02:30:41,120 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6022, 1.5199, 1.7499, 1.7752, 1.6782, 3.5019, 1.3323, 1.6191], + device='cuda:4'), covar=tensor([0.1029, 0.1820, 0.1192, 0.1063, 0.1704, 0.0225, 0.1591, 0.1847], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0082, 0.0078, 0.0080, 0.0093, 0.0083, 0.0086, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 02:30:44,604 INFO [finetune.py:976] (4/7) Epoch 4, batch 350, loss[loss=0.2584, simple_loss=0.3176, pruned_loss=0.09954, over 4809.00 frames. ], tot_loss[loss=0.2451, simple_loss=0.2982, pruned_loss=0.096, over 792011.52 frames. ], batch size: 38, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:30:47,077 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7479, 1.6149, 1.2599, 1.4527, 1.5154, 1.4829, 1.4713, 2.2473], + device='cuda:4'), covar=tensor([0.9235, 0.8759, 0.6968, 0.9317, 0.7584, 0.4632, 0.8619, 0.3131], + device='cuda:4'), in_proj_covar=tensor([0.0274, 0.0250, 0.0218, 0.0283, 0.0235, 0.0196, 0.0239, 0.0187], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 02:30:51,476 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4658, 1.2722, 1.2214, 1.4709, 1.7681, 1.4498, 1.0113, 1.2542], + device='cuda:4'), covar=tensor([0.2479, 0.2484, 0.2068, 0.1889, 0.1679, 0.1320, 0.2872, 0.1965], + device='cuda:4'), in_proj_covar=tensor([0.0232, 0.0208, 0.0198, 0.0183, 0.0232, 0.0173, 0.0213, 0.0186], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 02:30:53,138 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17545.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:31:09,666 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5507, 1.3502, 1.3542, 1.6306, 2.0467, 1.6761, 1.2821, 1.2547], + device='cuda:4'), covar=tensor([0.2864, 0.2918, 0.2474, 0.2189, 0.2452, 0.1640, 0.3572, 0.2445], + device='cuda:4'), in_proj_covar=tensor([0.0232, 0.0208, 0.0198, 0.0183, 0.0232, 0.0173, 0.0213, 0.0187], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 02:31:10,223 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17562.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 02:31:16,314 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=17572.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 02:31:17,561 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3846, 2.0153, 1.6237, 0.7562, 1.8100, 1.8589, 1.6964, 1.9429], + device='cuda:4'), covar=tensor([0.0951, 0.1004, 0.1814, 0.2795, 0.1644, 0.2675, 0.2493, 0.0988], + device='cuda:4'), in_proj_covar=tensor([0.0166, 0.0200, 0.0204, 0.0190, 0.0217, 0.0210, 0.0218, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 02:31:23,350 INFO [finetune.py:976] (4/7) Epoch 4, batch 400, loss[loss=0.2105, simple_loss=0.276, pruned_loss=0.07249, over 4740.00 frames. ], tot_loss[loss=0.2443, simple_loss=0.2983, pruned_loss=0.09514, over 829504.12 frames. ], batch size: 54, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:31:46,313 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=17610.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:31:51,079 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.188e+02 1.790e+02 1.987e+02 2.567e+02 5.687e+02, threshold=3.975e+02, percent-clipped=1.0 +2023-03-26 02:32:09,440 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-26 02:32:10,369 INFO [finetune.py:976] (4/7) Epoch 4, batch 450, loss[loss=0.2303, simple_loss=0.2867, pruned_loss=0.08693, over 4910.00 frames. ], tot_loss[loss=0.2412, simple_loss=0.2948, pruned_loss=0.0938, over 853986.04 frames. ], batch size: 43, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:32:39,410 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0609, 1.9338, 1.5992, 2.0036, 1.9014, 1.7848, 1.8338, 2.9294], + device='cuda:4'), covar=tensor([1.0044, 1.1332, 0.7772, 1.0562, 0.9448, 0.5843, 1.0499, 0.3045], + device='cuda:4'), in_proj_covar=tensor([0.0275, 0.0250, 0.0218, 0.0282, 0.0235, 0.0196, 0.0239, 0.0187], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 02:33:00,231 INFO [finetune.py:976] (4/7) Epoch 4, batch 500, loss[loss=0.2325, simple_loss=0.2849, pruned_loss=0.09004, over 4823.00 frames. ], tot_loss[loss=0.2383, simple_loss=0.2917, pruned_loss=0.0925, over 877772.26 frames. ], batch size: 39, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:33:11,506 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17700.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:33:14,326 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6228, 1.5420, 1.5381, 1.6117, 1.0252, 3.2853, 1.2758, 1.7363], + device='cuda:4'), covar=tensor([0.3390, 0.2405, 0.1996, 0.2182, 0.2092, 0.0200, 0.2704, 0.1405], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0111, 0.0116, 0.0119, 0.0116, 0.0097, 0.0101, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 02:33:24,347 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.264e+02 1.808e+02 2.091e+02 2.485e+02 4.480e+02, threshold=4.181e+02, percent-clipped=1.0 +2023-03-26 02:33:24,434 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17718.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:33:31,863 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17730.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:33:34,061 INFO [finetune.py:976] (4/7) Epoch 4, batch 550, loss[loss=0.2184, simple_loss=0.2774, pruned_loss=0.07967, over 4838.00 frames. ], tot_loss[loss=0.2368, simple_loss=0.2895, pruned_loss=0.09202, over 895471.90 frames. ], batch size: 30, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:33:37,769 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17739.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 02:33:52,059 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3335, 2.8805, 2.8108, 1.2278, 3.0465, 2.1872, 0.7016, 1.8268], + device='cuda:4'), covar=tensor([0.2479, 0.2470, 0.1925, 0.3564, 0.1429, 0.1237, 0.4206, 0.1828], + device='cuda:4'), in_proj_covar=tensor([0.0157, 0.0170, 0.0165, 0.0130, 0.0156, 0.0122, 0.0147, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 02:34:03,931 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17761.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:34:17,628 INFO [finetune.py:976] (4/7) Epoch 4, batch 600, loss[loss=0.3134, simple_loss=0.3559, pruned_loss=0.1354, over 4746.00 frames. ], tot_loss[loss=0.2373, simple_loss=0.2898, pruned_loss=0.09245, over 907860.86 frames. ], batch size: 54, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:34:22,580 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17791.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:34:41,399 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.036e+02 1.729e+02 2.101e+02 2.480e+02 7.519e+02, threshold=4.202e+02, percent-clipped=3.0 +2023-03-26 02:34:47,578 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5852, 1.4976, 1.3192, 1.2127, 1.9066, 2.0330, 1.7667, 1.3398], + device='cuda:4'), covar=tensor([0.0337, 0.0443, 0.0666, 0.0484, 0.0254, 0.0392, 0.0286, 0.0515], + device='cuda:4'), in_proj_covar=tensor([0.0085, 0.0115, 0.0138, 0.0118, 0.0105, 0.0100, 0.0091, 0.0110], + device='cuda:4'), out_proj_covar=tensor([6.6272e-05, 9.0534e-05, 1.1125e-04, 9.2892e-05, 8.3515e-05, 7.4240e-05, + 7.0491e-05, 8.5760e-05], device='cuda:4') +2023-03-26 02:34:50,417 INFO [finetune.py:976] (4/7) Epoch 4, batch 650, loss[loss=0.24, simple_loss=0.2937, pruned_loss=0.09319, over 4902.00 frames. ], tot_loss[loss=0.2419, simple_loss=0.2948, pruned_loss=0.09451, over 920240.53 frames. ], batch size: 36, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:34:56,263 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6785, 1.5382, 2.0282, 1.8775, 2.0408, 4.1284, 1.4350, 1.9361], + device='cuda:4'), covar=tensor([0.1025, 0.1788, 0.1327, 0.1106, 0.1411, 0.0201, 0.1532, 0.1660], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0081, 0.0078, 0.0080, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 02:34:59,969 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17845.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:35:20,730 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6813, 1.5530, 1.2246, 1.3084, 1.4769, 1.4519, 1.3746, 2.1662], + device='cuda:4'), covar=tensor([0.8818, 0.8430, 0.6925, 0.8864, 0.7359, 0.4719, 0.8914, 0.2943], + device='cuda:4'), in_proj_covar=tensor([0.0274, 0.0249, 0.0218, 0.0283, 0.0235, 0.0195, 0.0239, 0.0187], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 02:35:21,371 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.62 vs. limit=2.0 +2023-03-26 02:35:31,913 INFO [finetune.py:976] (4/7) Epoch 4, batch 700, loss[loss=0.2351, simple_loss=0.3012, pruned_loss=0.08456, over 4883.00 frames. ], tot_loss[loss=0.2424, simple_loss=0.2961, pruned_loss=0.09436, over 928389.31 frames. ], batch size: 32, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:35:46,369 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=17893.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:36:18,034 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.240e+02 1.804e+02 2.027e+02 2.461e+02 4.855e+02, threshold=4.055e+02, percent-clipped=2.0 +2023-03-26 02:36:34,753 INFO [finetune.py:976] (4/7) Epoch 4, batch 750, loss[loss=0.2781, simple_loss=0.321, pruned_loss=0.1176, over 4899.00 frames. ], tot_loss[loss=0.2417, simple_loss=0.2963, pruned_loss=0.09357, over 935757.92 frames. ], batch size: 46, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:36:58,573 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9221, 1.6369, 1.5128, 1.5725, 1.6430, 1.6034, 1.5703, 2.3240], + device='cuda:4'), covar=tensor([0.8107, 0.8467, 0.6096, 0.8343, 0.7031, 0.4449, 0.8502, 0.2658], + device='cuda:4'), in_proj_covar=tensor([0.0274, 0.0249, 0.0218, 0.0283, 0.0234, 0.0196, 0.0239, 0.0187], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 02:37:08,575 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17960.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:37:30,390 INFO [finetune.py:976] (4/7) Epoch 4, batch 800, loss[loss=0.2575, simple_loss=0.2833, pruned_loss=0.1159, over 3957.00 frames. ], tot_loss[loss=0.2406, simple_loss=0.2955, pruned_loss=0.09282, over 940112.20 frames. ], batch size: 17, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:38:05,564 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.155e+02 1.784e+02 2.191e+02 2.808e+02 5.190e+02, threshold=4.382e+02, percent-clipped=3.0 +2023-03-26 02:38:05,661 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18018.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:38:08,536 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18021.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:38:20,834 INFO [finetune.py:976] (4/7) Epoch 4, batch 850, loss[loss=0.2298, simple_loss=0.2825, pruned_loss=0.08856, over 4864.00 frames. ], tot_loss[loss=0.2376, simple_loss=0.2924, pruned_loss=0.09136, over 944751.79 frames. ], batch size: 34, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:38:23,811 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 02:38:26,575 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18039.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 02:38:33,270 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.8085, 1.8601, 2.0383, 1.4387, 1.7777, 2.1773, 2.0723, 1.8127], + device='cuda:4'), covar=tensor([0.0995, 0.0727, 0.0418, 0.0569, 0.0509, 0.0541, 0.0365, 0.0558], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0159, 0.0118, 0.0138, 0.0134, 0.0123, 0.0147, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.8557e-05, 1.1816e-04, 8.5457e-05, 1.0136e-04, 9.7346e-05, 9.1182e-05, + 1.0997e-04, 1.0781e-04], device='cuda:4') +2023-03-26 02:38:38,032 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0279, 1.7526, 2.2797, 1.4782, 1.9587, 2.1943, 1.6998, 2.4243], + device='cuda:4'), covar=tensor([0.1601, 0.2322, 0.1710, 0.2412, 0.1136, 0.1694, 0.2703, 0.1064], + device='cuda:4'), in_proj_covar=tensor([0.0207, 0.0207, 0.0204, 0.0199, 0.0183, 0.0228, 0.0217, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 02:38:39,174 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18056.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:38:45,705 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18066.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:38:57,816 INFO [finetune.py:976] (4/7) Epoch 4, batch 900, loss[loss=0.2278, simple_loss=0.2735, pruned_loss=0.09103, over 4840.00 frames. ], tot_loss[loss=0.2335, simple_loss=0.288, pruned_loss=0.08946, over 946952.41 frames. ], batch size: 33, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:38:59,685 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18086.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:39:00,271 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18087.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 02:39:19,670 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.218e+02 1.720e+02 1.921e+02 2.312e+02 4.297e+02, threshold=3.842e+02, percent-clipped=0.0 +2023-03-26 02:39:35,995 INFO [finetune.py:976] (4/7) Epoch 4, batch 950, loss[loss=0.2818, simple_loss=0.3362, pruned_loss=0.1137, over 4814.00 frames. ], tot_loss[loss=0.234, simple_loss=0.288, pruned_loss=0.08998, over 949482.87 frames. ], batch size: 41, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:39:46,536 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-26 02:40:16,964 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6388, 0.8528, 1.4646, 1.3514, 1.3055, 1.2800, 1.2000, 1.3856], + device='cuda:4'), covar=tensor([0.7249, 1.0745, 0.8877, 0.9559, 1.0333, 0.7813, 1.2344, 0.7729], + device='cuda:4'), in_proj_covar=tensor([0.0227, 0.0252, 0.0254, 0.0263, 0.0241, 0.0216, 0.0277, 0.0219], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 02:40:21,705 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7975, 1.8115, 1.9738, 1.2880, 2.0883, 2.1152, 1.9811, 1.4646], + device='cuda:4'), covar=tensor([0.0797, 0.0803, 0.0822, 0.1144, 0.0604, 0.0837, 0.0877, 0.1679], + device='cuda:4'), in_proj_covar=tensor([0.0138, 0.0131, 0.0143, 0.0128, 0.0109, 0.0141, 0.0146, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 02:40:22,966 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18174.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:40:24,042 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18175.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:40:29,234 INFO [finetune.py:976] (4/7) Epoch 4, batch 1000, loss[loss=0.2591, simple_loss=0.3101, pruned_loss=0.1041, over 4816.00 frames. ], tot_loss[loss=0.2379, simple_loss=0.2924, pruned_loss=0.09173, over 951639.78 frames. ], batch size: 38, lr: 3.97e-03, grad_scale: 64.0 +2023-03-26 02:41:07,308 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.077e+02 1.690e+02 2.099e+02 2.471e+02 3.966e+02, threshold=4.198e+02, percent-clipped=1.0 +2023-03-26 02:41:14,727 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0877, 0.9500, 0.9996, 0.3363, 0.7023, 1.1392, 1.2355, 1.0419], + device='cuda:4'), covar=tensor([0.0962, 0.0554, 0.0505, 0.0648, 0.0541, 0.0476, 0.0327, 0.0537], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0159, 0.0117, 0.0138, 0.0134, 0.0123, 0.0147, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.8591e-05, 1.1804e-04, 8.5311e-05, 1.0151e-04, 9.6769e-05, 9.0953e-05, + 1.0995e-04, 1.0752e-04], device='cuda:4') +2023-03-26 02:41:28,638 INFO [finetune.py:976] (4/7) Epoch 4, batch 1050, loss[loss=0.1974, simple_loss=0.2382, pruned_loss=0.07824, over 4236.00 frames. ], tot_loss[loss=0.2375, simple_loss=0.2929, pruned_loss=0.09104, over 949921.07 frames. ], batch size: 18, lr: 3.97e-03, grad_scale: 64.0 +2023-03-26 02:41:29,968 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18235.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:41:30,594 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18236.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:41:33,076 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9338, 1.7911, 1.4411, 1.7384, 1.7202, 1.6248, 1.6354, 2.5449], + device='cuda:4'), covar=tensor([0.9400, 1.0033, 0.7539, 1.0471, 0.8299, 0.5528, 0.9497, 0.3010], + device='cuda:4'), in_proj_covar=tensor([0.0274, 0.0250, 0.0218, 0.0282, 0.0234, 0.0196, 0.0238, 0.0187], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 02:42:18,624 INFO [finetune.py:976] (4/7) Epoch 4, batch 1100, loss[loss=0.2486, simple_loss=0.3019, pruned_loss=0.09769, over 4878.00 frames. ], tot_loss[loss=0.2387, simple_loss=0.294, pruned_loss=0.09173, over 950802.31 frames. ], batch size: 32, lr: 3.97e-03, grad_scale: 64.0 +2023-03-26 02:42:39,981 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9405, 1.8299, 1.6740, 1.8957, 1.3922, 4.5366, 1.7226, 2.3040], + device='cuda:4'), covar=tensor([0.3322, 0.2256, 0.2071, 0.2132, 0.1874, 0.0108, 0.2341, 0.1264], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0111, 0.0116, 0.0119, 0.0115, 0.0096, 0.0100, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 02:42:49,579 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18316.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:42:50,750 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.170e+02 1.851e+02 2.220e+02 2.754e+02 4.687e+02, threshold=4.440e+02, percent-clipped=1.0 +2023-03-26 02:43:08,652 INFO [finetune.py:976] (4/7) Epoch 4, batch 1150, loss[loss=0.2185, simple_loss=0.2768, pruned_loss=0.0801, over 4899.00 frames. ], tot_loss[loss=0.2387, simple_loss=0.2945, pruned_loss=0.09149, over 952339.05 frames. ], batch size: 36, lr: 3.97e-03, grad_scale: 64.0 +2023-03-26 02:43:14,713 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-03-26 02:43:24,241 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 02:43:25,362 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18356.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:44:03,368 INFO [finetune.py:976] (4/7) Epoch 4, batch 1200, loss[loss=0.2039, simple_loss=0.2682, pruned_loss=0.0698, over 4787.00 frames. ], tot_loss[loss=0.2366, simple_loss=0.292, pruned_loss=0.09064, over 952259.83 frames. ], batch size: 29, lr: 3.97e-03, grad_scale: 64.0 +2023-03-26 02:44:05,826 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18386.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:44:28,301 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18404.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:44:47,474 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.754e+02 2.082e+02 2.492e+02 3.668e+02, threshold=4.164e+02, percent-clipped=0.0 +2023-03-26 02:45:07,145 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-03-26 02:45:07,910 INFO [finetune.py:976] (4/7) Epoch 4, batch 1250, loss[loss=0.2404, simple_loss=0.2974, pruned_loss=0.09168, over 4913.00 frames. ], tot_loss[loss=0.234, simple_loss=0.2892, pruned_loss=0.08941, over 952699.87 frames. ], batch size: 36, lr: 3.97e-03, grad_scale: 64.0 +2023-03-26 02:45:09,084 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18434.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:45:19,468 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-26 02:45:27,387 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8379, 1.9205, 1.6303, 1.3594, 2.0783, 2.2366, 2.1068, 1.8391], + device='cuda:4'), covar=tensor([0.0416, 0.0409, 0.0600, 0.0481, 0.0396, 0.0530, 0.0347, 0.0477], + device='cuda:4'), in_proj_covar=tensor([0.0084, 0.0113, 0.0138, 0.0117, 0.0104, 0.0099, 0.0091, 0.0109], + device='cuda:4'), out_proj_covar=tensor([6.5697e-05, 8.9649e-05, 1.1115e-04, 9.2050e-05, 8.2700e-05, 7.3874e-05, + 7.0370e-05, 8.5705e-05], device='cuda:4') +2023-03-26 02:45:45,612 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.66 vs. limit=5.0 +2023-03-26 02:45:47,098 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0824, 2.2712, 2.1846, 1.4238, 2.5030, 2.4885, 2.1979, 1.9839], + device='cuda:4'), covar=tensor([0.0743, 0.0592, 0.0823, 0.1032, 0.0396, 0.0652, 0.0777, 0.0976], + device='cuda:4'), in_proj_covar=tensor([0.0138, 0.0132, 0.0143, 0.0128, 0.0110, 0.0141, 0.0146, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 02:45:59,526 INFO [finetune.py:976] (4/7) Epoch 4, batch 1300, loss[loss=0.2568, simple_loss=0.3058, pruned_loss=0.104, over 4826.00 frames. ], tot_loss[loss=0.2316, simple_loss=0.286, pruned_loss=0.08861, over 953829.06 frames. ], batch size: 33, lr: 3.97e-03, grad_scale: 64.0 +2023-03-26 02:46:10,682 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18490.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:46:22,651 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18508.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 02:46:29,711 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.658e+02 2.008e+02 2.632e+02 4.281e+02, threshold=4.017e+02, percent-clipped=1.0 +2023-03-26 02:46:36,933 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18530.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:46:37,498 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18531.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:46:41,948 INFO [finetune.py:976] (4/7) Epoch 4, batch 1350, loss[loss=0.1966, simple_loss=0.2527, pruned_loss=0.07024, over 4833.00 frames. ], tot_loss[loss=0.2328, simple_loss=0.2869, pruned_loss=0.08939, over 953873.16 frames. ], batch size: 30, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:46:55,232 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18551.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:47:11,808 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18569.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 02:47:24,184 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 02:47:25,815 INFO [finetune.py:976] (4/7) Epoch 4, batch 1400, loss[loss=0.2272, simple_loss=0.2922, pruned_loss=0.08116, over 4908.00 frames. ], tot_loss[loss=0.2367, simple_loss=0.2913, pruned_loss=0.09107, over 955724.35 frames. ], batch size: 36, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:47:25,914 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6581, 1.5042, 1.5692, 1.5302, 0.9437, 3.6662, 1.2815, 1.9144], + device='cuda:4'), covar=tensor([0.3294, 0.2416, 0.1980, 0.2207, 0.2105, 0.0159, 0.2644, 0.1312], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0112, 0.0116, 0.0119, 0.0116, 0.0096, 0.0100, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 02:47:51,395 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8852, 1.8192, 1.4923, 1.8626, 1.9421, 1.6013, 2.3758, 1.9098], + device='cuda:4'), covar=tensor([0.2069, 0.4151, 0.4732, 0.4238, 0.3298, 0.2315, 0.4234, 0.2700], + device='cuda:4'), in_proj_covar=tensor([0.0163, 0.0193, 0.0237, 0.0253, 0.0221, 0.0185, 0.0208, 0.0187], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 02:47:53,784 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18616.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:47:55,464 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.253e+02 1.863e+02 2.207e+02 2.712e+02 5.337e+02, threshold=4.415e+02, percent-clipped=2.0 +2023-03-26 02:47:55,684 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-26 02:48:10,033 INFO [finetune.py:976] (4/7) Epoch 4, batch 1450, loss[loss=0.2168, simple_loss=0.28, pruned_loss=0.07685, over 4822.00 frames. ], tot_loss[loss=0.238, simple_loss=0.2931, pruned_loss=0.09143, over 957587.39 frames. ], batch size: 33, lr: 3.97e-03, grad_scale: 32.0 +2023-03-26 02:48:13,834 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-26 02:48:42,204 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2475, 2.1041, 2.5106, 1.0454, 2.5684, 2.8271, 2.3425, 2.2556], + device='cuda:4'), covar=tensor([0.1041, 0.0692, 0.0416, 0.0784, 0.0640, 0.0432, 0.0399, 0.0540], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0157, 0.0117, 0.0137, 0.0132, 0.0121, 0.0146, 0.0144], + device='cuda:4'), out_proj_covar=tensor([9.7599e-05, 1.1684e-04, 8.4878e-05, 1.0078e-04, 9.5852e-05, 8.9531e-05, + 1.0850e-04, 1.0666e-04], device='cuda:4') +2023-03-26 02:48:43,323 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18664.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:48:49,443 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18674.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:48:56,988 INFO [finetune.py:976] (4/7) Epoch 4, batch 1500, loss[loss=0.3014, simple_loss=0.3372, pruned_loss=0.1328, over 4902.00 frames. ], tot_loss[loss=0.2388, simple_loss=0.2934, pruned_loss=0.09208, over 956117.49 frames. ], batch size: 37, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:49:36,303 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.377e+02 1.810e+02 2.122e+02 2.509e+02 6.153e+02, threshold=4.245e+02, percent-clipped=1.0 +2023-03-26 02:49:54,648 INFO [finetune.py:976] (4/7) Epoch 4, batch 1550, loss[loss=0.2581, simple_loss=0.3025, pruned_loss=0.1068, over 4878.00 frames. ], tot_loss[loss=0.2396, simple_loss=0.2941, pruned_loss=0.09251, over 955402.68 frames. ], batch size: 35, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:49:55,938 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.0348, 4.3830, 4.5838, 4.8691, 4.6946, 4.4800, 5.1358, 1.4333], + device='cuda:4'), covar=tensor([0.0689, 0.0704, 0.0664, 0.0730, 0.1286, 0.1371, 0.0574, 0.5244], + device='cuda:4'), in_proj_covar=tensor([0.0359, 0.0243, 0.0276, 0.0291, 0.0337, 0.0284, 0.0307, 0.0299], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 02:49:55,977 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18735.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:50:31,085 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 02:50:39,058 INFO [finetune.py:976] (4/7) Epoch 4, batch 1600, loss[loss=0.2518, simple_loss=0.2901, pruned_loss=0.1068, over 4154.00 frames. ], tot_loss[loss=0.2375, simple_loss=0.2918, pruned_loss=0.09163, over 956950.87 frames. ], batch size: 65, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:51:19,526 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.645e+02 2.011e+02 2.399e+02 5.772e+02, threshold=4.021e+02, percent-clipped=1.0 +2023-03-26 02:51:30,350 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18830.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:51:30,963 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18831.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:51:32,083 INFO [finetune.py:976] (4/7) Epoch 4, batch 1650, loss[loss=0.2068, simple_loss=0.2643, pruned_loss=0.07465, over 4793.00 frames. ], tot_loss[loss=0.2335, simple_loss=0.2877, pruned_loss=0.08967, over 956992.79 frames. ], batch size: 29, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:51:48,111 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18846.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:52:05,189 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18864.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 02:52:13,737 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0174, 1.8796, 1.8336, 2.0845, 1.3435, 4.5831, 1.7857, 2.4025], + device='cuda:4'), covar=tensor([0.3501, 0.2396, 0.1968, 0.2291, 0.1850, 0.0095, 0.2503, 0.1293], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0112, 0.0117, 0.0120, 0.0117, 0.0097, 0.0101, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 02:52:23,040 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18878.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:52:23,634 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=18879.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:52:25,971 INFO [finetune.py:976] (4/7) Epoch 4, batch 1700, loss[loss=0.2043, simple_loss=0.263, pruned_loss=0.07277, over 4838.00 frames. ], tot_loss[loss=0.2326, simple_loss=0.2861, pruned_loss=0.08958, over 956753.08 frames. ], batch size: 33, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:52:31,217 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-03-26 02:53:00,728 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.073e+02 1.767e+02 2.149e+02 2.599e+02 5.673e+02, threshold=4.299e+02, percent-clipped=2.0 +2023-03-26 02:53:07,541 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18930.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:53:09,238 INFO [finetune.py:976] (4/7) Epoch 4, batch 1750, loss[loss=0.2683, simple_loss=0.3199, pruned_loss=0.1083, over 4908.00 frames. ], tot_loss[loss=0.2351, simple_loss=0.2888, pruned_loss=0.0907, over 955984.25 frames. ], batch size: 36, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:53:11,227 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9402, 2.6811, 2.4127, 3.0512, 2.9079, 2.5665, 3.4517, 2.9009], + device='cuda:4'), covar=tensor([0.1398, 0.3224, 0.3520, 0.3206, 0.2554, 0.1677, 0.2956, 0.2218], + device='cuda:4'), in_proj_covar=tensor([0.0165, 0.0194, 0.0238, 0.0255, 0.0223, 0.0185, 0.0210, 0.0188], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 02:53:33,391 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7723, 3.7889, 3.6542, 2.0270, 3.9068, 2.7636, 0.8023, 2.6202], + device='cuda:4'), covar=tensor([0.2411, 0.1477, 0.1556, 0.2838, 0.0891, 0.1021, 0.4446, 0.1428], + device='cuda:4'), in_proj_covar=tensor([0.0155, 0.0170, 0.0164, 0.0129, 0.0156, 0.0123, 0.0146, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 02:53:52,363 INFO [finetune.py:976] (4/7) Epoch 4, batch 1800, loss[loss=0.188, simple_loss=0.2438, pruned_loss=0.06615, over 4694.00 frames. ], tot_loss[loss=0.237, simple_loss=0.2915, pruned_loss=0.09128, over 955788.73 frames. ], batch size: 23, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:53:57,497 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18991.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:54:06,968 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1889, 1.8038, 1.3465, 0.5959, 1.6652, 1.7820, 1.5201, 1.6914], + device='cuda:4'), covar=tensor([0.0816, 0.0988, 0.1701, 0.2231, 0.1419, 0.2380, 0.2423, 0.1008], + device='cuda:4'), in_proj_covar=tensor([0.0167, 0.0199, 0.0202, 0.0189, 0.0216, 0.0208, 0.0218, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 02:54:32,252 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.317e+02 1.869e+02 2.115e+02 2.590e+02 5.981e+02, threshold=4.230e+02, percent-clipped=1.0 +2023-03-26 02:54:39,920 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3402, 1.3201, 1.5945, 1.7444, 1.4908, 3.1848, 1.1713, 1.4137], + device='cuda:4'), covar=tensor([0.1025, 0.1650, 0.1303, 0.1013, 0.1517, 0.0254, 0.1482, 0.1689], + device='cuda:4'), in_proj_covar=tensor([0.0079, 0.0082, 0.0078, 0.0080, 0.0093, 0.0083, 0.0086, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 02:54:42,231 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6239, 1.4643, 2.0221, 3.2834, 2.2571, 2.3303, 1.0150, 2.5462], + device='cuda:4'), covar=tensor([0.1797, 0.1646, 0.1326, 0.0552, 0.0793, 0.1461, 0.1942, 0.0642], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0119, 0.0137, 0.0167, 0.0104, 0.0143, 0.0129, 0.0106], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 02:54:50,224 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19030.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:54:52,013 INFO [finetune.py:976] (4/7) Epoch 4, batch 1850, loss[loss=0.2098, simple_loss=0.2549, pruned_loss=0.08236, over 4343.00 frames. ], tot_loss[loss=0.2376, simple_loss=0.2926, pruned_loss=0.0913, over 955045.12 frames. ], batch size: 19, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:55:39,897 INFO [finetune.py:976] (4/7) Epoch 4, batch 1900, loss[loss=0.2179, simple_loss=0.2918, pruned_loss=0.07201, over 4744.00 frames. ], tot_loss[loss=0.24, simple_loss=0.2952, pruned_loss=0.0924, over 957664.28 frames. ], batch size: 54, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:56:12,754 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.301e+02 1.752e+02 2.082e+02 2.658e+02 3.786e+02, threshold=4.164e+02, percent-clipped=0.0 +2023-03-26 02:56:25,542 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=19129.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 02:56:33,397 INFO [finetune.py:976] (4/7) Epoch 4, batch 1950, loss[loss=0.1552, simple_loss=0.217, pruned_loss=0.04667, over 4723.00 frames. ], tot_loss[loss=0.2361, simple_loss=0.2915, pruned_loss=0.0903, over 959105.98 frames. ], batch size: 23, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:56:41,418 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19146.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:56:53,415 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19164.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 02:57:09,559 INFO [finetune.py:976] (4/7) Epoch 4, batch 2000, loss[loss=0.2081, simple_loss=0.2642, pruned_loss=0.07598, over 4901.00 frames. ], tot_loss[loss=0.2323, simple_loss=0.2875, pruned_loss=0.08851, over 957395.55 frames. ], batch size: 46, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:57:14,700 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=19190.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 02:57:17,078 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19194.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:57:18,463 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-26 02:57:35,168 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19212.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 02:57:39,344 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.251e+02 1.681e+02 2.009e+02 2.396e+02 5.395e+02, threshold=4.017e+02, percent-clipped=3.0 +2023-03-26 02:57:49,441 INFO [finetune.py:976] (4/7) Epoch 4, batch 2050, loss[loss=0.2511, simple_loss=0.2795, pruned_loss=0.1113, over 4157.00 frames. ], tot_loss[loss=0.2288, simple_loss=0.2837, pruned_loss=0.08694, over 955226.16 frames. ], batch size: 18, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:58:31,808 INFO [finetune.py:976] (4/7) Epoch 4, batch 2100, loss[loss=0.2085, simple_loss=0.2711, pruned_loss=0.07297, over 4753.00 frames. ], tot_loss[loss=0.2285, simple_loss=0.2831, pruned_loss=0.08694, over 952352.28 frames. ], batch size: 27, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 02:58:34,244 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19286.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:58:59,636 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0770, 1.8860, 1.5354, 1.8401, 1.7827, 1.7660, 1.7527, 2.6257], + device='cuda:4'), covar=tensor([0.8990, 0.9513, 0.7168, 0.9190, 0.7868, 0.4792, 0.9048, 0.2917], + device='cuda:4'), in_proj_covar=tensor([0.0276, 0.0252, 0.0219, 0.0283, 0.0235, 0.0196, 0.0240, 0.0188], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:4') +2023-03-26 02:59:09,862 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.737e+02 2.057e+02 2.371e+02 3.601e+02, threshold=4.115e+02, percent-clipped=0.0 +2023-03-26 02:59:26,248 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19330.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 02:59:27,977 INFO [finetune.py:976] (4/7) Epoch 4, batch 2150, loss[loss=0.2058, simple_loss=0.2795, pruned_loss=0.06611, over 4783.00 frames. ], tot_loss[loss=0.2308, simple_loss=0.2864, pruned_loss=0.08761, over 951813.01 frames. ], batch size: 29, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:00:09,714 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2218, 1.4301, 0.8761, 2.2873, 2.3476, 1.8093, 1.7331, 2.0812], + device='cuda:4'), covar=tensor([0.1505, 0.2225, 0.2378, 0.1184, 0.2024, 0.2041, 0.1427, 0.2098], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0099, 0.0117, 0.0094, 0.0125, 0.0097, 0.0101, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 03:00:10,913 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19378.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:00:13,946 INFO [finetune.py:976] (4/7) Epoch 4, batch 2200, loss[loss=0.2401, simple_loss=0.2927, pruned_loss=0.09374, over 4779.00 frames. ], tot_loss[loss=0.2339, simple_loss=0.2897, pruned_loss=0.08911, over 952698.13 frames. ], batch size: 26, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:00:33,371 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=19394.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:00:46,430 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8352, 1.3085, 1.6839, 1.5855, 1.4521, 1.4469, 1.4975, 1.5288], + device='cuda:4'), covar=tensor([0.7167, 1.0466, 0.8260, 1.0178, 1.0251, 0.8109, 1.2411, 0.7888], + device='cuda:4'), in_proj_covar=tensor([0.0229, 0.0252, 0.0257, 0.0263, 0.0243, 0.0218, 0.0278, 0.0221], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:00:47,396 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=2.03 vs. limit=2.0 +2023-03-26 03:01:05,355 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.065e+02 1.822e+02 2.092e+02 2.549e+02 4.918e+02, threshold=4.184e+02, percent-clipped=2.0 +2023-03-26 03:01:19,365 INFO [finetune.py:976] (4/7) Epoch 4, batch 2250, loss[loss=0.2419, simple_loss=0.3092, pruned_loss=0.08728, over 4808.00 frames. ], tot_loss[loss=0.2366, simple_loss=0.2921, pruned_loss=0.09054, over 951824.85 frames. ], batch size: 45, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:01:46,315 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=19455.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:02:09,425 INFO [finetune.py:976] (4/7) Epoch 4, batch 2300, loss[loss=0.2001, simple_loss=0.2741, pruned_loss=0.06301, over 4748.00 frames. ], tot_loss[loss=0.2342, simple_loss=0.2905, pruned_loss=0.0889, over 953316.88 frames. ], batch size: 27, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:02:12,900 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19485.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 03:02:31,191 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.3553, 3.7489, 3.9089, 4.2047, 4.0662, 3.8765, 4.4092, 1.4503], + device='cuda:4'), covar=tensor([0.0678, 0.0716, 0.0817, 0.0779, 0.1101, 0.1425, 0.0663, 0.4985], + device='cuda:4'), in_proj_covar=tensor([0.0359, 0.0245, 0.0278, 0.0293, 0.0339, 0.0286, 0.0311, 0.0301], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:02:34,858 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.1379, 2.6079, 2.4260, 1.4089, 2.6016, 2.2035, 1.8716, 2.4043], + device='cuda:4'), covar=tensor([0.0903, 0.1013, 0.1908, 0.2526, 0.2142, 0.2601, 0.2359, 0.1304], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0199, 0.0202, 0.0190, 0.0217, 0.0210, 0.0219, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:02:37,771 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.717e+02 2.025e+02 2.639e+02 4.089e+02, threshold=4.050e+02, percent-clipped=0.0 +2023-03-26 03:02:53,440 INFO [finetune.py:976] (4/7) Epoch 4, batch 2350, loss[loss=0.2189, simple_loss=0.2832, pruned_loss=0.07726, over 4765.00 frames. ], tot_loss[loss=0.232, simple_loss=0.2881, pruned_loss=0.08797, over 954098.83 frames. ], batch size: 27, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:03:27,719 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 03:03:37,781 INFO [finetune.py:976] (4/7) Epoch 4, batch 2400, loss[loss=0.1942, simple_loss=0.236, pruned_loss=0.07617, over 3967.00 frames. ], tot_loss[loss=0.229, simple_loss=0.2846, pruned_loss=0.08673, over 954872.54 frames. ], batch size: 17, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:03:40,239 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19586.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:04:10,036 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=19618.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:04:10,537 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.651e+02 1.936e+02 2.390e+02 3.810e+02, threshold=3.872e+02, percent-clipped=0.0 +2023-03-26 03:04:16,133 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-03-26 03:04:19,620 INFO [finetune.py:976] (4/7) Epoch 4, batch 2450, loss[loss=0.2781, simple_loss=0.3073, pruned_loss=0.1244, over 4794.00 frames. ], tot_loss[loss=0.2257, simple_loss=0.2809, pruned_loss=0.08521, over 955005.28 frames. ], batch size: 29, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:04:20,303 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19634.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:04:25,721 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5664, 1.4888, 2.0061, 1.2564, 1.6121, 1.7827, 1.5726, 2.0562], + device='cuda:4'), covar=tensor([0.1518, 0.2446, 0.1251, 0.1820, 0.0999, 0.1497, 0.2614, 0.0889], + device='cuda:4'), in_proj_covar=tensor([0.0208, 0.0208, 0.0205, 0.0199, 0.0183, 0.0226, 0.0217, 0.0206], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:04:44,694 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8205, 4.3061, 4.0880, 2.2293, 4.5428, 3.1771, 0.7618, 2.9780], + device='cuda:4'), covar=tensor([0.2762, 0.1721, 0.1384, 0.3314, 0.0682, 0.0994, 0.4770, 0.1634], + device='cuda:4'), in_proj_covar=tensor([0.0155, 0.0170, 0.0163, 0.0129, 0.0155, 0.0122, 0.0146, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 03:04:59,950 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=19679.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:05:02,278 INFO [finetune.py:976] (4/7) Epoch 4, batch 2500, loss[loss=0.1777, simple_loss=0.2454, pruned_loss=0.05505, over 4711.00 frames. ], tot_loss[loss=0.2292, simple_loss=0.2838, pruned_loss=0.08728, over 954293.86 frames. ], batch size: 23, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:05:20,322 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8859, 1.1914, 1.6726, 1.5879, 1.5109, 1.4830, 1.4866, 1.4856], + device='cuda:4'), covar=tensor([0.6680, 1.0540, 0.8827, 0.9801, 1.0815, 0.8206, 1.2476, 0.8190], + device='cuda:4'), in_proj_covar=tensor([0.0230, 0.0253, 0.0257, 0.0264, 0.0243, 0.0219, 0.0279, 0.0222], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:05:26,483 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8823, 2.6279, 2.1519, 1.2985, 2.2808, 2.1879, 1.7436, 2.2577], + device='cuda:4'), covar=tensor([0.0726, 0.0844, 0.1549, 0.2089, 0.1561, 0.2081, 0.2343, 0.1067], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0199, 0.0202, 0.0190, 0.0217, 0.0210, 0.0218, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:05:28,390 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5796, 1.5999, 1.3755, 1.3119, 1.8484, 1.9446, 1.6224, 1.4504], + device='cuda:4'), covar=tensor([0.0304, 0.0336, 0.0501, 0.0349, 0.0217, 0.0359, 0.0342, 0.0388], + device='cuda:4'), in_proj_covar=tensor([0.0084, 0.0113, 0.0137, 0.0117, 0.0104, 0.0098, 0.0091, 0.0109], + device='cuda:4'), out_proj_covar=tensor([6.5924e-05, 8.9450e-05, 1.1015e-04, 9.2538e-05, 8.2176e-05, 7.3192e-05, + 7.0261e-05, 8.4880e-05], device='cuda:4') +2023-03-26 03:05:30,655 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.296e+02 1.721e+02 2.075e+02 2.470e+02 4.533e+02, threshold=4.150e+02, percent-clipped=4.0 +2023-03-26 03:05:45,256 INFO [finetune.py:976] (4/7) Epoch 4, batch 2550, loss[loss=0.2354, simple_loss=0.2936, pruned_loss=0.08862, over 4879.00 frames. ], tot_loss[loss=0.2315, simple_loss=0.2868, pruned_loss=0.08807, over 954664.78 frames. ], batch size: 34, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:05:58,474 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19750.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:06:19,448 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.7699, 3.2638, 3.4167, 3.6434, 3.5184, 3.3075, 3.8161, 1.3285], + device='cuda:4'), covar=tensor([0.0772, 0.0871, 0.0897, 0.0991, 0.1236, 0.1595, 0.0850, 0.4962], + device='cuda:4'), in_proj_covar=tensor([0.0362, 0.0247, 0.0280, 0.0295, 0.0342, 0.0288, 0.0313, 0.0303], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:06:31,487 INFO [finetune.py:976] (4/7) Epoch 4, batch 2600, loss[loss=0.2073, simple_loss=0.2651, pruned_loss=0.07471, over 4793.00 frames. ], tot_loss[loss=0.2335, simple_loss=0.2888, pruned_loss=0.08911, over 954180.45 frames. ], batch size: 29, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:06:33,327 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19785.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 03:07:15,919 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.335e+02 1.804e+02 2.222e+02 2.871e+02 4.406e+02, threshold=4.445e+02, percent-clipped=2.0 +2023-03-26 03:07:33,028 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 03:07:34,471 INFO [finetune.py:976] (4/7) Epoch 4, batch 2650, loss[loss=0.1891, simple_loss=0.2517, pruned_loss=0.06324, over 4830.00 frames. ], tot_loss[loss=0.2357, simple_loss=0.2908, pruned_loss=0.09034, over 954859.17 frames. ], batch size: 49, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:07:34,539 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=19833.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 03:08:34,280 INFO [finetune.py:976] (4/7) Epoch 4, batch 2700, loss[loss=0.1872, simple_loss=0.2437, pruned_loss=0.06535, over 4780.00 frames. ], tot_loss[loss=0.2327, simple_loss=0.2882, pruned_loss=0.08863, over 956322.78 frames. ], batch size: 51, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:09:16,013 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.181e+02 1.716e+02 2.005e+02 2.489e+02 3.950e+02, threshold=4.009e+02, percent-clipped=0.0 +2023-03-26 03:09:16,792 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.7188, 1.6155, 1.8233, 1.1043, 1.6640, 1.9705, 1.8932, 1.6172], + device='cuda:4'), covar=tensor([0.0958, 0.0738, 0.0446, 0.0556, 0.0480, 0.0465, 0.0350, 0.0572], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0158, 0.0117, 0.0137, 0.0133, 0.0121, 0.0147, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.7610e-05, 1.1739e-04, 8.5354e-05, 1.0028e-04, 9.6238e-05, 8.9772e-05, + 1.0945e-04, 1.0728e-04], device='cuda:4') +2023-03-26 03:09:26,957 INFO [finetune.py:976] (4/7) Epoch 4, batch 2750, loss[loss=0.2325, simple_loss=0.2926, pruned_loss=0.0862, over 4814.00 frames. ], tot_loss[loss=0.2306, simple_loss=0.2856, pruned_loss=0.08783, over 955690.34 frames. ], batch size: 38, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:09:33,739 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-26 03:09:42,023 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5254, 1.1647, 1.3341, 1.1740, 1.6122, 1.6302, 1.4726, 1.2832], + device='cuda:4'), covar=tensor([0.0253, 0.0372, 0.0583, 0.0357, 0.0240, 0.0326, 0.0282, 0.0432], + device='cuda:4'), in_proj_covar=tensor([0.0085, 0.0114, 0.0138, 0.0118, 0.0104, 0.0099, 0.0092, 0.0109], + device='cuda:4'), out_proj_covar=tensor([6.6597e-05, 9.0119e-05, 1.1132e-04, 9.3364e-05, 8.2578e-05, 7.3857e-05, + 7.0627e-05, 8.5450e-05], device='cuda:4') +2023-03-26 03:09:54,791 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19974.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:10:00,207 INFO [finetune.py:976] (4/7) Epoch 4, batch 2800, loss[loss=0.1935, simple_loss=0.2474, pruned_loss=0.0698, over 4383.00 frames. ], tot_loss[loss=0.2278, simple_loss=0.2824, pruned_loss=0.08659, over 957992.31 frames. ], batch size: 19, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:10:20,551 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7369, 1.4996, 1.4714, 1.2613, 1.8598, 1.5239, 1.7273, 1.7156], + device='cuda:4'), covar=tensor([0.2064, 0.3570, 0.4547, 0.3794, 0.3173, 0.2313, 0.3384, 0.2892], + device='cuda:4'), in_proj_covar=tensor([0.0165, 0.0193, 0.0237, 0.0254, 0.0223, 0.0186, 0.0210, 0.0188], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:10:24,993 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.781e+02 2.091e+02 2.518e+02 3.954e+02, threshold=4.183e+02, percent-clipped=0.0 +2023-03-26 03:10:26,235 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20020.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:10:34,540 INFO [finetune.py:976] (4/7) Epoch 4, batch 2850, loss[loss=0.1802, simple_loss=0.2479, pruned_loss=0.05627, over 4811.00 frames. ], tot_loss[loss=0.2277, simple_loss=0.2822, pruned_loss=0.08656, over 956214.15 frames. ], batch size: 38, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:10:45,821 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20050.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:11:21,584 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20081.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:11:22,669 INFO [finetune.py:976] (4/7) Epoch 4, batch 2900, loss[loss=0.2678, simple_loss=0.3287, pruned_loss=0.1034, over 4806.00 frames. ], tot_loss[loss=0.2304, simple_loss=0.2854, pruned_loss=0.0877, over 955953.80 frames. ], batch size: 41, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:11:31,830 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20089.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:11:42,700 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20098.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:11:56,192 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20112.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:12:05,858 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.342e+02 1.862e+02 2.201e+02 2.748e+02 4.534e+02, threshold=4.402e+02, percent-clipped=1.0 +2023-03-26 03:12:27,412 INFO [finetune.py:976] (4/7) Epoch 4, batch 2950, loss[loss=0.2596, simple_loss=0.3074, pruned_loss=0.1059, over 4905.00 frames. ], tot_loss[loss=0.2325, simple_loss=0.2883, pruned_loss=0.08833, over 955655.19 frames. ], batch size: 37, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:12:48,254 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20150.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:13:10,789 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20173.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:13:18,822 INFO [finetune.py:976] (4/7) Epoch 4, batch 3000, loss[loss=0.2355, simple_loss=0.2862, pruned_loss=0.09237, over 4714.00 frames. ], tot_loss[loss=0.2332, simple_loss=0.2891, pruned_loss=0.08863, over 951546.57 frames. ], batch size: 23, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:13:18,822 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 03:13:34,008 INFO [finetune.py:1010] (4/7) Epoch 4, validation: loss=0.169, simple_loss=0.2409, pruned_loss=0.04857, over 2265189.00 frames. +2023-03-26 03:13:34,009 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6301MB +2023-03-26 03:13:36,443 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7262, 1.5538, 2.1311, 1.5727, 2.0008, 1.9846, 1.5022, 2.2897], + device='cuda:4'), covar=tensor([0.1632, 0.2452, 0.1699, 0.2374, 0.0953, 0.1748, 0.2961, 0.0872], + device='cuda:4'), in_proj_covar=tensor([0.0209, 0.0209, 0.0206, 0.0199, 0.0184, 0.0227, 0.0218, 0.0206], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:13:47,431 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6550, 3.6973, 3.5602, 1.5948, 3.8394, 2.9095, 0.7058, 2.5011], + device='cuda:4'), covar=tensor([0.2504, 0.1715, 0.1532, 0.3440, 0.0955, 0.0923, 0.4830, 0.1621], + device='cuda:4'), in_proj_covar=tensor([0.0155, 0.0170, 0.0163, 0.0128, 0.0155, 0.0121, 0.0145, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 03:13:58,689 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20204.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:14:18,156 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.180e+02 1.917e+02 2.134e+02 2.804e+02 4.274e+02, threshold=4.268e+02, percent-clipped=0.0 +2023-03-26 03:14:20,139 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.5167, 1.4440, 1.5744, 1.0167, 1.4733, 1.7634, 1.7209, 1.4421], + device='cuda:4'), covar=tensor([0.0949, 0.0671, 0.0433, 0.0578, 0.0435, 0.0508, 0.0312, 0.0586], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0159, 0.0119, 0.0138, 0.0134, 0.0122, 0.0149, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.8240e-05, 1.1784e-04, 8.6317e-05, 1.0124e-04, 9.6866e-05, 9.0571e-05, + 1.1082e-04, 1.0787e-04], device='cuda:4') +2023-03-26 03:14:23,621 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5208, 1.3662, 1.8593, 2.9478, 2.0922, 2.1164, 1.0480, 2.3035], + device='cuda:4'), covar=tensor([0.1820, 0.1519, 0.1287, 0.0612, 0.0847, 0.1397, 0.1772, 0.0719], + device='cuda:4'), in_proj_covar=tensor([0.0104, 0.0119, 0.0137, 0.0167, 0.0104, 0.0143, 0.0130, 0.0105], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 03:14:27,242 INFO [finetune.py:976] (4/7) Epoch 4, batch 3050, loss[loss=0.2436, simple_loss=0.2992, pruned_loss=0.09401, over 4803.00 frames. ], tot_loss[loss=0.2343, simple_loss=0.2906, pruned_loss=0.08902, over 953275.40 frames. ], batch size: 45, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:15:06,993 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20265.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:15:13,539 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.0735, 4.3884, 4.5517, 4.8881, 4.7703, 4.4787, 5.1790, 1.4756], + device='cuda:4'), covar=tensor([0.0759, 0.0763, 0.0874, 0.0896, 0.1369, 0.1538, 0.0628, 0.5735], + device='cuda:4'), in_proj_covar=tensor([0.0358, 0.0243, 0.0277, 0.0292, 0.0338, 0.0285, 0.0308, 0.0299], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:15:16,638 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20274.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:15:24,345 INFO [finetune.py:976] (4/7) Epoch 4, batch 3100, loss[loss=0.2433, simple_loss=0.2859, pruned_loss=0.1003, over 4151.00 frames. ], tot_loss[loss=0.2314, simple_loss=0.288, pruned_loss=0.08744, over 954406.28 frames. ], batch size: 65, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:16:01,241 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.111e+02 1.549e+02 1.969e+02 2.570e+02 5.632e+02, threshold=3.937e+02, percent-clipped=1.0 +2023-03-26 03:16:03,107 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20322.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:16:14,158 INFO [finetune.py:976] (4/7) Epoch 4, batch 3150, loss[loss=0.2091, simple_loss=0.2704, pruned_loss=0.07387, over 4784.00 frames. ], tot_loss[loss=0.2297, simple_loss=0.2854, pruned_loss=0.08697, over 954156.95 frames. ], batch size: 29, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:16:56,929 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20376.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:17:01,161 INFO [finetune.py:976] (4/7) Epoch 4, batch 3200, loss[loss=0.1873, simple_loss=0.2546, pruned_loss=0.06005, over 4752.00 frames. ], tot_loss[loss=0.2264, simple_loss=0.282, pruned_loss=0.08533, over 955212.89 frames. ], batch size: 27, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:17:04,823 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20388.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:17:40,201 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20418.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:17:40,662 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.143e+02 1.625e+02 1.959e+02 2.342e+02 5.079e+02, threshold=3.919e+02, percent-clipped=3.0 +2023-03-26 03:17:48,787 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.54 vs. limit=5.0 +2023-03-26 03:17:58,485 INFO [finetune.py:976] (4/7) Epoch 4, batch 3250, loss[loss=0.2621, simple_loss=0.32, pruned_loss=0.1021, over 4819.00 frames. ], tot_loss[loss=0.2272, simple_loss=0.2827, pruned_loss=0.08588, over 955435.34 frames. ], batch size: 40, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:18:06,351 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20445.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:18:08,887 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20449.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:18:21,894 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20468.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:18:26,360 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 03:18:29,606 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20479.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:18:30,836 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8008, 1.5296, 1.4914, 1.3271, 1.8513, 1.5412, 1.7242, 1.7318], + device='cuda:4'), covar=tensor([0.1840, 0.3144, 0.3995, 0.3192, 0.2944, 0.2011, 0.3279, 0.2462], + device='cuda:4'), in_proj_covar=tensor([0.0166, 0.0194, 0.0237, 0.0255, 0.0225, 0.0187, 0.0210, 0.0188], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:18:31,929 INFO [finetune.py:976] (4/7) Epoch 4, batch 3300, loss[loss=0.2333, simple_loss=0.2956, pruned_loss=0.08548, over 4811.00 frames. ], tot_loss[loss=0.2294, simple_loss=0.2851, pruned_loss=0.08681, over 954002.27 frames. ], batch size: 45, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:18:38,652 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6343, 1.6300, 1.7356, 0.9478, 1.6203, 1.9250, 1.9147, 1.5232], + device='cuda:4'), covar=tensor([0.0949, 0.0666, 0.0491, 0.0684, 0.0438, 0.0435, 0.0314, 0.0698], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0157, 0.0117, 0.0136, 0.0132, 0.0121, 0.0147, 0.0144], + device='cuda:4'), out_proj_covar=tensor([9.7030e-05, 1.1620e-04, 8.5347e-05, 9.9530e-05, 9.5473e-05, 8.9651e-05, + 1.0954e-04, 1.0648e-04], device='cuda:4') +2023-03-26 03:19:09,460 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.164e+02 1.750e+02 2.039e+02 2.534e+02 4.074e+02, threshold=4.078e+02, percent-clipped=2.0 +2023-03-26 03:19:29,492 INFO [finetune.py:976] (4/7) Epoch 4, batch 3350, loss[loss=0.1835, simple_loss=0.2589, pruned_loss=0.05404, over 4796.00 frames. ], tot_loss[loss=0.2307, simple_loss=0.2867, pruned_loss=0.08739, over 952593.12 frames. ], batch size: 29, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:19:41,091 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 03:19:59,031 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20560.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:20:13,279 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5266, 1.4255, 1.3833, 1.5070, 0.9513, 3.2148, 1.2138, 1.6360], + device='cuda:4'), covar=tensor([0.3341, 0.2413, 0.2125, 0.2296, 0.2042, 0.0192, 0.2813, 0.1418], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0113, 0.0117, 0.0121, 0.0117, 0.0097, 0.0101, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 03:20:17,267 INFO [finetune.py:976] (4/7) Epoch 4, batch 3400, loss[loss=0.2366, simple_loss=0.3053, pruned_loss=0.08393, over 4918.00 frames. ], tot_loss[loss=0.2314, simple_loss=0.2876, pruned_loss=0.08762, over 954458.70 frames. ], batch size: 33, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:20:20,438 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20588.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:20:37,710 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0126, 1.9168, 1.5100, 1.9379, 2.0208, 1.6396, 2.4070, 1.9288], + device='cuda:4'), covar=tensor([0.1902, 0.3702, 0.4446, 0.3993, 0.3158, 0.2183, 0.4149, 0.2601], + device='cuda:4'), in_proj_covar=tensor([0.0167, 0.0196, 0.0238, 0.0256, 0.0226, 0.0188, 0.0211, 0.0190], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:20:43,798 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6241, 1.4396, 1.3419, 1.5757, 1.9964, 1.5900, 1.1575, 1.3845], + device='cuda:4'), covar=tensor([0.2413, 0.2340, 0.2076, 0.1898, 0.1801, 0.1306, 0.2947, 0.1879], + device='cuda:4'), in_proj_covar=tensor([0.0232, 0.0209, 0.0198, 0.0184, 0.0234, 0.0174, 0.0215, 0.0187], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:20:46,764 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.150e+02 1.735e+02 2.048e+02 2.538e+02 3.974e+02, threshold=4.096e+02, percent-clipped=0.0 +2023-03-26 03:21:05,099 INFO [finetune.py:976] (4/7) Epoch 4, batch 3450, loss[loss=0.213, simple_loss=0.2679, pruned_loss=0.07906, over 4211.00 frames. ], tot_loss[loss=0.2306, simple_loss=0.2873, pruned_loss=0.08691, over 953567.48 frames. ], batch size: 18, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:21:22,516 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20649.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:21:23,456 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 03:21:23,805 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-03-26 03:21:36,108 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5878, 1.5839, 1.2452, 1.3104, 1.8011, 1.8460, 1.6477, 1.3913], + device='cuda:4'), covar=tensor([0.0318, 0.0388, 0.0593, 0.0434, 0.0231, 0.0348, 0.0301, 0.0442], + device='cuda:4'), in_proj_covar=tensor([0.0085, 0.0113, 0.0137, 0.0118, 0.0104, 0.0099, 0.0091, 0.0109], + device='cuda:4'), out_proj_covar=tensor([6.6470e-05, 8.9354e-05, 1.1040e-04, 9.2979e-05, 8.2841e-05, 7.3459e-05, + 6.9986e-05, 8.4903e-05], device='cuda:4') +2023-03-26 03:21:45,688 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20676.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:21:51,374 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4182, 1.3063, 1.3025, 1.3357, 0.7421, 2.2082, 0.7441, 1.2282], + device='cuda:4'), covar=tensor([0.3357, 0.2465, 0.2125, 0.2439, 0.2224, 0.0357, 0.2735, 0.1455], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0112, 0.0117, 0.0120, 0.0117, 0.0097, 0.0101, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 03:21:51,862 INFO [finetune.py:976] (4/7) Epoch 4, batch 3500, loss[loss=0.2524, simple_loss=0.3065, pruned_loss=0.09919, over 4814.00 frames. ], tot_loss[loss=0.2294, simple_loss=0.2854, pruned_loss=0.08674, over 954712.60 frames. ], batch size: 39, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:21:54,260 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8292, 4.0271, 3.8653, 2.0339, 4.2140, 3.0654, 0.9099, 2.9371], + device='cuda:4'), covar=tensor([0.2328, 0.1769, 0.1404, 0.3167, 0.0884, 0.0898, 0.4505, 0.1467], + device='cuda:4'), in_proj_covar=tensor([0.0154, 0.0170, 0.0162, 0.0128, 0.0155, 0.0121, 0.0145, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 03:22:14,141 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9192, 1.2594, 0.8734, 1.5747, 2.1210, 1.3724, 1.5299, 1.7502], + device='cuda:4'), covar=tensor([0.1465, 0.2138, 0.2373, 0.1328, 0.2101, 0.2219, 0.1520, 0.1950], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0100, 0.0119, 0.0095, 0.0127, 0.0099, 0.0102, 0.0096], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 03:22:31,512 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.236e+02 1.688e+02 2.022e+02 2.523e+02 5.341e+02, threshold=4.043e+02, percent-clipped=2.0 +2023-03-26 03:22:35,129 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20724.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:22:43,661 INFO [finetune.py:976] (4/7) Epoch 4, batch 3550, loss[loss=0.2198, simple_loss=0.2678, pruned_loss=0.08594, over 4855.00 frames. ], tot_loss[loss=0.2278, simple_loss=0.2829, pruned_loss=0.08632, over 954487.30 frames. ], batch size: 34, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:22:45,820 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 03:22:56,121 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20744.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:22:56,750 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20745.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:23:24,109 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20768.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:23:33,941 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20774.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:23:41,430 INFO [finetune.py:976] (4/7) Epoch 4, batch 3600, loss[loss=0.1626, simple_loss=0.2308, pruned_loss=0.04715, over 4774.00 frames. ], tot_loss[loss=0.2256, simple_loss=0.2807, pruned_loss=0.08523, over 954071.88 frames. ], batch size: 26, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:23:52,913 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20793.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:24:24,359 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20816.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:24:31,499 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.828e+01 1.831e+02 2.152e+02 2.506e+02 5.159e+02, threshold=4.304e+02, percent-clipped=1.0 +2023-03-26 03:24:50,039 INFO [finetune.py:976] (4/7) Epoch 4, batch 3650, loss[loss=0.2981, simple_loss=0.3563, pruned_loss=0.12, over 4825.00 frames. ], tot_loss[loss=0.2284, simple_loss=0.2835, pruned_loss=0.08669, over 954846.56 frames. ], batch size: 40, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:24:54,539 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8505, 2.5642, 1.9822, 1.1047, 2.3215, 2.1806, 1.8929, 2.2080], + device='cuda:4'), covar=tensor([0.0890, 0.0882, 0.1848, 0.2286, 0.1604, 0.2035, 0.2125, 0.1045], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0200, 0.0203, 0.0190, 0.0217, 0.0210, 0.0219, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:25:24,817 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20860.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:25:52,757 INFO [finetune.py:976] (4/7) Epoch 4, batch 3700, loss[loss=0.2565, simple_loss=0.3164, pruned_loss=0.09833, over 4897.00 frames. ], tot_loss[loss=0.2315, simple_loss=0.2871, pruned_loss=0.08794, over 955886.20 frames. ], batch size: 43, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:26:17,163 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=20908.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:26:24,268 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.096e+02 1.710e+02 2.022e+02 2.626e+02 5.956e+02, threshold=4.044e+02, percent-clipped=2.0 +2023-03-26 03:26:34,614 INFO [finetune.py:976] (4/7) Epoch 4, batch 3750, loss[loss=0.2701, simple_loss=0.3097, pruned_loss=0.1152, over 4771.00 frames. ], tot_loss[loss=0.234, simple_loss=0.2894, pruned_loss=0.08926, over 952455.58 frames. ], batch size: 26, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:26:46,570 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20944.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:27:33,089 INFO [finetune.py:976] (4/7) Epoch 4, batch 3800, loss[loss=0.2256, simple_loss=0.2885, pruned_loss=0.08137, over 4821.00 frames. ], tot_loss[loss=0.2346, simple_loss=0.2908, pruned_loss=0.08921, over 953428.76 frames. ], batch size: 30, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:27:43,288 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.92 vs. limit=5.0 +2023-03-26 03:28:14,225 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.257e+02 1.718e+02 1.983e+02 2.372e+02 3.493e+02, threshold=3.966e+02, percent-clipped=0.0 +2023-03-26 03:28:29,875 INFO [finetune.py:976] (4/7) Epoch 4, batch 3850, loss[loss=0.2072, simple_loss=0.2616, pruned_loss=0.07638, over 4821.00 frames. ], tot_loss[loss=0.234, simple_loss=0.2893, pruned_loss=0.08938, over 951603.47 frames. ], batch size: 39, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:28:37,706 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21044.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:28:45,302 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21052.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:28:54,217 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0140, 1.8080, 2.2526, 3.4998, 2.5314, 2.6189, 1.1268, 2.7489], + device='cuda:4'), covar=tensor([0.1545, 0.1436, 0.1271, 0.0594, 0.0805, 0.1949, 0.1779, 0.0576], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0119, 0.0137, 0.0168, 0.0105, 0.0144, 0.0130, 0.0105], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 03:28:54,386 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 03:29:08,439 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21074.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:29:19,707 INFO [finetune.py:976] (4/7) Epoch 4, batch 3900, loss[loss=0.1909, simple_loss=0.2532, pruned_loss=0.06424, over 4811.00 frames. ], tot_loss[loss=0.231, simple_loss=0.2859, pruned_loss=0.088, over 951968.08 frames. ], batch size: 25, lr: 3.96e-03, grad_scale: 64.0 +2023-03-26 03:29:22,152 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1411, 2.0263, 2.1343, 1.1121, 2.2668, 2.5996, 2.2086, 2.0272], + device='cuda:4'), covar=tensor([0.0995, 0.0746, 0.0613, 0.0730, 0.0596, 0.0406, 0.0490, 0.0640], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0158, 0.0118, 0.0137, 0.0132, 0.0121, 0.0147, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.7895e-05, 1.1697e-04, 8.6095e-05, 1.0039e-04, 9.5726e-05, 8.9614e-05, + 1.0988e-04, 1.0793e-04], device='cuda:4') +2023-03-26 03:29:26,805 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21092.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:29:43,722 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21113.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:29:46,794 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21118.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:29:47,851 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.709e+02 1.984e+02 2.370e+02 6.134e+02, threshold=3.968e+02, percent-clipped=2.0 +2023-03-26 03:29:49,149 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21122.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:29:59,344 INFO [finetune.py:976] (4/7) Epoch 4, batch 3950, loss[loss=0.2046, simple_loss=0.2613, pruned_loss=0.07396, over 4909.00 frames. ], tot_loss[loss=0.2282, simple_loss=0.2828, pruned_loss=0.08684, over 954412.76 frames. ], batch size: 36, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:30:00,067 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4035, 2.1709, 1.6519, 0.7155, 1.8670, 1.9056, 1.6763, 1.9204], + device='cuda:4'), covar=tensor([0.0922, 0.0914, 0.1643, 0.2318, 0.1562, 0.3081, 0.2534, 0.1045], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0201, 0.0203, 0.0190, 0.0217, 0.0211, 0.0219, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:30:41,989 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8690, 1.3763, 1.7003, 1.6893, 1.5165, 1.5196, 1.6189, 1.5886], + device='cuda:4'), covar=tensor([0.7140, 1.0205, 0.8209, 0.9310, 1.0410, 0.7083, 1.1261, 0.7857], + device='cuda:4'), in_proj_covar=tensor([0.0229, 0.0252, 0.0256, 0.0262, 0.0242, 0.0218, 0.0277, 0.0221], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:30:43,160 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1826, 3.6107, 3.8345, 4.0383, 3.9320, 3.7237, 4.2373, 1.3461], + device='cuda:4'), covar=tensor([0.0900, 0.0909, 0.0972, 0.1157, 0.1283, 0.1476, 0.0738, 0.5492], + device='cuda:4'), in_proj_covar=tensor([0.0357, 0.0242, 0.0275, 0.0292, 0.0336, 0.0284, 0.0306, 0.0299], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:30:47,968 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21179.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:30:50,289 INFO [finetune.py:976] (4/7) Epoch 4, batch 4000, loss[loss=0.2683, simple_loss=0.3081, pruned_loss=0.1142, over 4133.00 frames. ], tot_loss[loss=0.2267, simple_loss=0.2815, pruned_loss=0.08594, over 954882.32 frames. ], batch size: 65, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:30:57,539 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8497, 1.6487, 1.6536, 1.8898, 2.4170, 1.9381, 1.3143, 1.5469], + device='cuda:4'), covar=tensor([0.2472, 0.2464, 0.2089, 0.2083, 0.1931, 0.1289, 0.3164, 0.2025], + device='cuda:4'), in_proj_covar=tensor([0.0233, 0.0210, 0.0199, 0.0185, 0.0236, 0.0175, 0.0216, 0.0188], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:31:19,052 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.740e+02 2.148e+02 2.427e+02 4.357e+02, threshold=4.296e+02, percent-clipped=1.0 +2023-03-26 03:31:27,485 INFO [finetune.py:976] (4/7) Epoch 4, batch 4050, loss[loss=0.1751, simple_loss=0.2284, pruned_loss=0.06089, over 4239.00 frames. ], tot_loss[loss=0.2296, simple_loss=0.2845, pruned_loss=0.08732, over 951886.73 frames. ], batch size: 18, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:31:29,781 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=5.00 vs. limit=5.0 +2023-03-26 03:31:35,246 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21244.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:31:43,738 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-26 03:32:12,118 INFO [finetune.py:976] (4/7) Epoch 4, batch 4100, loss[loss=0.2482, simple_loss=0.3119, pruned_loss=0.09226, over 4744.00 frames. ], tot_loss[loss=0.2327, simple_loss=0.2883, pruned_loss=0.08852, over 951428.71 frames. ], batch size: 59, lr: 3.96e-03, grad_scale: 32.0 +2023-03-26 03:32:15,309 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.8145, 3.3157, 3.4793, 3.6926, 3.5328, 3.3393, 3.8833, 1.2769], + device='cuda:4'), covar=tensor([0.0886, 0.0860, 0.0842, 0.0932, 0.1367, 0.1434, 0.0831, 0.4933], + device='cuda:4'), in_proj_covar=tensor([0.0357, 0.0242, 0.0276, 0.0293, 0.0336, 0.0284, 0.0305, 0.0299], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:32:23,704 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21292.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:32:42,818 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7148, 2.1992, 1.8602, 0.9209, 2.1223, 1.9782, 1.5595, 1.9414], + device='cuda:4'), covar=tensor([0.0656, 0.1193, 0.1813, 0.2413, 0.1796, 0.2329, 0.2576, 0.1247], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0201, 0.0204, 0.0191, 0.0219, 0.0211, 0.0220, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:32:44,667 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21312.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:32:45,317 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8587, 1.4989, 2.2471, 1.4778, 2.0530, 2.1448, 1.4822, 2.2909], + device='cuda:4'), covar=tensor([0.1482, 0.2188, 0.1222, 0.1925, 0.0903, 0.1399, 0.3001, 0.0853], + device='cuda:4'), in_proj_covar=tensor([0.0206, 0.0206, 0.0203, 0.0196, 0.0183, 0.0225, 0.0215, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:32:46,559 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6546, 1.5195, 1.5051, 1.5732, 1.2706, 3.5104, 1.5473, 2.1190], + device='cuda:4'), covar=tensor([0.3499, 0.2483, 0.2093, 0.2283, 0.1767, 0.0168, 0.2606, 0.1319], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0112, 0.0116, 0.0120, 0.0116, 0.0097, 0.0101, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 03:32:53,585 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.825e+02 2.074e+02 2.571e+02 5.101e+02, threshold=4.147e+02, percent-clipped=2.0 +2023-03-26 03:33:04,188 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21328.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:33:07,032 INFO [finetune.py:976] (4/7) Epoch 4, batch 4150, loss[loss=0.2145, simple_loss=0.2818, pruned_loss=0.07361, over 4790.00 frames. ], tot_loss[loss=0.2332, simple_loss=0.289, pruned_loss=0.08874, over 953345.45 frames. ], batch size: 29, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:33:36,551 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1283, 2.0443, 2.7727, 1.6391, 2.3819, 2.6127, 1.9506, 2.5967], + device='cuda:4'), covar=tensor([0.1832, 0.2258, 0.1641, 0.2802, 0.1124, 0.1765, 0.2641, 0.1233], + device='cuda:4'), in_proj_covar=tensor([0.0207, 0.0208, 0.0204, 0.0198, 0.0184, 0.0227, 0.0216, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:33:45,525 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21373.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:33:52,601 INFO [finetune.py:976] (4/7) Epoch 4, batch 4200, loss[loss=0.1965, simple_loss=0.2708, pruned_loss=0.06114, over 4757.00 frames. ], tot_loss[loss=0.2327, simple_loss=0.2892, pruned_loss=0.08813, over 952896.19 frames. ], batch size: 27, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:34:02,008 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21389.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:34:21,944 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21408.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:34:39,319 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.725e+02 2.024e+02 2.533e+02 3.913e+02, threshold=4.049e+02, percent-clipped=0.0 +2023-03-26 03:34:39,427 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3516, 2.8868, 2.7894, 1.3599, 3.0303, 2.0466, 0.6993, 1.8822], + device='cuda:4'), covar=tensor([0.2408, 0.2213, 0.1796, 0.3417, 0.1338, 0.1224, 0.4250, 0.1757], + device='cuda:4'), in_proj_covar=tensor([0.0154, 0.0170, 0.0162, 0.0128, 0.0155, 0.0121, 0.0145, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 03:34:50,757 INFO [finetune.py:976] (4/7) Epoch 4, batch 4250, loss[loss=0.2199, simple_loss=0.2745, pruned_loss=0.08263, over 4907.00 frames. ], tot_loss[loss=0.2318, simple_loss=0.288, pruned_loss=0.08785, over 955241.78 frames. ], batch size: 36, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:35:18,394 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21474.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:35:24,280 INFO [finetune.py:976] (4/7) Epoch 4, batch 4300, loss[loss=0.24, simple_loss=0.2949, pruned_loss=0.09251, over 4904.00 frames. ], tot_loss[loss=0.2291, simple_loss=0.2847, pruned_loss=0.08672, over 954233.05 frames. ], batch size: 43, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:35:53,224 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.244e+02 1.671e+02 2.065e+02 2.560e+02 4.445e+02, threshold=4.130e+02, percent-clipped=1.0 +2023-03-26 03:36:01,105 INFO [finetune.py:976] (4/7) Epoch 4, batch 4350, loss[loss=0.2309, simple_loss=0.2856, pruned_loss=0.08809, over 4901.00 frames. ], tot_loss[loss=0.2257, simple_loss=0.2812, pruned_loss=0.08507, over 956958.02 frames. ], batch size: 35, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:36:40,600 INFO [finetune.py:976] (4/7) Epoch 4, batch 4400, loss[loss=0.2881, simple_loss=0.3428, pruned_loss=0.1167, over 4901.00 frames. ], tot_loss[loss=0.2283, simple_loss=0.2836, pruned_loss=0.08653, over 954491.91 frames. ], batch size: 43, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:36:49,029 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21595.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:36:58,800 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-26 03:37:05,948 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21613.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:37:15,011 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.167e+02 1.721e+02 2.036e+02 2.627e+02 4.967e+02, threshold=4.072e+02, percent-clipped=2.0 +2023-03-26 03:37:22,899 INFO [finetune.py:976] (4/7) Epoch 4, batch 4450, loss[loss=0.2778, simple_loss=0.3306, pruned_loss=0.1125, over 4900.00 frames. ], tot_loss[loss=0.2332, simple_loss=0.2887, pruned_loss=0.08881, over 955832.71 frames. ], batch size: 43, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:37:38,564 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21656.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:37:46,219 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5695, 1.4903, 1.9393, 1.8261, 1.7209, 3.9991, 1.3969, 1.7196], + device='cuda:4'), covar=tensor([0.1069, 0.1889, 0.1395, 0.1167, 0.1730, 0.0201, 0.1737, 0.1906], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0081, 0.0078, 0.0080, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 03:37:51,478 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21668.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:38:01,189 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21674.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:38:12,009 INFO [finetune.py:976] (4/7) Epoch 4, batch 4500, loss[loss=0.1755, simple_loss=0.2258, pruned_loss=0.06258, over 4069.00 frames. ], tot_loss[loss=0.2339, simple_loss=0.2895, pruned_loss=0.08917, over 953246.75 frames. ], batch size: 17, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:38:12,684 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21684.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:38:19,928 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.2180, 1.2992, 1.3797, 0.6304, 1.1535, 1.5011, 1.5846, 1.3147], + device='cuda:4'), covar=tensor([0.0751, 0.0441, 0.0441, 0.0488, 0.0400, 0.0437, 0.0224, 0.0477], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0156, 0.0118, 0.0136, 0.0131, 0.0121, 0.0146, 0.0144], + device='cuda:4'), out_proj_covar=tensor([9.7097e-05, 1.1626e-04, 8.5461e-05, 9.9530e-05, 9.4829e-05, 8.9338e-05, + 1.0889e-04, 1.0683e-04], device='cuda:4') +2023-03-26 03:38:40,207 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4054, 1.3766, 1.4474, 0.7629, 1.5606, 1.4509, 1.4564, 1.3001], + device='cuda:4'), covar=tensor([0.0653, 0.0683, 0.0699, 0.1017, 0.0701, 0.0795, 0.0656, 0.1113], + device='cuda:4'), in_proj_covar=tensor([0.0139, 0.0133, 0.0145, 0.0129, 0.0111, 0.0144, 0.0148, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:38:40,804 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5345, 1.4943, 1.9348, 1.7056, 1.7682, 3.9113, 1.4607, 1.7970], + device='cuda:4'), covar=tensor([0.0993, 0.1644, 0.1306, 0.1057, 0.1475, 0.0203, 0.1408, 0.1714], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0081, 0.0078, 0.0080, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 03:38:41,417 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21708.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:38:49,075 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.156e+02 1.721e+02 2.032e+02 2.543e+02 5.339e+02, threshold=4.063e+02, percent-clipped=3.0 +2023-03-26 03:38:58,599 INFO [finetune.py:976] (4/7) Epoch 4, batch 4550, loss[loss=0.2516, simple_loss=0.3077, pruned_loss=0.09779, over 4841.00 frames. ], tot_loss[loss=0.2352, simple_loss=0.291, pruned_loss=0.08968, over 953009.41 frames. ], batch size: 49, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:39:02,566 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-26 03:39:09,001 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7344, 1.5801, 1.9423, 1.9406, 1.8194, 4.0789, 1.5109, 1.8425], + device='cuda:4'), covar=tensor([0.0941, 0.1706, 0.1321, 0.0992, 0.1486, 0.0189, 0.1444, 0.1684], + device='cuda:4'), in_proj_covar=tensor([0.0079, 0.0081, 0.0078, 0.0080, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 03:39:14,116 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21756.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:39:18,473 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21763.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:39:30,507 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21774.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:39:41,977 INFO [finetune.py:976] (4/7) Epoch 4, batch 4600, loss[loss=0.2213, simple_loss=0.2774, pruned_loss=0.08258, over 4854.00 frames. ], tot_loss[loss=0.2336, simple_loss=0.2898, pruned_loss=0.08871, over 954628.40 frames. ], batch size: 44, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:39:51,169 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8461, 1.9172, 1.8722, 1.2564, 2.1215, 2.1353, 1.9930, 1.6716], + device='cuda:4'), covar=tensor([0.0639, 0.0584, 0.0722, 0.0992, 0.0561, 0.0685, 0.0619, 0.1118], + device='cuda:4'), in_proj_covar=tensor([0.0139, 0.0134, 0.0146, 0.0129, 0.0112, 0.0144, 0.0148, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:40:05,678 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3710, 2.0136, 1.6318, 0.7698, 1.9049, 1.8706, 1.6859, 1.9233], + device='cuda:4'), covar=tensor([0.0848, 0.1007, 0.1538, 0.2257, 0.1469, 0.2433, 0.2250, 0.0968], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0203, 0.0204, 0.0192, 0.0220, 0.0211, 0.0221, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:40:23,613 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.33 vs. limit=5.0 +2023-03-26 03:40:25,681 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.219e+02 1.789e+02 2.167e+02 2.687e+02 4.147e+02, threshold=4.334e+02, percent-clipped=1.0 +2023-03-26 03:40:32,332 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=21822.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:40:33,583 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21824.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:40:44,471 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-26 03:40:44,876 INFO [finetune.py:976] (4/7) Epoch 4, batch 4650, loss[loss=0.1766, simple_loss=0.2442, pruned_loss=0.05452, over 4852.00 frames. ], tot_loss[loss=0.23, simple_loss=0.2862, pruned_loss=0.08688, over 954316.31 frames. ], batch size: 44, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:41:06,183 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21850.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:41:19,442 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4112, 1.2908, 1.6565, 2.4569, 1.6865, 2.1216, 1.0018, 2.0264], + device='cuda:4'), covar=tensor([0.1723, 0.1600, 0.1088, 0.0648, 0.0944, 0.1129, 0.1597, 0.0700], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0119, 0.0137, 0.0167, 0.0104, 0.0143, 0.0130, 0.0104], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 03:41:28,342 INFO [finetune.py:976] (4/7) Epoch 4, batch 4700, loss[loss=0.2227, simple_loss=0.2693, pruned_loss=0.08804, over 4833.00 frames. ], tot_loss[loss=0.2273, simple_loss=0.2832, pruned_loss=0.0857, over 955887.21 frames. ], batch size: 30, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:41:34,301 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9284, 3.7403, 3.7246, 1.9282, 3.8784, 2.8518, 0.9136, 2.6718], + device='cuda:4'), covar=tensor([0.2029, 0.1755, 0.1316, 0.3037, 0.1009, 0.0914, 0.4372, 0.1464], + device='cuda:4'), in_proj_covar=tensor([0.0155, 0.0172, 0.0164, 0.0129, 0.0156, 0.0122, 0.0147, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 03:41:51,082 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 03:42:02,057 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21911.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:42:08,439 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.749e+02 2.081e+02 2.546e+02 7.973e+02, threshold=4.162e+02, percent-clipped=1.0 +2023-03-26 03:42:16,911 INFO [finetune.py:976] (4/7) Epoch 4, batch 4750, loss[loss=0.2394, simple_loss=0.2861, pruned_loss=0.09637, over 4830.00 frames. ], tot_loss[loss=0.2257, simple_loss=0.2815, pruned_loss=0.0849, over 956205.76 frames. ], batch size: 33, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:42:19,423 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1311, 1.7768, 1.9298, 1.8551, 1.6891, 1.7469, 1.8008, 1.8807], + device='cuda:4'), covar=tensor([0.6040, 0.8800, 0.6605, 0.8540, 0.9713, 0.7516, 1.1544, 0.5996], + device='cuda:4'), in_proj_covar=tensor([0.0229, 0.0252, 0.0256, 0.0261, 0.0242, 0.0218, 0.0278, 0.0222], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:42:26,254 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-03-26 03:42:29,496 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21951.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:42:46,661 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21968.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:42:53,026 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21969.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:42:53,748 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21970.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:43:02,052 INFO [finetune.py:976] (4/7) Epoch 4, batch 4800, loss[loss=0.2231, simple_loss=0.2772, pruned_loss=0.08457, over 4699.00 frames. ], tot_loss[loss=0.2288, simple_loss=0.2843, pruned_loss=0.08662, over 953730.82 frames. ], batch size: 23, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:43:02,777 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21984.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:43:34,457 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22016.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:43:37,302 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.143e+02 1.769e+02 1.987e+02 2.631e+02 5.032e+02, threshold=3.974e+02, percent-clipped=2.0 +2023-03-26 03:43:44,128 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22031.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:43:44,662 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22032.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:43:45,185 INFO [finetune.py:976] (4/7) Epoch 4, batch 4850, loss[loss=0.1909, simple_loss=0.2457, pruned_loss=0.06803, over 4750.00 frames. ], tot_loss[loss=0.2306, simple_loss=0.2874, pruned_loss=0.08692, over 953643.77 frames. ], batch size: 23, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:44:19,577 INFO [finetune.py:976] (4/7) Epoch 4, batch 4900, loss[loss=0.2438, simple_loss=0.2948, pruned_loss=0.09638, over 4091.00 frames. ], tot_loss[loss=0.2324, simple_loss=0.2891, pruned_loss=0.08781, over 952647.10 frames. ], batch size: 65, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:45:00,547 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22119.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:45:01,044 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.093e+02 1.764e+02 2.232e+02 2.515e+02 4.523e+02, threshold=4.464e+02, percent-clipped=3.0 +2023-03-26 03:45:18,652 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22130.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:45:20,458 INFO [finetune.py:976] (4/7) Epoch 4, batch 4950, loss[loss=0.228, simple_loss=0.2868, pruned_loss=0.08457, over 4811.00 frames. ], tot_loss[loss=0.2324, simple_loss=0.2896, pruned_loss=0.0876, over 951572.04 frames. ], batch size: 39, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:45:21,775 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5938, 1.4679, 2.0335, 3.1307, 2.2054, 2.1148, 0.9279, 2.4465], + device='cuda:4'), covar=tensor([0.1839, 0.1559, 0.1269, 0.0597, 0.0848, 0.1616, 0.1925, 0.0684], + device='cuda:4'), in_proj_covar=tensor([0.0104, 0.0120, 0.0138, 0.0168, 0.0105, 0.0145, 0.0131, 0.0106], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 03:45:49,690 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7048, 1.4918, 1.4216, 1.5084, 1.8418, 1.8875, 1.6669, 1.3391], + device='cuda:4'), covar=tensor([0.0274, 0.0361, 0.0530, 0.0347, 0.0212, 0.0398, 0.0359, 0.0466], + device='cuda:4'), in_proj_covar=tensor([0.0086, 0.0114, 0.0138, 0.0118, 0.0105, 0.0100, 0.0092, 0.0109], + device='cuda:4'), out_proj_covar=tensor([6.7145e-05, 9.0111e-05, 1.1114e-04, 9.3564e-05, 8.3453e-05, 7.4348e-05, + 7.0399e-05, 8.5395e-05], device='cuda:4') +2023-03-26 03:46:12,777 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22176.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:46:19,507 INFO [finetune.py:976] (4/7) Epoch 4, batch 5000, loss[loss=0.1988, simple_loss=0.2524, pruned_loss=0.07262, over 4748.00 frames. ], tot_loss[loss=0.2308, simple_loss=0.2876, pruned_loss=0.08698, over 952679.15 frames. ], batch size: 27, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:46:24,517 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22191.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:46:35,051 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22206.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:46:40,542 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22215.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:46:43,479 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.178e+02 1.598e+02 2.028e+02 2.482e+02 4.524e+02, threshold=4.056e+02, percent-clipped=1.0 +2023-03-26 03:46:59,738 INFO [finetune.py:976] (4/7) Epoch 4, batch 5050, loss[loss=0.1968, simple_loss=0.2643, pruned_loss=0.06458, over 4795.00 frames. ], tot_loss[loss=0.2271, simple_loss=0.2838, pruned_loss=0.08521, over 954470.66 frames. ], batch size: 51, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:47:02,281 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22237.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:47:02,291 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22237.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:47:16,634 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22251.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:47:26,346 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1668, 3.6327, 3.8295, 4.0076, 3.9485, 3.7175, 4.2380, 1.5174], + device='cuda:4'), covar=tensor([0.0659, 0.0738, 0.0773, 0.0802, 0.0999, 0.1113, 0.0589, 0.4487], + device='cuda:4'), in_proj_covar=tensor([0.0358, 0.0243, 0.0275, 0.0292, 0.0336, 0.0283, 0.0304, 0.0300], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:47:28,156 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22269.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:47:32,950 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22276.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:47:40,348 INFO [finetune.py:976] (4/7) Epoch 4, batch 5100, loss[loss=0.2185, simple_loss=0.2689, pruned_loss=0.08407, over 4923.00 frames. ], tot_loss[loss=0.2245, simple_loss=0.2805, pruned_loss=0.08426, over 954051.92 frames. ], batch size: 38, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:47:49,915 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22298.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:47:50,456 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22299.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:47:51,706 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7068, 1.6502, 1.8265, 1.8916, 1.8480, 3.3084, 1.6089, 1.8334], + device='cuda:4'), covar=tensor([0.0943, 0.1636, 0.1024, 0.0954, 0.1456, 0.0301, 0.1317, 0.1453], + device='cuda:4'), in_proj_covar=tensor([0.0079, 0.0082, 0.0078, 0.0080, 0.0093, 0.0084, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 03:48:03,281 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22317.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:48:05,002 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.872e+01 1.620e+02 1.853e+02 2.165e+02 3.345e+02, threshold=3.706e+02, percent-clipped=0.0 +2023-03-26 03:48:08,733 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22326.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:48:13,379 INFO [finetune.py:976] (4/7) Epoch 4, batch 5150, loss[loss=0.225, simple_loss=0.2813, pruned_loss=0.08438, over 4841.00 frames. ], tot_loss[loss=0.2243, simple_loss=0.28, pruned_loss=0.08437, over 954197.74 frames. ], batch size: 25, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:48:49,145 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5457, 1.4310, 1.4785, 1.5077, 1.0513, 3.5433, 1.3830, 1.8628], + device='cuda:4'), covar=tensor([0.3610, 0.2483, 0.2159, 0.2380, 0.2030, 0.0167, 0.2756, 0.1437], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0112, 0.0116, 0.0119, 0.0115, 0.0096, 0.0100, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 03:48:51,907 INFO [finetune.py:976] (4/7) Epoch 4, batch 5200, loss[loss=0.2787, simple_loss=0.3419, pruned_loss=0.1078, over 4843.00 frames. ], tot_loss[loss=0.2266, simple_loss=0.283, pruned_loss=0.0851, over 954152.65 frames. ], batch size: 47, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:48:53,220 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22385.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:49:04,596 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2916, 2.1912, 1.6990, 0.8727, 1.8695, 1.7311, 1.5713, 1.9158], + device='cuda:4'), covar=tensor([0.0895, 0.0832, 0.1860, 0.2302, 0.1540, 0.2582, 0.2396, 0.0988], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0201, 0.0204, 0.0192, 0.0220, 0.0210, 0.0222, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:49:26,403 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22419.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:49:26,885 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.279e+02 1.811e+02 2.155e+02 2.737e+02 4.498e+02, threshold=4.310e+02, percent-clipped=4.0 +2023-03-26 03:49:40,609 INFO [finetune.py:976] (4/7) Epoch 4, batch 5250, loss[loss=0.2461, simple_loss=0.3018, pruned_loss=0.09522, over 4811.00 frames. ], tot_loss[loss=0.2291, simple_loss=0.2857, pruned_loss=0.08625, over 954283.50 frames. ], batch size: 39, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:49:54,894 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22446.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:50:18,143 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22467.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:50:27,826 INFO [finetune.py:976] (4/7) Epoch 4, batch 5300, loss[loss=0.2293, simple_loss=0.2879, pruned_loss=0.08539, over 4886.00 frames. ], tot_loss[loss=0.232, simple_loss=0.2885, pruned_loss=0.08775, over 953562.92 frames. ], batch size: 35, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:50:30,462 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22486.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:50:49,037 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22506.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:51:10,353 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.196e+02 1.644e+02 2.088e+02 2.525e+02 4.526e+02, threshold=4.176e+02, percent-clipped=1.0 +2023-03-26 03:51:19,333 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.88 vs. limit=5.0 +2023-03-26 03:51:29,049 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22532.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:51:29,575 INFO [finetune.py:976] (4/7) Epoch 4, batch 5350, loss[loss=0.1895, simple_loss=0.2453, pruned_loss=0.06683, over 4789.00 frames. ], tot_loss[loss=0.2315, simple_loss=0.2884, pruned_loss=0.08727, over 953278.02 frames. ], batch size: 25, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:51:43,431 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22554.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:52:10,734 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22571.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:52:20,638 INFO [finetune.py:976] (4/7) Epoch 4, batch 5400, loss[loss=0.207, simple_loss=0.2587, pruned_loss=0.07761, over 4840.00 frames. ], tot_loss[loss=0.2285, simple_loss=0.2856, pruned_loss=0.08571, over 954200.39 frames. ], batch size: 30, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:52:21,978 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0878, 1.9297, 1.5266, 1.9546, 1.9735, 1.7431, 2.3031, 2.0527], + device='cuda:4'), covar=tensor([0.1697, 0.3281, 0.3898, 0.3533, 0.3114, 0.1880, 0.4648, 0.2437], + device='cuda:4'), in_proj_covar=tensor([0.0165, 0.0194, 0.0236, 0.0254, 0.0226, 0.0187, 0.0209, 0.0188], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:52:27,246 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22593.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:52:45,273 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.025e+02 1.598e+02 1.961e+02 2.261e+02 4.832e+02, threshold=3.922e+02, percent-clipped=2.0 +2023-03-26 03:52:48,830 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5132, 2.3600, 1.9327, 2.6145, 2.5803, 2.0552, 2.9976, 2.4709], + device='cuda:4'), covar=tensor([0.1826, 0.3676, 0.4292, 0.3988, 0.3025, 0.2118, 0.3932, 0.2558], + device='cuda:4'), in_proj_covar=tensor([0.0166, 0.0194, 0.0236, 0.0254, 0.0226, 0.0187, 0.0209, 0.0189], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:52:50,436 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22626.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:52:54,549 INFO [finetune.py:976] (4/7) Epoch 4, batch 5450, loss[loss=0.1638, simple_loss=0.2283, pruned_loss=0.04961, over 4752.00 frames. ], tot_loss[loss=0.2248, simple_loss=0.2815, pruned_loss=0.08402, over 954667.21 frames. ], batch size: 26, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:52:58,341 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22639.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:53:01,386 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5948, 1.3882, 1.3354, 1.5792, 1.7847, 1.5750, 0.9267, 1.3449], + device='cuda:4'), covar=tensor([0.2340, 0.2384, 0.2113, 0.1987, 0.1662, 0.1259, 0.3177, 0.2016], + device='cuda:4'), in_proj_covar=tensor([0.0231, 0.0209, 0.0198, 0.0184, 0.0234, 0.0174, 0.0215, 0.0187], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:53:08,766 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1807, 3.6114, 3.8039, 4.0620, 3.9370, 3.6601, 4.2689, 1.5368], + device='cuda:4'), covar=tensor([0.0765, 0.0848, 0.0956, 0.0934, 0.1156, 0.1610, 0.0748, 0.4620], + device='cuda:4'), in_proj_covar=tensor([0.0360, 0.0244, 0.0276, 0.0294, 0.0340, 0.0286, 0.0307, 0.0300], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:53:19,417 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9990, 1.8308, 1.4202, 1.8017, 1.7174, 1.6308, 1.6603, 2.5245], + device='cuda:4'), covar=tensor([0.8119, 0.8460, 0.6792, 0.8441, 0.7078, 0.4673, 0.8750, 0.2924], + device='cuda:4'), in_proj_covar=tensor([0.0278, 0.0254, 0.0220, 0.0284, 0.0237, 0.0198, 0.0242, 0.0193], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:53:20,558 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7037, 3.6214, 3.4882, 1.6121, 3.8041, 2.8096, 0.8585, 2.5333], + device='cuda:4'), covar=tensor([0.2828, 0.1949, 0.1461, 0.3335, 0.0886, 0.0941, 0.4518, 0.1635], + device='cuda:4'), in_proj_covar=tensor([0.0154, 0.0171, 0.0161, 0.0128, 0.0155, 0.0122, 0.0145, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 03:53:25,492 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3553, 2.0858, 1.9157, 1.6153, 2.3261, 2.8106, 2.4874, 2.0510], + device='cuda:4'), covar=tensor([0.0272, 0.0472, 0.0495, 0.0452, 0.0334, 0.0421, 0.0332, 0.0406], + device='cuda:4'), in_proj_covar=tensor([0.0087, 0.0115, 0.0140, 0.0120, 0.0106, 0.0101, 0.0093, 0.0110], + device='cuda:4'), out_proj_covar=tensor([6.8134e-05, 9.0740e-05, 1.1223e-04, 9.4949e-05, 8.4329e-05, 7.5349e-05, + 7.1088e-05, 8.6445e-05], device='cuda:4') +2023-03-26 03:53:30,708 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22674.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:53:37,508 INFO [finetune.py:976] (4/7) Epoch 4, batch 5500, loss[loss=0.2553, simple_loss=0.298, pruned_loss=0.1063, over 4838.00 frames. ], tot_loss[loss=0.2234, simple_loss=0.2792, pruned_loss=0.08376, over 955445.37 frames. ], batch size: 47, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:53:39,395 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4837, 1.5724, 1.9161, 1.9152, 1.8457, 3.6088, 1.4732, 1.7418], + device='cuda:4'), covar=tensor([0.1034, 0.1778, 0.1085, 0.1044, 0.1411, 0.0229, 0.1456, 0.1649], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0082, 0.0078, 0.0080, 0.0093, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 03:53:48,430 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22700.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 03:54:00,992 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.350e+02 1.787e+02 2.088e+02 2.580e+02 6.017e+02, threshold=4.176e+02, percent-clipped=5.0 +2023-03-26 03:54:16,190 INFO [finetune.py:976] (4/7) Epoch 4, batch 5550, loss[loss=0.2329, simple_loss=0.2823, pruned_loss=0.09177, over 4877.00 frames. ], tot_loss[loss=0.2259, simple_loss=0.2812, pruned_loss=0.08529, over 954354.32 frames. ], batch size: 31, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:54:23,409 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22741.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:54:27,876 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.96 vs. limit=5.0 +2023-03-26 03:54:55,504 INFO [finetune.py:976] (4/7) Epoch 4, batch 5600, loss[loss=0.2139, simple_loss=0.2629, pruned_loss=0.08241, over 4689.00 frames. ], tot_loss[loss=0.2278, simple_loss=0.284, pruned_loss=0.08577, over 955634.05 frames. ], batch size: 23, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:54:57,305 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22786.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:55:01,955 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22794.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:55:19,050 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6265, 1.1235, 0.8875, 1.5606, 2.0609, 1.1096, 1.3668, 1.5951], + device='cuda:4'), covar=tensor([0.1683, 0.2248, 0.2252, 0.1297, 0.2046, 0.2149, 0.1561, 0.2052], + device='cuda:4'), in_proj_covar=tensor([0.0092, 0.0099, 0.0117, 0.0093, 0.0125, 0.0097, 0.0101, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 03:55:28,583 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.180e+02 1.750e+02 2.147e+02 2.459e+02 4.993e+02, threshold=4.295e+02, percent-clipped=1.0 +2023-03-26 03:55:31,579 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22825.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:55:40,681 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22832.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:55:41,206 INFO [finetune.py:976] (4/7) Epoch 4, batch 5650, loss[loss=0.1708, simple_loss=0.2375, pruned_loss=0.05206, over 4729.00 frames. ], tot_loss[loss=0.2305, simple_loss=0.2876, pruned_loss=0.08666, over 956209.72 frames. ], batch size: 23, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:55:41,958 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22834.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:56:06,174 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22855.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:56:21,392 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22871.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:56:31,135 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22880.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:56:33,111 INFO [finetune.py:976] (4/7) Epoch 4, batch 5700, loss[loss=0.1521, simple_loss=0.2067, pruned_loss=0.04879, over 4563.00 frames. ], tot_loss[loss=0.2269, simple_loss=0.2828, pruned_loss=0.08547, over 936887.69 frames. ], batch size: 20, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:56:34,969 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22886.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:56:41,791 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22893.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:57:20,677 INFO [finetune.py:976] (4/7) Epoch 5, batch 0, loss[loss=0.2779, simple_loss=0.3292, pruned_loss=0.1133, over 4817.00 frames. ], tot_loss[loss=0.2779, simple_loss=0.3292, pruned_loss=0.1133, over 4817.00 frames. ], batch size: 38, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:57:20,678 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 03:57:30,738 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4664, 1.1756, 1.3061, 1.2761, 1.6676, 1.6103, 1.4077, 1.2677], + device='cuda:4'), covar=tensor([0.0314, 0.0335, 0.0549, 0.0346, 0.0263, 0.0383, 0.0358, 0.0415], + device='cuda:4'), in_proj_covar=tensor([0.0086, 0.0115, 0.0138, 0.0120, 0.0105, 0.0101, 0.0092, 0.0110], + device='cuda:4'), out_proj_covar=tensor([6.7570e-05, 9.0440e-05, 1.1130e-04, 9.4555e-05, 8.3660e-05, 7.5062e-05, + 7.0309e-05, 8.6112e-05], device='cuda:4') +2023-03-26 03:57:37,527 INFO [finetune.py:1010] (4/7) Epoch 5, validation: loss=0.1701, simple_loss=0.2413, pruned_loss=0.0494, over 2265189.00 frames. +2023-03-26 03:57:37,528 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6301MB +2023-03-26 03:57:48,823 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22919.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:57:49,349 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.844e+01 1.616e+02 1.840e+02 2.309e+02 3.969e+02, threshold=3.680e+02, percent-clipped=0.0 +2023-03-26 03:58:02,584 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=22941.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:58:12,868 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6700, 3.5723, 3.3892, 1.4355, 3.6317, 2.7691, 0.7087, 2.5155], + device='cuda:4'), covar=tensor([0.2666, 0.1545, 0.1652, 0.3307, 0.1018, 0.0954, 0.4427, 0.1319], + device='cuda:4'), in_proj_covar=tensor([0.0154, 0.0170, 0.0161, 0.0127, 0.0154, 0.0121, 0.0144, 0.0121], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 03:58:12,943 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7375, 1.6249, 1.4392, 1.7087, 2.1145, 1.6951, 1.2675, 1.3742], + device='cuda:4'), covar=tensor([0.2144, 0.2170, 0.1933, 0.1737, 0.1836, 0.1257, 0.2918, 0.1843], + device='cuda:4'), in_proj_covar=tensor([0.0233, 0.0210, 0.0198, 0.0185, 0.0236, 0.0175, 0.0215, 0.0188], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 03:58:13,501 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6311, 1.4097, 1.8219, 1.9245, 1.6106, 3.2840, 1.3538, 1.5366], + device='cuda:4'), covar=tensor([0.0971, 0.1832, 0.1261, 0.1028, 0.1650, 0.0342, 0.1616, 0.1875], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0081, 0.0077, 0.0080, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 03:58:29,020 INFO [finetune.py:976] (4/7) Epoch 5, batch 50, loss[loss=0.2025, simple_loss=0.2653, pruned_loss=0.06987, over 4777.00 frames. ], tot_loss[loss=0.235, simple_loss=0.2895, pruned_loss=0.09026, over 214996.28 frames. ], batch size: 29, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:59:00,324 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4750, 1.4794, 1.6948, 1.7938, 1.6063, 3.3440, 1.2851, 1.5528], + device='cuda:4'), covar=tensor([0.1063, 0.1880, 0.1111, 0.1073, 0.1731, 0.0252, 0.1587, 0.1680], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0082, 0.0078, 0.0080, 0.0093, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 03:59:05,723 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22995.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 03:59:17,859 INFO [finetune.py:976] (4/7) Epoch 5, batch 100, loss[loss=0.2006, simple_loss=0.2407, pruned_loss=0.08029, over 4010.00 frames. ], tot_loss[loss=0.224, simple_loss=0.2793, pruned_loss=0.08434, over 378324.89 frames. ], batch size: 17, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 03:59:23,736 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.764e+02 2.029e+02 2.456e+02 6.922e+02, threshold=4.057e+02, percent-clipped=5.0 +2023-03-26 03:59:29,847 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7123, 3.6633, 3.4033, 1.7021, 3.7616, 2.7550, 0.7351, 2.5487], + device='cuda:4'), covar=tensor([0.2232, 0.1734, 0.1495, 0.3082, 0.0897, 0.0982, 0.4413, 0.1485], + device='cuda:4'), in_proj_covar=tensor([0.0154, 0.0170, 0.0161, 0.0127, 0.0154, 0.0121, 0.0145, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 03:59:36,987 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23041.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 03:59:51,327 INFO [finetune.py:976] (4/7) Epoch 5, batch 150, loss[loss=0.2097, simple_loss=0.2711, pruned_loss=0.07413, over 4903.00 frames. ], tot_loss[loss=0.2221, simple_loss=0.2764, pruned_loss=0.0839, over 504852.69 frames. ], batch size: 37, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 04:00:08,778 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23089.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:00:28,367 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23108.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:00:30,575 INFO [finetune.py:976] (4/7) Epoch 5, batch 200, loss[loss=0.1829, simple_loss=0.2465, pruned_loss=0.05967, over 4900.00 frames. ], tot_loss[loss=0.218, simple_loss=0.2731, pruned_loss=0.08149, over 603548.82 frames. ], batch size: 32, lr: 3.95e-03, grad_scale: 64.0 +2023-03-26 04:00:42,592 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.047e+02 1.662e+02 1.994e+02 2.595e+02 4.858e+02, threshold=3.989e+02, percent-clipped=3.0 +2023-03-26 04:01:01,382 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23150.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:01:09,514 INFO [finetune.py:976] (4/7) Epoch 5, batch 250, loss[loss=0.2318, simple_loss=0.2889, pruned_loss=0.08738, over 4928.00 frames. ], tot_loss[loss=0.2193, simple_loss=0.2759, pruned_loss=0.08131, over 682735.14 frames. ], batch size: 38, lr: 3.95e-03, grad_scale: 64.0 +2023-03-26 04:01:23,283 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23169.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:01:31,047 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23181.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:02:00,762 INFO [finetune.py:976] (4/7) Epoch 5, batch 300, loss[loss=0.2436, simple_loss=0.311, pruned_loss=0.08806, over 4896.00 frames. ], tot_loss[loss=0.2235, simple_loss=0.2815, pruned_loss=0.08279, over 744411.38 frames. ], batch size: 35, lr: 3.95e-03, grad_scale: 64.0 +2023-03-26 04:02:11,927 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23217.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:02:20,054 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.235e+02 1.736e+02 2.142e+02 2.597e+02 5.294e+02, threshold=4.284e+02, percent-clipped=3.0 +2023-03-26 04:02:31,236 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23229.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:02:58,880 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4251, 1.2136, 1.3193, 1.3723, 1.6496, 1.5935, 1.4448, 1.2458], + device='cuda:4'), covar=tensor([0.0320, 0.0324, 0.0527, 0.0285, 0.0237, 0.0340, 0.0294, 0.0383], + device='cuda:4'), in_proj_covar=tensor([0.0087, 0.0114, 0.0138, 0.0119, 0.0105, 0.0101, 0.0091, 0.0110], + device='cuda:4'), out_proj_covar=tensor([6.8083e-05, 9.0113e-05, 1.1118e-04, 9.4212e-05, 8.3568e-05, 7.5368e-05, + 7.0088e-05, 8.6010e-05], device='cuda:4') +2023-03-26 04:03:03,005 INFO [finetune.py:976] (4/7) Epoch 5, batch 350, loss[loss=0.229, simple_loss=0.285, pruned_loss=0.08654, over 4915.00 frames. ], tot_loss[loss=0.2277, simple_loss=0.2852, pruned_loss=0.08512, over 792517.47 frames. ], batch size: 38, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 04:03:20,914 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23278.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:03:34,243 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23290.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:03:41,699 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23295.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:03:49,946 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4563, 1.3053, 1.3671, 1.3502, 0.8981, 2.8855, 1.0673, 1.5387], + device='cuda:4'), covar=tensor([0.3271, 0.2418, 0.2107, 0.2383, 0.2117, 0.0228, 0.2901, 0.1365], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0112, 0.0116, 0.0120, 0.0116, 0.0097, 0.0100, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 04:03:51,625 INFO [finetune.py:976] (4/7) Epoch 5, batch 400, loss[loss=0.2085, simple_loss=0.2815, pruned_loss=0.06774, over 4916.00 frames. ], tot_loss[loss=0.2287, simple_loss=0.2864, pruned_loss=0.08551, over 828621.15 frames. ], batch size: 38, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 04:03:58,184 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.903e+01 1.731e+02 2.116e+02 2.565e+02 5.981e+02, threshold=4.232e+02, percent-clipped=1.0 +2023-03-26 04:03:59,475 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9945, 1.5171, 1.7780, 1.8086, 1.6213, 1.6700, 1.7345, 1.6590], + device='cuda:4'), covar=tensor([0.7110, 0.9713, 0.7556, 0.9380, 0.9516, 0.7504, 1.1748, 0.7069], + device='cuda:4'), in_proj_covar=tensor([0.0230, 0.0252, 0.0258, 0.0261, 0.0242, 0.0220, 0.0278, 0.0224], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:04:08,212 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8386, 1.6386, 1.4434, 1.4859, 1.5962, 1.6246, 1.5676, 2.2412], + device='cuda:4'), covar=tensor([0.8143, 0.8290, 0.6294, 0.7651, 0.6507, 0.4614, 0.7303, 0.3127], + device='cuda:4'), in_proj_covar=tensor([0.0279, 0.0253, 0.0219, 0.0283, 0.0236, 0.0198, 0.0241, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:04:13,595 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23343.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:04:24,904 INFO [finetune.py:976] (4/7) Epoch 5, batch 450, loss[loss=0.2086, simple_loss=0.2739, pruned_loss=0.07164, over 4821.00 frames. ], tot_loss[loss=0.2267, simple_loss=0.2843, pruned_loss=0.08452, over 854782.15 frames. ], batch size: 39, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 04:04:54,678 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7547, 1.2697, 0.9741, 1.7048, 2.1050, 1.3374, 1.5522, 1.7357], + device='cuda:4'), covar=tensor([0.1934, 0.2777, 0.2501, 0.1474, 0.2213, 0.2538, 0.1782, 0.2508], + device='cuda:4'), in_proj_covar=tensor([0.0092, 0.0099, 0.0117, 0.0093, 0.0124, 0.0097, 0.0101, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 04:05:10,555 INFO [finetune.py:976] (4/7) Epoch 5, batch 500, loss[loss=0.2167, simple_loss=0.2676, pruned_loss=0.08296, over 4916.00 frames. ], tot_loss[loss=0.2248, simple_loss=0.2816, pruned_loss=0.08396, over 875786.77 frames. ], batch size: 37, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 04:05:16,627 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.006e+02 1.775e+02 2.029e+02 2.615e+02 5.539e+02, threshold=4.057e+02, percent-clipped=1.0 +2023-03-26 04:05:42,945 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23450.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:05:43,582 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6289, 1.5820, 1.5508, 1.6012, 1.1881, 2.7868, 1.2534, 1.8269], + device='cuda:4'), covar=tensor([0.2916, 0.2071, 0.1688, 0.1971, 0.1709, 0.0277, 0.2639, 0.1082], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0112, 0.0116, 0.0120, 0.0116, 0.0097, 0.0100, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 04:05:52,277 INFO [finetune.py:976] (4/7) Epoch 5, batch 550, loss[loss=0.1722, simple_loss=0.2375, pruned_loss=0.05346, over 4775.00 frames. ], tot_loss[loss=0.2228, simple_loss=0.2792, pruned_loss=0.08322, over 893718.18 frames. ], batch size: 27, lr: 3.95e-03, grad_scale: 32.0 +2023-03-26 04:05:54,203 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23464.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:06:15,747 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23481.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:06:23,755 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23488.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:06:30,290 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23498.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:06:38,164 INFO [finetune.py:976] (4/7) Epoch 5, batch 600, loss[loss=0.2291, simple_loss=0.2902, pruned_loss=0.08397, over 4901.00 frames. ], tot_loss[loss=0.2231, simple_loss=0.2794, pruned_loss=0.08336, over 908672.13 frames. ], batch size: 43, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:06:44,765 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.759e+02 2.043e+02 2.434e+02 4.744e+02, threshold=4.086e+02, percent-clipped=2.0 +2023-03-26 04:06:51,178 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23529.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:07:18,144 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8234, 1.6813, 1.4079, 1.5735, 1.6145, 1.5491, 1.5433, 2.3196], + device='cuda:4'), covar=tensor([0.7511, 0.7923, 0.5920, 0.7618, 0.6495, 0.3920, 0.7569, 0.2568], + device='cuda:4'), in_proj_covar=tensor([0.0279, 0.0254, 0.0219, 0.0283, 0.0236, 0.0198, 0.0241, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:07:18,729 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23549.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 04:07:21,778 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8011, 1.6898, 1.5225, 1.8588, 2.2388, 1.8714, 1.3526, 1.4454], + device='cuda:4'), covar=tensor([0.2523, 0.2389, 0.2151, 0.1931, 0.2146, 0.1287, 0.3058, 0.2075], + device='cuda:4'), in_proj_covar=tensor([0.0231, 0.0208, 0.0197, 0.0183, 0.0234, 0.0173, 0.0213, 0.0186], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:07:24,194 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0629, 1.7745, 1.8402, 0.8915, 2.0065, 2.3119, 1.9588, 1.7908], + device='cuda:4'), covar=tensor([0.1076, 0.0773, 0.0662, 0.0841, 0.0521, 0.0655, 0.0521, 0.0738], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0157, 0.0118, 0.0136, 0.0132, 0.0122, 0.0148, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.7685e-05, 1.1674e-04, 8.5797e-05, 9.9650e-05, 9.5370e-05, 9.0146e-05, + 1.1007e-04, 1.0755e-04], device='cuda:4') +2023-03-26 04:07:25,886 INFO [finetune.py:976] (4/7) Epoch 5, batch 650, loss[loss=0.2419, simple_loss=0.3177, pruned_loss=0.08305, over 4738.00 frames. ], tot_loss[loss=0.2267, simple_loss=0.2835, pruned_loss=0.08496, over 919419.56 frames. ], batch size: 59, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:07:33,760 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23573.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:07:43,054 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23585.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:08:14,419 INFO [finetune.py:976] (4/7) Epoch 5, batch 700, loss[loss=0.3013, simple_loss=0.3516, pruned_loss=0.1255, over 4808.00 frames. ], tot_loss[loss=0.2286, simple_loss=0.2857, pruned_loss=0.0858, over 926503.07 frames. ], batch size: 39, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:08:30,908 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.701e+02 2.127e+02 2.576e+02 5.648e+02, threshold=4.253e+02, percent-clipped=2.0 +2023-03-26 04:08:31,081 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1515, 1.9522, 1.6533, 2.2551, 2.0214, 1.7904, 1.9158, 2.9613], + device='cuda:4'), covar=tensor([0.7843, 0.9678, 0.6729, 0.8410, 0.7397, 0.4779, 0.9199, 0.2670], + device='cuda:4'), in_proj_covar=tensor([0.0278, 0.0252, 0.0218, 0.0282, 0.0234, 0.0197, 0.0240, 0.0193], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:09:25,211 INFO [finetune.py:976] (4/7) Epoch 5, batch 750, loss[loss=0.1854, simple_loss=0.2348, pruned_loss=0.06795, over 4278.00 frames. ], tot_loss[loss=0.23, simple_loss=0.2867, pruned_loss=0.08665, over 932704.09 frames. ], batch size: 18, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:09:51,354 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.81 vs. limit=5.0 +2023-03-26 04:10:02,098 INFO [finetune.py:976] (4/7) Epoch 5, batch 800, loss[loss=0.2287, simple_loss=0.2642, pruned_loss=0.09661, over 4271.00 frames. ], tot_loss[loss=0.2273, simple_loss=0.2844, pruned_loss=0.08507, over 935982.17 frames. ], batch size: 18, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:10:05,282 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.4594, 1.5908, 1.6766, 0.9448, 1.5657, 1.8867, 1.7813, 1.4313], + device='cuda:4'), covar=tensor([0.0970, 0.0565, 0.0440, 0.0564, 0.0429, 0.0417, 0.0295, 0.0616], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0157, 0.0118, 0.0136, 0.0132, 0.0121, 0.0147, 0.0144], + device='cuda:4'), out_proj_covar=tensor([9.7337e-05, 1.1634e-04, 8.5358e-05, 9.9464e-05, 9.5297e-05, 8.9786e-05, + 1.0927e-04, 1.0697e-04], device='cuda:4') +2023-03-26 04:10:08,708 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.149e+02 1.635e+02 1.954e+02 2.429e+02 4.773e+02, threshold=3.908e+02, percent-clipped=1.0 +2023-03-26 04:10:56,824 INFO [finetune.py:976] (4/7) Epoch 5, batch 850, loss[loss=0.1746, simple_loss=0.2301, pruned_loss=0.05961, over 4834.00 frames. ], tot_loss[loss=0.2259, simple_loss=0.283, pruned_loss=0.08439, over 941111.55 frames. ], batch size: 33, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:11:04,290 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23764.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:11:18,772 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9191, 1.3247, 0.8071, 1.7514, 2.2287, 1.3409, 1.7927, 1.8426], + device='cuda:4'), covar=tensor([0.1470, 0.2092, 0.2329, 0.1203, 0.1862, 0.2315, 0.1364, 0.1886], + device='cuda:4'), in_proj_covar=tensor([0.0092, 0.0099, 0.0117, 0.0094, 0.0125, 0.0098, 0.0101, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 04:11:54,660 INFO [finetune.py:976] (4/7) Epoch 5, batch 900, loss[loss=0.2225, simple_loss=0.2653, pruned_loss=0.08982, over 4813.00 frames. ], tot_loss[loss=0.2234, simple_loss=0.2805, pruned_loss=0.08316, over 946319.80 frames. ], batch size: 25, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:11:55,342 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23812.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:12:00,784 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.642e+02 1.957e+02 2.389e+02 4.840e+02, threshold=3.913e+02, percent-clipped=2.0 +2023-03-26 04:12:03,446 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 04:12:18,886 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7135, 1.8195, 1.7460, 1.0239, 1.9812, 1.8453, 1.8099, 1.6092], + device='cuda:4'), covar=tensor([0.0736, 0.0676, 0.0821, 0.1034, 0.0569, 0.0852, 0.0667, 0.1128], + device='cuda:4'), in_proj_covar=tensor([0.0138, 0.0132, 0.0143, 0.0127, 0.0111, 0.0142, 0.0145, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:12:21,818 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23844.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 04:12:37,547 INFO [finetune.py:976] (4/7) Epoch 5, batch 950, loss[loss=0.1715, simple_loss=0.2379, pruned_loss=0.05253, over 4750.00 frames. ], tot_loss[loss=0.222, simple_loss=0.2782, pruned_loss=0.08293, over 947212.08 frames. ], batch size: 27, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:12:44,927 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23873.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:12:52,632 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23885.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:13:27,665 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4909, 1.5257, 2.0070, 1.9125, 1.8418, 4.1122, 1.3742, 1.7815], + device='cuda:4'), covar=tensor([0.1068, 0.1771, 0.1260, 0.1040, 0.1554, 0.0193, 0.1546, 0.1720], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0082, 0.0078, 0.0080, 0.0093, 0.0084, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 04:13:28,831 INFO [finetune.py:976] (4/7) Epoch 5, batch 1000, loss[loss=0.2959, simple_loss=0.3264, pruned_loss=0.1328, over 4156.00 frames. ], tot_loss[loss=0.2242, simple_loss=0.2808, pruned_loss=0.08378, over 949902.59 frames. ], batch size: 17, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:13:30,091 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7422, 1.4056, 1.0652, 0.3012, 1.2928, 1.4919, 1.3583, 1.4550], + device='cuda:4'), covar=tensor([0.0941, 0.0925, 0.1441, 0.2168, 0.1477, 0.2746, 0.2622, 0.0915], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0200, 0.0202, 0.0190, 0.0216, 0.0208, 0.0220, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:13:38,586 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.702e+02 2.066e+02 2.385e+02 5.722e+02, threshold=4.131e+02, percent-clipped=3.0 +2023-03-26 04:13:38,656 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23921.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:13:47,710 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=23933.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:14:14,858 INFO [finetune.py:976] (4/7) Epoch 5, batch 1050, loss[loss=0.2491, simple_loss=0.2992, pruned_loss=0.09944, over 4805.00 frames. ], tot_loss[loss=0.2259, simple_loss=0.2835, pruned_loss=0.08414, over 951724.49 frames. ], batch size: 25, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:14:40,671 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7993, 1.1939, 0.8020, 1.6110, 1.9992, 1.4312, 1.6111, 1.6390], + device='cuda:4'), covar=tensor([0.1611, 0.2319, 0.2537, 0.1331, 0.2149, 0.2410, 0.1503, 0.2051], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0100, 0.0118, 0.0094, 0.0125, 0.0098, 0.0102, 0.0096], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 04:15:19,193 INFO [finetune.py:976] (4/7) Epoch 5, batch 1100, loss[loss=0.2237, simple_loss=0.2826, pruned_loss=0.08236, over 4736.00 frames. ], tot_loss[loss=0.2254, simple_loss=0.2836, pruned_loss=0.08363, over 952917.76 frames. ], batch size: 54, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:15:28,406 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.244e+02 1.825e+02 2.109e+02 2.589e+02 5.024e+02, threshold=4.219e+02, percent-clipped=4.0 +2023-03-26 04:15:38,219 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24037.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:15:41,482 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.97 vs. limit=5.0 +2023-03-26 04:15:54,484 INFO [finetune.py:976] (4/7) Epoch 5, batch 1150, loss[loss=0.265, simple_loss=0.3227, pruned_loss=0.1036, over 4881.00 frames. ], tot_loss[loss=0.2282, simple_loss=0.286, pruned_loss=0.08523, over 953597.92 frames. ], batch size: 35, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:16:13,041 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1236, 1.9719, 1.6264, 2.1702, 2.1307, 1.7379, 2.5420, 2.1198], + device='cuda:4'), covar=tensor([0.1868, 0.3587, 0.4297, 0.3672, 0.3060, 0.2189, 0.3383, 0.2512], + device='cuda:4'), in_proj_covar=tensor([0.0165, 0.0194, 0.0237, 0.0255, 0.0227, 0.0188, 0.0211, 0.0190], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:16:18,941 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24098.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:16:28,015 INFO [finetune.py:976] (4/7) Epoch 5, batch 1200, loss[loss=0.2618, simple_loss=0.3145, pruned_loss=0.1045, over 4892.00 frames. ], tot_loss[loss=0.2274, simple_loss=0.2848, pruned_loss=0.08499, over 953301.72 frames. ], batch size: 35, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:16:37,231 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.193e+02 1.721e+02 2.129e+02 2.606e+02 7.150e+02, threshold=4.257e+02, percent-clipped=3.0 +2023-03-26 04:16:47,795 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8902, 1.6848, 1.6461, 1.7783, 1.5384, 4.3430, 1.7666, 2.4170], + device='cuda:4'), covar=tensor([0.3316, 0.2303, 0.2028, 0.2164, 0.1639, 0.0101, 0.2578, 0.1261], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0112, 0.0116, 0.0120, 0.0116, 0.0097, 0.0101, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 04:16:48,451 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9126, 1.7457, 1.4558, 1.6912, 1.6695, 1.5984, 1.6105, 2.3758], + device='cuda:4'), covar=tensor([0.7786, 0.8166, 0.6409, 0.8078, 0.7128, 0.4446, 0.7470, 0.2623], + device='cuda:4'), in_proj_covar=tensor([0.0279, 0.0253, 0.0219, 0.0282, 0.0235, 0.0198, 0.0241, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:16:51,894 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=24144.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 04:16:59,626 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.90 vs. limit=5.0 +2023-03-26 04:17:02,374 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5656, 1.3620, 1.3182, 1.4196, 1.8250, 1.7422, 1.5785, 1.3058], + device='cuda:4'), covar=tensor([0.0286, 0.0365, 0.0602, 0.0352, 0.0237, 0.0389, 0.0283, 0.0410], + device='cuda:4'), in_proj_covar=tensor([0.0088, 0.0114, 0.0140, 0.0120, 0.0106, 0.0102, 0.0092, 0.0110], + device='cuda:4'), out_proj_covar=tensor([6.8518e-05, 9.0098e-05, 1.1272e-04, 9.4643e-05, 8.3695e-05, 7.5926e-05, + 7.0531e-05, 8.6393e-05], device='cuda:4') +2023-03-26 04:17:03,456 INFO [finetune.py:976] (4/7) Epoch 5, batch 1250, loss[loss=0.2159, simple_loss=0.2729, pruned_loss=0.07948, over 4765.00 frames. ], tot_loss[loss=0.2243, simple_loss=0.2817, pruned_loss=0.08345, over 954757.00 frames. ], batch size: 23, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:17:08,430 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-26 04:17:28,512 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=24192.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:17:42,424 INFO [finetune.py:976] (4/7) Epoch 5, batch 1300, loss[loss=0.2358, simple_loss=0.2892, pruned_loss=0.09115, over 4854.00 frames. ], tot_loss[loss=0.2215, simple_loss=0.2784, pruned_loss=0.08235, over 954332.86 frames. ], batch size: 44, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:17:56,617 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.607e+02 1.851e+02 2.364e+02 3.844e+02, threshold=3.702e+02, percent-clipped=0.0 +2023-03-26 04:18:34,749 INFO [finetune.py:976] (4/7) Epoch 5, batch 1350, loss[loss=0.2302, simple_loss=0.2822, pruned_loss=0.08907, over 4820.00 frames. ], tot_loss[loss=0.2205, simple_loss=0.2773, pruned_loss=0.08182, over 956764.85 frames. ], batch size: 30, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:19:09,075 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3919, 0.9163, 0.8534, 1.2928, 1.8303, 0.8485, 1.1057, 1.3150], + device='cuda:4'), covar=tensor([0.1620, 0.2389, 0.2017, 0.1368, 0.1997, 0.2285, 0.1736, 0.2029], + device='cuda:4'), in_proj_covar=tensor([0.0092, 0.0098, 0.0117, 0.0093, 0.0124, 0.0097, 0.0101, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 04:19:12,805 INFO [finetune.py:976] (4/7) Epoch 5, batch 1400, loss[loss=0.2633, simple_loss=0.3333, pruned_loss=0.09668, over 4925.00 frames. ], tot_loss[loss=0.2257, simple_loss=0.2825, pruned_loss=0.0845, over 956196.48 frames. ], batch size: 42, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:19:21,620 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.173e+02 1.714e+02 2.138e+02 2.571e+02 4.877e+02, threshold=4.276e+02, percent-clipped=6.0 +2023-03-26 04:19:24,541 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 04:19:56,901 INFO [finetune.py:976] (4/7) Epoch 5, batch 1450, loss[loss=0.1557, simple_loss=0.2113, pruned_loss=0.05003, over 4076.00 frames. ], tot_loss[loss=0.2271, simple_loss=0.2846, pruned_loss=0.08484, over 954763.00 frames. ], batch size: 17, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:20:05,773 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7154, 1.5505, 1.9585, 1.2380, 1.7215, 1.9502, 1.5562, 2.1325], + device='cuda:4'), covar=tensor([0.1242, 0.2162, 0.1451, 0.1865, 0.0973, 0.1272, 0.2484, 0.0856], + device='cuda:4'), in_proj_covar=tensor([0.0203, 0.0203, 0.0200, 0.0194, 0.0181, 0.0221, 0.0213, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:20:20,196 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=24393.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:20:23,901 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0069, 1.9158, 1.8878, 2.1527, 2.5086, 2.0550, 1.7941, 1.7145], + device='cuda:4'), covar=tensor([0.2405, 0.2312, 0.1914, 0.1701, 0.1784, 0.1210, 0.2701, 0.2006], + device='cuda:4'), in_proj_covar=tensor([0.0233, 0.0209, 0.0198, 0.0185, 0.0236, 0.0174, 0.0215, 0.0187], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:20:31,767 INFO [finetune.py:976] (4/7) Epoch 5, batch 1500, loss[loss=0.2552, simple_loss=0.313, pruned_loss=0.09872, over 4811.00 frames. ], tot_loss[loss=0.2285, simple_loss=0.2863, pruned_loss=0.08541, over 956107.89 frames. ], batch size: 40, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:20:32,584 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.53 vs. limit=5.0 +2023-03-26 04:20:38,324 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.776e+02 2.138e+02 2.564e+02 4.291e+02, threshold=4.276e+02, percent-clipped=1.0 +2023-03-26 04:21:13,454 INFO [finetune.py:976] (4/7) Epoch 5, batch 1550, loss[loss=0.2462, simple_loss=0.3021, pruned_loss=0.09519, over 4828.00 frames. ], tot_loss[loss=0.2293, simple_loss=0.2873, pruned_loss=0.08566, over 955667.69 frames. ], batch size: 33, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:21:47,123 INFO [finetune.py:976] (4/7) Epoch 5, batch 1600, loss[loss=0.2313, simple_loss=0.2746, pruned_loss=0.09397, over 4831.00 frames. ], tot_loss[loss=0.227, simple_loss=0.2843, pruned_loss=0.08485, over 955379.29 frames. ], batch size: 30, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:21:58,799 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.205e+02 1.770e+02 2.018e+02 2.552e+02 5.194e+02, threshold=4.037e+02, percent-clipped=4.0 +2023-03-26 04:22:06,108 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4537, 1.3400, 1.5836, 1.7163, 1.3782, 3.1805, 1.1656, 1.4205], + device='cuda:4'), covar=tensor([0.1228, 0.2168, 0.1345, 0.1201, 0.2064, 0.0323, 0.2077, 0.2253], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0081, 0.0077, 0.0079, 0.0092, 0.0083, 0.0085, 0.0078], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 04:22:14,470 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8322, 1.5523, 2.3622, 3.4986, 2.5289, 2.4776, 0.9978, 2.7464], + device='cuda:4'), covar=tensor([0.1736, 0.1567, 0.1200, 0.0539, 0.0765, 0.1395, 0.1988, 0.0764], + device='cuda:4'), in_proj_covar=tensor([0.0104, 0.0119, 0.0137, 0.0167, 0.0103, 0.0143, 0.0129, 0.0105], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 04:22:19,302 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4344, 1.3408, 1.3560, 1.3606, 0.9225, 2.3432, 0.7491, 1.3811], + device='cuda:4'), covar=tensor([0.4146, 0.3071, 0.2389, 0.3001, 0.1923, 0.0441, 0.2774, 0.1338], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0113, 0.0117, 0.0121, 0.0116, 0.0097, 0.0100, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 04:22:21,138 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24547.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:22:33,857 INFO [finetune.py:976] (4/7) Epoch 5, batch 1650, loss[loss=0.2356, simple_loss=0.2714, pruned_loss=0.0999, over 4328.00 frames. ], tot_loss[loss=0.2234, simple_loss=0.2807, pruned_loss=0.08306, over 956180.13 frames. ], batch size: 19, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:22:43,187 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24569.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:22:56,043 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1319, 2.3758, 2.0587, 1.4325, 2.4309, 2.3822, 2.3399, 1.9266], + device='cuda:4'), covar=tensor([0.0550, 0.0451, 0.0736, 0.0907, 0.0454, 0.0586, 0.0515, 0.0977], + device='cuda:4'), in_proj_covar=tensor([0.0140, 0.0134, 0.0145, 0.0129, 0.0112, 0.0144, 0.0147, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:23:17,430 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24608.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:23:22,442 INFO [finetune.py:976] (4/7) Epoch 5, batch 1700, loss[loss=0.2381, simple_loss=0.2899, pruned_loss=0.09319, over 4899.00 frames. ], tot_loss[loss=0.2208, simple_loss=0.2774, pruned_loss=0.08206, over 954943.57 frames. ], batch size: 32, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:23:31,255 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.673e+02 1.915e+02 2.251e+02 4.027e+02, threshold=3.830e+02, percent-clipped=0.0 +2023-03-26 04:23:42,124 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24630.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:23:55,212 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.6766, 3.2900, 3.4680, 3.3778, 3.2153, 3.1241, 3.7861, 1.3120], + device='cuda:4'), covar=tensor([0.1615, 0.1709, 0.1502, 0.2298, 0.2620, 0.2780, 0.1522, 0.7346], + device='cuda:4'), in_proj_covar=tensor([0.0358, 0.0243, 0.0274, 0.0292, 0.0338, 0.0285, 0.0304, 0.0299], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:23:56,960 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2325, 2.4384, 2.1617, 1.6097, 2.5514, 2.4921, 2.3839, 2.0953], + device='cuda:4'), covar=tensor([0.0621, 0.0514, 0.0731, 0.0888, 0.0552, 0.0599, 0.0579, 0.0947], + device='cuda:4'), in_proj_covar=tensor([0.0140, 0.0134, 0.0145, 0.0128, 0.0112, 0.0144, 0.0147, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:24:06,429 INFO [finetune.py:976] (4/7) Epoch 5, batch 1750, loss[loss=0.1788, simple_loss=0.2362, pruned_loss=0.06073, over 4764.00 frames. ], tot_loss[loss=0.2223, simple_loss=0.2788, pruned_loss=0.0829, over 954746.02 frames. ], batch size: 26, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:24:28,020 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=24693.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:24:39,306 INFO [finetune.py:976] (4/7) Epoch 5, batch 1800, loss[loss=0.2427, simple_loss=0.2976, pruned_loss=0.09394, over 4796.00 frames. ], tot_loss[loss=0.2245, simple_loss=0.2818, pruned_loss=0.08356, over 954295.71 frames. ], batch size: 45, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:24:45,828 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.243e+02 1.800e+02 2.166e+02 2.491e+02 4.201e+02, threshold=4.331e+02, percent-clipped=2.0 +2023-03-26 04:24:45,930 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5962, 1.3646, 1.9122, 3.1103, 2.1395, 2.2724, 0.8217, 2.5310], + device='cuda:4'), covar=tensor([0.1916, 0.1641, 0.1471, 0.0678, 0.0879, 0.1358, 0.2162, 0.0688], + device='cuda:4'), in_proj_covar=tensor([0.0104, 0.0118, 0.0136, 0.0166, 0.0103, 0.0143, 0.0129, 0.0105], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 04:24:59,913 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=24741.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:25:12,946 INFO [finetune.py:976] (4/7) Epoch 5, batch 1850, loss[loss=0.2494, simple_loss=0.3067, pruned_loss=0.09603, over 4897.00 frames. ], tot_loss[loss=0.226, simple_loss=0.2837, pruned_loss=0.08413, over 953084.59 frames. ], batch size: 37, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:25:15,482 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24765.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:25:46,396 INFO [finetune.py:976] (4/7) Epoch 5, batch 1900, loss[loss=0.2656, simple_loss=0.3202, pruned_loss=0.1055, over 4903.00 frames. ], tot_loss[loss=0.2268, simple_loss=0.2846, pruned_loss=0.08447, over 954352.72 frames. ], batch size: 37, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:25:52,457 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.150e+02 1.802e+02 2.061e+02 2.489e+02 6.200e+02, threshold=4.122e+02, percent-clipped=1.0 +2023-03-26 04:25:57,967 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24826.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:26:20,824 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24848.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:26:28,191 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 04:26:29,632 INFO [finetune.py:976] (4/7) Epoch 5, batch 1950, loss[loss=0.2114, simple_loss=0.2739, pruned_loss=0.07449, over 4815.00 frames. ], tot_loss[loss=0.2244, simple_loss=0.2824, pruned_loss=0.08322, over 954295.73 frames. ], batch size: 25, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:26:33,993 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2704, 2.9049, 2.7122, 1.1672, 2.9771, 2.2222, 0.6743, 1.8668], + device='cuda:4'), covar=tensor([0.2562, 0.2439, 0.2039, 0.3922, 0.1501, 0.1229, 0.4520, 0.1765], + device='cuda:4'), in_proj_covar=tensor([0.0157, 0.0173, 0.0166, 0.0130, 0.0156, 0.0124, 0.0147, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 04:26:42,985 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5132, 1.3695, 1.4120, 1.3729, 0.8878, 2.3543, 0.7939, 1.3362], + device='cuda:4'), covar=tensor([0.4161, 0.2957, 0.2380, 0.3102, 0.2043, 0.0418, 0.2644, 0.1367], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0113, 0.0117, 0.0121, 0.0117, 0.0098, 0.0100, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 04:26:56,934 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 04:26:57,655 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=24903.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:27:01,821 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24909.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:27:02,946 INFO [finetune.py:976] (4/7) Epoch 5, batch 2000, loss[loss=0.2224, simple_loss=0.2786, pruned_loss=0.08305, over 4814.00 frames. ], tot_loss[loss=0.2226, simple_loss=0.28, pruned_loss=0.08259, over 954943.03 frames. ], batch size: 41, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:27:13,005 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.611e+02 2.012e+02 2.424e+02 3.709e+02, threshold=4.024e+02, percent-clipped=0.0 +2023-03-26 04:27:15,547 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=24925.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:27:50,326 INFO [finetune.py:976] (4/7) Epoch 5, batch 2050, loss[loss=0.2438, simple_loss=0.2946, pruned_loss=0.09652, over 4875.00 frames. ], tot_loss[loss=0.2191, simple_loss=0.2762, pruned_loss=0.08094, over 955555.45 frames. ], batch size: 34, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:28:12,195 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24995.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:28:23,338 INFO [finetune.py:976] (4/7) Epoch 5, batch 2100, loss[loss=0.2565, simple_loss=0.2881, pruned_loss=0.1124, over 4191.00 frames. ], tot_loss[loss=0.2187, simple_loss=0.276, pruned_loss=0.08066, over 957248.68 frames. ], batch size: 18, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:28:29,967 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7894, 1.6885, 1.3498, 1.5956, 1.5755, 1.4899, 1.5445, 2.3038], + device='cuda:4'), covar=tensor([0.7344, 0.7822, 0.6068, 0.7204, 0.6213, 0.4218, 0.7202, 0.2841], + device='cuda:4'), in_proj_covar=tensor([0.0280, 0.0255, 0.0221, 0.0283, 0.0237, 0.0200, 0.0243, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:28:39,038 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.247e+01 1.663e+02 1.989e+02 2.457e+02 4.446e+02, threshold=3.978e+02, percent-clipped=1.0 +2023-03-26 04:28:51,253 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-26 04:29:06,637 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-03-26 04:29:08,258 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=25056.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 04:29:11,162 INFO [finetune.py:976] (4/7) Epoch 5, batch 2150, loss[loss=0.2009, simple_loss=0.2721, pruned_loss=0.06482, over 4904.00 frames. ], tot_loss[loss=0.2209, simple_loss=0.2792, pruned_loss=0.08128, over 956369.58 frames. ], batch size: 35, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:29:27,070 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2012, 1.7988, 1.9492, 2.0152, 1.7722, 1.8613, 1.9949, 1.8345], + device='cuda:4'), covar=tensor([0.6108, 0.8825, 0.6742, 0.8048, 0.9725, 0.6431, 1.0318, 0.6886], + device='cuda:4'), in_proj_covar=tensor([0.0229, 0.0250, 0.0256, 0.0260, 0.0242, 0.0219, 0.0276, 0.0223], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:29:39,868 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5192, 1.3443, 1.8604, 2.6257, 1.7953, 2.0752, 1.1424, 2.2051], + device='cuda:4'), covar=tensor([0.1734, 0.1777, 0.1284, 0.0874, 0.0900, 0.2653, 0.1698, 0.0746], + device='cuda:4'), in_proj_covar=tensor([0.0104, 0.0119, 0.0136, 0.0166, 0.0103, 0.0143, 0.0129, 0.0104], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 04:29:45,151 INFO [finetune.py:976] (4/7) Epoch 5, batch 2200, loss[loss=0.2539, simple_loss=0.3234, pruned_loss=0.09218, over 4805.00 frames. ], tot_loss[loss=0.2234, simple_loss=0.2822, pruned_loss=0.08231, over 955179.19 frames. ], batch size: 40, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:29:47,757 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7894, 1.6507, 1.4644, 1.4209, 1.5788, 1.5321, 1.5539, 2.3136], + device='cuda:4'), covar=tensor([0.6168, 0.5955, 0.4645, 0.5903, 0.5202, 0.3463, 0.5655, 0.2211], + device='cuda:4'), in_proj_covar=tensor([0.0280, 0.0255, 0.0221, 0.0283, 0.0237, 0.0200, 0.0243, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:29:52,262 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.691e+02 1.983e+02 2.301e+02 4.176e+02, threshold=3.967e+02, percent-clipped=1.0 +2023-03-26 04:29:52,344 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25121.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:29:55,615 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-26 04:30:04,980 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-26 04:30:18,607 INFO [finetune.py:976] (4/7) Epoch 5, batch 2250, loss[loss=0.2232, simple_loss=0.2891, pruned_loss=0.07865, over 4800.00 frames. ], tot_loss[loss=0.2269, simple_loss=0.2851, pruned_loss=0.08438, over 954452.52 frames. ], batch size: 51, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:30:22,248 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9567, 1.6383, 1.7376, 1.9055, 1.6434, 1.6815, 1.8071, 1.6840], + device='cuda:4'), covar=tensor([0.6404, 0.8197, 0.6879, 0.8185, 0.9072, 0.6594, 1.0464, 0.6491], + device='cuda:4'), in_proj_covar=tensor([0.0228, 0.0249, 0.0254, 0.0259, 0.0241, 0.0218, 0.0274, 0.0222], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:30:46,341 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25203.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:30:47,787 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25204.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:30:52,132 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-26 04:30:53,042 INFO [finetune.py:976] (4/7) Epoch 5, batch 2300, loss[loss=0.2433, simple_loss=0.2938, pruned_loss=0.09639, over 4109.00 frames. ], tot_loss[loss=0.2257, simple_loss=0.2846, pruned_loss=0.08343, over 953324.49 frames. ], batch size: 65, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:31:04,693 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7064, 1.2478, 0.9197, 1.5720, 2.0372, 1.2291, 1.5197, 1.7741], + device='cuda:4'), covar=tensor([0.1557, 0.2165, 0.2259, 0.1296, 0.2172, 0.2356, 0.1502, 0.1965], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0098, 0.0116, 0.0093, 0.0124, 0.0097, 0.0101, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 04:31:05,174 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.234e+02 1.840e+02 2.117e+02 2.638e+02 5.911e+02, threshold=4.234e+02, percent-clipped=5.0 +2023-03-26 04:31:13,444 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0759, 2.3311, 2.0266, 1.3653, 2.3684, 2.2219, 2.1195, 1.9111], + device='cuda:4'), covar=tensor([0.0647, 0.0499, 0.0804, 0.1029, 0.0415, 0.0813, 0.0650, 0.0893], + device='cuda:4'), in_proj_covar=tensor([0.0138, 0.0132, 0.0143, 0.0126, 0.0110, 0.0142, 0.0144, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:31:14,023 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25225.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:31:35,993 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25251.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:31:42,892 INFO [finetune.py:976] (4/7) Epoch 5, batch 2350, loss[loss=0.2051, simple_loss=0.2587, pruned_loss=0.07578, over 4795.00 frames. ], tot_loss[loss=0.2225, simple_loss=0.2812, pruned_loss=0.08188, over 953470.41 frames. ], batch size: 51, lr: 3.94e-03, grad_scale: 64.0 +2023-03-26 04:31:51,255 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25273.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:32:04,507 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 04:32:16,769 INFO [finetune.py:976] (4/7) Epoch 5, batch 2400, loss[loss=0.2024, simple_loss=0.2546, pruned_loss=0.07515, over 4701.00 frames. ], tot_loss[loss=0.2205, simple_loss=0.2784, pruned_loss=0.08128, over 953389.16 frames. ], batch size: 23, lr: 3.94e-03, grad_scale: 64.0 +2023-03-26 04:32:23,862 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.140e+02 1.631e+02 1.900e+02 2.318e+02 5.058e+02, threshold=3.799e+02, percent-clipped=1.0 +2023-03-26 04:32:58,160 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25351.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 04:33:04,183 INFO [finetune.py:976] (4/7) Epoch 5, batch 2450, loss[loss=0.2031, simple_loss=0.2576, pruned_loss=0.07428, over 4865.00 frames. ], tot_loss[loss=0.2181, simple_loss=0.2756, pruned_loss=0.08034, over 953889.11 frames. ], batch size: 49, lr: 3.94e-03, grad_scale: 64.0 +2023-03-26 04:33:15,361 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7811, 1.5440, 2.2109, 1.3837, 1.8621, 1.9226, 1.4411, 2.0108], + device='cuda:4'), covar=tensor([0.1486, 0.2098, 0.1467, 0.1996, 0.1088, 0.1718, 0.2721, 0.1180], + device='cuda:4'), in_proj_covar=tensor([0.0207, 0.0207, 0.0203, 0.0197, 0.0185, 0.0225, 0.0218, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:33:48,802 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 04:34:02,100 INFO [finetune.py:976] (4/7) Epoch 5, batch 2500, loss[loss=0.2107, simple_loss=0.2708, pruned_loss=0.07525, over 4817.00 frames. ], tot_loss[loss=0.2186, simple_loss=0.2759, pruned_loss=0.08062, over 951158.80 frames. ], batch size: 40, lr: 3.94e-03, grad_scale: 64.0 +2023-03-26 04:34:18,832 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.080e+02 1.793e+02 2.115e+02 2.620e+02 5.379e+02, threshold=4.229e+02, percent-clipped=6.0 +2023-03-26 04:34:18,939 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25421.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:34:26,162 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=25428.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 04:34:47,668 INFO [finetune.py:976] (4/7) Epoch 5, batch 2550, loss[loss=0.2652, simple_loss=0.327, pruned_loss=0.1017, over 4807.00 frames. ], tot_loss[loss=0.2246, simple_loss=0.2819, pruned_loss=0.08369, over 951769.05 frames. ], batch size: 41, lr: 3.94e-03, grad_scale: 64.0 +2023-03-26 04:34:53,578 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25469.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:35:07,250 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=25489.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 04:35:12,578 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2606, 1.0413, 1.4910, 2.2878, 1.5110, 2.1141, 0.9109, 1.9450], + device='cuda:4'), covar=tensor([0.2089, 0.2016, 0.1404, 0.1023, 0.1090, 0.1484, 0.1795, 0.0889], + device='cuda:4'), in_proj_covar=tensor([0.0104, 0.0119, 0.0136, 0.0166, 0.0103, 0.0143, 0.0128, 0.0105], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 04:35:16,192 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25504.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:35:20,836 INFO [finetune.py:976] (4/7) Epoch 5, batch 2600, loss[loss=0.2349, simple_loss=0.2958, pruned_loss=0.08697, over 4850.00 frames. ], tot_loss[loss=0.2254, simple_loss=0.2829, pruned_loss=0.08399, over 951424.22 frames. ], batch size: 31, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:35:21,561 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6367, 1.4068, 2.2706, 3.2067, 2.2109, 2.3477, 1.0700, 2.4933], + device='cuda:4'), covar=tensor([0.1792, 0.1624, 0.1164, 0.0608, 0.0787, 0.1868, 0.1896, 0.0627], + device='cuda:4'), in_proj_covar=tensor([0.0105, 0.0120, 0.0137, 0.0167, 0.0103, 0.0144, 0.0129, 0.0105], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 04:35:28,041 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.693e+02 2.088e+02 2.425e+02 4.415e+02, threshold=4.177e+02, percent-clipped=1.0 +2023-03-26 04:35:39,514 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-03-26 04:35:48,649 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25552.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:35:54,566 INFO [finetune.py:976] (4/7) Epoch 5, batch 2650, loss[loss=0.2202, simple_loss=0.2694, pruned_loss=0.08549, over 4782.00 frames. ], tot_loss[loss=0.2276, simple_loss=0.2851, pruned_loss=0.08502, over 950414.51 frames. ], batch size: 29, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:36:33,947 INFO [finetune.py:976] (4/7) Epoch 5, batch 2700, loss[loss=0.2028, simple_loss=0.2619, pruned_loss=0.07178, over 4889.00 frames. ], tot_loss[loss=0.2245, simple_loss=0.283, pruned_loss=0.08302, over 951385.93 frames. ], batch size: 35, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:36:50,930 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.205e+02 1.711e+02 2.002e+02 2.331e+02 3.948e+02, threshold=4.004e+02, percent-clipped=0.0 +2023-03-26 04:37:20,934 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7889, 3.9552, 3.7537, 1.7404, 4.0863, 3.0497, 1.0222, 2.6682], + device='cuda:4'), covar=tensor([0.2239, 0.1867, 0.1500, 0.3383, 0.0896, 0.0963, 0.4283, 0.1477], + device='cuda:4'), in_proj_covar=tensor([0.0157, 0.0173, 0.0165, 0.0130, 0.0157, 0.0123, 0.0147, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 04:37:26,223 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25651.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 04:37:32,347 INFO [finetune.py:976] (4/7) Epoch 5, batch 2750, loss[loss=0.2065, simple_loss=0.259, pruned_loss=0.07698, over 4866.00 frames. ], tot_loss[loss=0.2219, simple_loss=0.2796, pruned_loss=0.08215, over 949521.77 frames. ], batch size: 31, lr: 3.94e-03, grad_scale: 32.0 +2023-03-26 04:37:45,311 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.83 vs. limit=5.0 +2023-03-26 04:37:47,304 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1476, 2.4210, 2.3354, 1.6377, 2.4605, 2.3663, 2.3063, 2.0044], + device='cuda:4'), covar=tensor([0.0715, 0.0496, 0.0700, 0.0983, 0.0444, 0.0864, 0.0717, 0.0967], + device='cuda:4'), in_proj_covar=tensor([0.0138, 0.0132, 0.0143, 0.0127, 0.0110, 0.0143, 0.0145, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:37:58,589 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=25699.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:37:59,511 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-26 04:38:06,507 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8384, 1.3180, 0.8134, 1.7442, 2.1386, 1.5277, 1.6381, 1.7470], + device='cuda:4'), covar=tensor([0.1527, 0.2103, 0.2453, 0.1219, 0.2028, 0.2177, 0.1462, 0.2015], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0098, 0.0116, 0.0093, 0.0123, 0.0097, 0.0100, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 04:38:07,651 INFO [finetune.py:976] (4/7) Epoch 5, batch 2800, loss[loss=0.1915, simple_loss=0.244, pruned_loss=0.06955, over 4763.00 frames. ], tot_loss[loss=0.2184, simple_loss=0.2758, pruned_loss=0.08051, over 951720.27 frames. ], batch size: 27, lr: 3.93e-03, grad_scale: 32.0 +2023-03-26 04:38:23,888 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.133e+02 1.611e+02 1.888e+02 2.301e+02 3.388e+02, threshold=3.776e+02, percent-clipped=0.0 +2023-03-26 04:38:33,638 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9907, 1.3440, 1.7938, 1.7891, 1.5929, 1.5770, 1.7109, 1.6398], + device='cuda:4'), covar=tensor([0.5324, 0.8125, 0.6147, 0.7358, 0.7936, 0.6042, 0.8879, 0.6083], + device='cuda:4'), in_proj_covar=tensor([0.0227, 0.0248, 0.0254, 0.0258, 0.0241, 0.0218, 0.0274, 0.0222], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:39:02,536 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-03-26 04:39:03,403 INFO [finetune.py:976] (4/7) Epoch 5, batch 2850, loss[loss=0.1611, simple_loss=0.2263, pruned_loss=0.04794, over 4715.00 frames. ], tot_loss[loss=0.2175, simple_loss=0.2745, pruned_loss=0.08025, over 950452.75 frames. ], batch size: 23, lr: 3.93e-03, grad_scale: 32.0 +2023-03-26 04:39:18,522 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25784.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 04:39:37,534 INFO [finetune.py:976] (4/7) Epoch 5, batch 2900, loss[loss=0.237, simple_loss=0.306, pruned_loss=0.08407, over 4839.00 frames. ], tot_loss[loss=0.2191, simple_loss=0.2762, pruned_loss=0.08102, over 949952.39 frames. ], batch size: 49, lr: 3.93e-03, grad_scale: 32.0 +2023-03-26 04:39:41,960 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1256, 2.6560, 2.1215, 1.8013, 2.6241, 2.5969, 2.3238, 2.1258], + device='cuda:4'), covar=tensor([0.0913, 0.0610, 0.1002, 0.1032, 0.0657, 0.0907, 0.0845, 0.1076], + device='cuda:4'), in_proj_covar=tensor([0.0138, 0.0133, 0.0143, 0.0127, 0.0110, 0.0143, 0.0145, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:39:44,762 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.812e+02 2.065e+02 2.463e+02 5.082e+02, threshold=4.130e+02, percent-clipped=4.0 +2023-03-26 04:40:02,033 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6679, 1.5091, 1.5462, 1.6100, 1.0061, 3.6880, 1.4687, 1.8392], + device='cuda:4'), covar=tensor([0.3466, 0.2518, 0.2058, 0.2353, 0.2063, 0.0158, 0.2624, 0.1386], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0114, 0.0117, 0.0122, 0.0117, 0.0098, 0.0101, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 04:40:09,029 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=25858.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 04:40:10,731 INFO [finetune.py:976] (4/7) Epoch 5, batch 2950, loss[loss=0.2375, simple_loss=0.2994, pruned_loss=0.08782, over 4906.00 frames. ], tot_loss[loss=0.2226, simple_loss=0.2805, pruned_loss=0.08235, over 950206.38 frames. ], batch size: 37, lr: 3.93e-03, grad_scale: 32.0 +2023-03-26 04:40:43,943 INFO [finetune.py:976] (4/7) Epoch 5, batch 3000, loss[loss=0.2247, simple_loss=0.2893, pruned_loss=0.08011, over 4869.00 frames. ], tot_loss[loss=0.2247, simple_loss=0.2827, pruned_loss=0.08336, over 950299.26 frames. ], batch size: 31, lr: 3.93e-03, grad_scale: 32.0 +2023-03-26 04:40:43,943 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 04:40:54,558 INFO [finetune.py:1010] (4/7) Epoch 5, validation: loss=0.1652, simple_loss=0.2371, pruned_loss=0.04667, over 2265189.00 frames. +2023-03-26 04:40:54,558 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6301MB +2023-03-26 04:41:00,085 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=25919.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 04:41:01,809 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.188e+02 1.739e+02 2.096e+02 2.435e+02 4.160e+02, threshold=4.193e+02, percent-clipped=2.0 +2023-03-26 04:41:05,105 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.69 vs. limit=2.0 +2023-03-26 04:41:27,995 INFO [finetune.py:976] (4/7) Epoch 5, batch 3050, loss[loss=0.2457, simple_loss=0.306, pruned_loss=0.09264, over 4880.00 frames. ], tot_loss[loss=0.224, simple_loss=0.2827, pruned_loss=0.08269, over 950233.25 frames. ], batch size: 32, lr: 3.93e-03, grad_scale: 32.0 +2023-03-26 04:41:43,368 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 04:42:08,229 INFO [finetune.py:976] (4/7) Epoch 5, batch 3100, loss[loss=0.1993, simple_loss=0.2539, pruned_loss=0.07234, over 4887.00 frames. ], tot_loss[loss=0.2218, simple_loss=0.2804, pruned_loss=0.08157, over 951036.19 frames. ], batch size: 32, lr: 3.93e-03, grad_scale: 32.0 +2023-03-26 04:42:25,411 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.603e+02 1.879e+02 2.413e+02 4.411e+02, threshold=3.758e+02, percent-clipped=2.0 +2023-03-26 04:43:10,319 INFO [finetune.py:976] (4/7) Epoch 5, batch 3150, loss[loss=0.24, simple_loss=0.2839, pruned_loss=0.09807, over 4830.00 frames. ], tot_loss[loss=0.2186, simple_loss=0.2765, pruned_loss=0.08038, over 950604.37 frames. ], batch size: 25, lr: 3.93e-03, grad_scale: 32.0 +2023-03-26 04:43:27,258 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7883, 1.6857, 1.3161, 1.6032, 1.7121, 1.4457, 2.3582, 1.7235], + device='cuda:4'), covar=tensor([0.1679, 0.2890, 0.4158, 0.3478, 0.3149, 0.1938, 0.2716, 0.2555], + device='cuda:4'), in_proj_covar=tensor([0.0165, 0.0192, 0.0234, 0.0253, 0.0227, 0.0187, 0.0209, 0.0189], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:43:30,920 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26084.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 04:43:42,324 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26102.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:43:49,558 INFO [finetune.py:976] (4/7) Epoch 5, batch 3200, loss[loss=0.2276, simple_loss=0.2824, pruned_loss=0.08633, over 4855.00 frames. ], tot_loss[loss=0.215, simple_loss=0.2727, pruned_loss=0.07863, over 952268.83 frames. ], batch size: 44, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:43:58,350 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.040e+02 1.599e+02 1.999e+02 2.453e+02 4.323e+02, threshold=3.997e+02, percent-clipped=1.0 +2023-03-26 04:44:04,857 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26132.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 04:44:37,849 INFO [finetune.py:976] (4/7) Epoch 5, batch 3250, loss[loss=0.1999, simple_loss=0.2638, pruned_loss=0.06796, over 4761.00 frames. ], tot_loss[loss=0.2155, simple_loss=0.273, pruned_loss=0.07904, over 950117.03 frames. ], batch size: 27, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:44:43,929 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26163.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:44:47,976 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26168.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:45:29,393 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.12 vs. limit=5.0 +2023-03-26 04:45:40,770 INFO [finetune.py:976] (4/7) Epoch 5, batch 3300, loss[loss=0.244, simple_loss=0.2845, pruned_loss=0.1018, over 4232.00 frames. ], tot_loss[loss=0.2199, simple_loss=0.2777, pruned_loss=0.08109, over 950269.24 frames. ], batch size: 18, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:45:47,650 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26214.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 04:46:00,061 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.165e+02 1.721e+02 2.149e+02 2.490e+02 4.939e+02, threshold=4.298e+02, percent-clipped=1.0 +2023-03-26 04:46:06,006 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26229.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:46:19,855 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26250.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:46:29,543 INFO [finetune.py:976] (4/7) Epoch 5, batch 3350, loss[loss=0.2489, simple_loss=0.3118, pruned_loss=0.09298, over 4843.00 frames. ], tot_loss[loss=0.2226, simple_loss=0.2806, pruned_loss=0.08227, over 949343.76 frames. ], batch size: 47, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:46:30,319 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.95 vs. limit=5.0 +2023-03-26 04:47:12,173 INFO [finetune.py:976] (4/7) Epoch 5, batch 3400, loss[loss=0.1729, simple_loss=0.2485, pruned_loss=0.04869, over 4803.00 frames. ], tot_loss[loss=0.223, simple_loss=0.2815, pruned_loss=0.08231, over 949379.95 frames. ], batch size: 51, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:47:12,314 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26311.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:47:19,911 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.624e+02 1.904e+02 2.361e+02 4.543e+02, threshold=3.807e+02, percent-clipped=1.0 +2023-03-26 04:47:53,149 INFO [finetune.py:976] (4/7) Epoch 5, batch 3450, loss[loss=0.2578, simple_loss=0.3024, pruned_loss=0.1066, over 4917.00 frames. ], tot_loss[loss=0.2223, simple_loss=0.2813, pruned_loss=0.08167, over 950697.31 frames. ], batch size: 38, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:48:02,763 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9951, 2.1896, 1.9123, 1.4653, 2.2878, 2.1839, 2.0942, 1.8486], + device='cuda:4'), covar=tensor([0.0741, 0.0584, 0.0860, 0.0984, 0.0484, 0.0759, 0.0708, 0.0996], + device='cuda:4'), in_proj_covar=tensor([0.0139, 0.0134, 0.0145, 0.0129, 0.0111, 0.0144, 0.0147, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:48:02,775 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4518, 2.1852, 2.0046, 0.9896, 2.1202, 1.8652, 1.6156, 1.9331], + device='cuda:4'), covar=tensor([0.0974, 0.0989, 0.1926, 0.2428, 0.1850, 0.2384, 0.2534, 0.1422], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0202, 0.0204, 0.0191, 0.0218, 0.0210, 0.0223, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:48:08,202 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 04:48:27,134 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0167, 1.2859, 0.7728, 1.9466, 2.4419, 1.7020, 1.6266, 1.8861], + device='cuda:4'), covar=tensor([0.1476, 0.2180, 0.2362, 0.1199, 0.1877, 0.2124, 0.1447, 0.1996], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0098, 0.0116, 0.0094, 0.0124, 0.0097, 0.0101, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 04:48:35,618 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7757, 1.8676, 1.5775, 1.4317, 2.2226, 2.1936, 1.9727, 1.8511], + device='cuda:4'), covar=tensor([0.0411, 0.0357, 0.0497, 0.0404, 0.0288, 0.0472, 0.0314, 0.0385], + device='cuda:4'), in_proj_covar=tensor([0.0086, 0.0111, 0.0136, 0.0116, 0.0104, 0.0100, 0.0090, 0.0108], + device='cuda:4'), out_proj_covar=tensor([6.7152e-05, 8.7641e-05, 1.0969e-04, 9.2133e-05, 8.2060e-05, 7.4241e-05, + 6.8915e-05, 8.4308e-05], device='cuda:4') +2023-03-26 04:48:37,958 INFO [finetune.py:976] (4/7) Epoch 5, batch 3500, loss[loss=0.2318, simple_loss=0.2839, pruned_loss=0.08984, over 4828.00 frames. ], tot_loss[loss=0.2204, simple_loss=0.2789, pruned_loss=0.08091, over 950546.88 frames. ], batch size: 38, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:48:43,888 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26414.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:48:45,745 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.3687, 3.0413, 2.7854, 1.5373, 2.9222, 2.4304, 2.1841, 2.5497], + device='cuda:4'), covar=tensor([0.0792, 0.0822, 0.1546, 0.2227, 0.1584, 0.2108, 0.2029, 0.1215], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0202, 0.0204, 0.0191, 0.0217, 0.0210, 0.0222, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:48:51,309 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.234e+02 1.586e+02 1.962e+02 2.229e+02 4.326e+02, threshold=3.925e+02, percent-clipped=1.0 +2023-03-26 04:49:23,749 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26458.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:49:25,515 INFO [finetune.py:976] (4/7) Epoch 5, batch 3550, loss[loss=0.1572, simple_loss=0.2271, pruned_loss=0.04365, over 4900.00 frames. ], tot_loss[loss=0.217, simple_loss=0.2754, pruned_loss=0.07934, over 950257.70 frames. ], batch size: 36, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:49:26,300 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9388, 1.2923, 1.7381, 1.7020, 1.5593, 1.5475, 1.6359, 1.6496], + device='cuda:4'), covar=tensor([0.5763, 0.7753, 0.6253, 0.7686, 0.8086, 0.6519, 0.9025, 0.5936], + device='cuda:4'), in_proj_covar=tensor([0.0230, 0.0248, 0.0256, 0.0259, 0.0242, 0.0219, 0.0275, 0.0224], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:49:34,681 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26475.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:49:38,328 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8483, 2.0310, 1.6451, 1.5295, 2.1834, 2.2528, 2.1584, 1.9408], + device='cuda:4'), covar=tensor([0.0311, 0.0326, 0.0538, 0.0365, 0.0290, 0.0416, 0.0239, 0.0330], + device='cuda:4'), in_proj_covar=tensor([0.0086, 0.0111, 0.0137, 0.0117, 0.0104, 0.0100, 0.0091, 0.0109], + device='cuda:4'), out_proj_covar=tensor([6.7265e-05, 8.7841e-05, 1.0982e-04, 9.2390e-05, 8.2240e-05, 7.4515e-05, + 6.9374e-05, 8.4677e-05], device='cuda:4') +2023-03-26 04:49:50,745 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6205, 1.6342, 2.2413, 3.4324, 2.4810, 2.5115, 1.1713, 2.7272], + device='cuda:4'), covar=tensor([0.1855, 0.1519, 0.1269, 0.0647, 0.0731, 0.1411, 0.1900, 0.0606], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0119, 0.0137, 0.0167, 0.0102, 0.0142, 0.0129, 0.0104], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 04:50:09,378 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.41 vs. limit=5.0 +2023-03-26 04:50:11,928 INFO [finetune.py:976] (4/7) Epoch 5, batch 3600, loss[loss=0.224, simple_loss=0.2672, pruned_loss=0.09038, over 4835.00 frames. ], tot_loss[loss=0.2138, simple_loss=0.2719, pruned_loss=0.07791, over 951003.42 frames. ], batch size: 33, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:50:13,872 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26514.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 04:50:19,781 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.118e+02 1.588e+02 1.920e+02 2.290e+02 3.397e+02, threshold=3.841e+02, percent-clipped=0.0 +2023-03-26 04:50:20,510 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26524.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:50:55,027 INFO [finetune.py:976] (4/7) Epoch 5, batch 3650, loss[loss=0.2678, simple_loss=0.332, pruned_loss=0.1018, over 4840.00 frames. ], tot_loss[loss=0.217, simple_loss=0.2747, pruned_loss=0.07958, over 952645.21 frames. ], batch size: 49, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:50:55,699 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26562.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 04:51:19,229 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 04:51:31,379 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26606.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:51:34,854 INFO [finetune.py:976] (4/7) Epoch 5, batch 3700, loss[loss=0.2331, simple_loss=0.2974, pruned_loss=0.08439, over 4831.00 frames. ], tot_loss[loss=0.2231, simple_loss=0.2808, pruned_loss=0.08266, over 952469.40 frames. ], batch size: 47, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:51:42,594 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.101e+02 1.866e+02 2.222e+02 2.784e+02 4.852e+02, threshold=4.444e+02, percent-clipped=6.0 +2023-03-26 04:52:07,673 INFO [finetune.py:976] (4/7) Epoch 5, batch 3750, loss[loss=0.2246, simple_loss=0.28, pruned_loss=0.08463, over 4819.00 frames. ], tot_loss[loss=0.2256, simple_loss=0.2834, pruned_loss=0.08391, over 952869.03 frames. ], batch size: 30, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:53:00,964 INFO [finetune.py:976] (4/7) Epoch 5, batch 3800, loss[loss=0.2392, simple_loss=0.3117, pruned_loss=0.08337, over 4809.00 frames. ], tot_loss[loss=0.2257, simple_loss=0.2843, pruned_loss=0.08355, over 952250.29 frames. ], batch size: 38, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:53:03,396 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5903, 1.7564, 1.6557, 0.9853, 1.9496, 1.7867, 1.7082, 1.5017], + device='cuda:4'), covar=tensor([0.0796, 0.0718, 0.0836, 0.1129, 0.0548, 0.0829, 0.0794, 0.1209], + device='cuda:4'), in_proj_covar=tensor([0.0138, 0.0133, 0.0143, 0.0128, 0.0111, 0.0143, 0.0145, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:53:08,704 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.710e+02 2.076e+02 2.649e+02 5.488e+02, threshold=4.152e+02, percent-clipped=2.0 +2023-03-26 04:53:15,518 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9312, 1.8265, 1.5343, 1.6159, 2.0271, 1.6733, 2.1321, 1.8585], + device='cuda:4'), covar=tensor([0.1767, 0.2934, 0.4147, 0.3575, 0.2997, 0.2035, 0.3249, 0.2384], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0195, 0.0238, 0.0256, 0.0232, 0.0190, 0.0211, 0.0192], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 04:53:57,421 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26758.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:53:59,695 INFO [finetune.py:976] (4/7) Epoch 5, batch 3850, loss[loss=0.2297, simple_loss=0.2791, pruned_loss=0.0902, over 4722.00 frames. ], tot_loss[loss=0.2238, simple_loss=0.2826, pruned_loss=0.08245, over 953067.24 frames. ], batch size: 59, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:54:11,279 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26770.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:55:00,185 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26806.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:55:03,123 INFO [finetune.py:976] (4/7) Epoch 5, batch 3900, loss[loss=0.2074, simple_loss=0.2493, pruned_loss=0.08281, over 4758.00 frames. ], tot_loss[loss=0.2201, simple_loss=0.2785, pruned_loss=0.08084, over 954626.14 frames. ], batch size: 59, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:55:04,016 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.74 vs. limit=2.0 +2023-03-26 04:55:21,757 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.223e+02 1.797e+02 2.146e+02 2.516e+02 4.177e+02, threshold=4.292e+02, percent-clipped=1.0 +2023-03-26 04:55:22,488 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26824.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:55:57,245 INFO [finetune.py:976] (4/7) Epoch 5, batch 3950, loss[loss=0.1918, simple_loss=0.2522, pruned_loss=0.06568, over 4766.00 frames. ], tot_loss[loss=0.218, simple_loss=0.2757, pruned_loss=0.08016, over 955867.49 frames. ], batch size: 54, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:56:06,712 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26872.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:56:29,756 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26904.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 04:56:30,927 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26906.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:56:34,935 INFO [finetune.py:976] (4/7) Epoch 5, batch 4000, loss[loss=0.2606, simple_loss=0.3245, pruned_loss=0.09835, over 4865.00 frames. ], tot_loss[loss=0.2181, simple_loss=0.2752, pruned_loss=0.08052, over 953344.61 frames. ], batch size: 44, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:56:43,169 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.227e+02 1.662e+02 2.009e+02 2.453e+02 3.802e+02, threshold=4.018e+02, percent-clipped=0.0 +2023-03-26 04:56:43,287 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5044, 1.6164, 2.0055, 1.8696, 1.8138, 3.7157, 1.3636, 1.8729], + device='cuda:4'), covar=tensor([0.0909, 0.1566, 0.1422, 0.0972, 0.1403, 0.0213, 0.1406, 0.1486], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0082, 0.0077, 0.0080, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 04:57:05,888 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26950.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:57:08,738 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=26954.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:57:18,754 INFO [finetune.py:976] (4/7) Epoch 5, batch 4050, loss[loss=0.208, simple_loss=0.2681, pruned_loss=0.07396, over 4854.00 frames. ], tot_loss[loss=0.2215, simple_loss=0.2792, pruned_loss=0.08196, over 953455.70 frames. ], batch size: 31, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:57:26,845 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26965.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 04:57:37,855 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 04:58:10,821 INFO [finetune.py:976] (4/7) Epoch 5, batch 4100, loss[loss=0.2051, simple_loss=0.2684, pruned_loss=0.07086, over 4921.00 frames. ], tot_loss[loss=0.2239, simple_loss=0.2821, pruned_loss=0.08284, over 952960.33 frames. ], batch size: 38, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:58:11,432 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=27011.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:58:19,553 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.223e+02 1.740e+02 2.111e+02 2.577e+02 5.326e+02, threshold=4.223e+02, percent-clipped=3.0 +2023-03-26 04:58:26,577 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6467, 3.8772, 3.6806, 1.9820, 3.9055, 2.7143, 0.7942, 2.7305], + device='cuda:4'), covar=tensor([0.2275, 0.1428, 0.1420, 0.3006, 0.0840, 0.1032, 0.4448, 0.1301], + device='cuda:4'), in_proj_covar=tensor([0.0156, 0.0173, 0.0164, 0.0129, 0.0155, 0.0123, 0.0146, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 04:58:58,627 INFO [finetune.py:976] (4/7) Epoch 5, batch 4150, loss[loss=0.2297, simple_loss=0.2927, pruned_loss=0.08329, over 4907.00 frames. ], tot_loss[loss=0.2255, simple_loss=0.2839, pruned_loss=0.08358, over 953431.31 frames. ], batch size: 36, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:59:05,645 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=27070.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:59:32,488 INFO [finetune.py:976] (4/7) Epoch 5, batch 4200, loss[loss=0.2161, simple_loss=0.2841, pruned_loss=0.0741, over 4897.00 frames. ], tot_loss[loss=0.2242, simple_loss=0.2833, pruned_loss=0.08258, over 954720.66 frames. ], batch size: 37, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 04:59:37,724 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=27118.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 04:59:41,614 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.682e+02 1.955e+02 2.295e+02 5.538e+02, threshold=3.911e+02, percent-clipped=3.0 +2023-03-26 04:59:57,812 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=27148.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:00:10,426 INFO [finetune.py:976] (4/7) Epoch 5, batch 4250, loss[loss=0.1701, simple_loss=0.2323, pruned_loss=0.05397, over 4778.00 frames. ], tot_loss[loss=0.2228, simple_loss=0.2814, pruned_loss=0.08212, over 953690.01 frames. ], batch size: 28, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:00:10,537 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9006, 1.7808, 1.6294, 1.9163, 1.2363, 4.4681, 1.7443, 2.2044], + device='cuda:4'), covar=tensor([0.3277, 0.2406, 0.2141, 0.2341, 0.1970, 0.0114, 0.2494, 0.1346], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0114, 0.0118, 0.0122, 0.0118, 0.0098, 0.0101, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 05:01:08,503 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=27209.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:01:09,571 INFO [finetune.py:976] (4/7) Epoch 5, batch 4300, loss[loss=0.2044, simple_loss=0.2778, pruned_loss=0.06552, over 4829.00 frames. ], tot_loss[loss=0.2188, simple_loss=0.2773, pruned_loss=0.08013, over 955152.92 frames. ], batch size: 40, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:01:26,945 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.146e+02 1.762e+02 2.023e+02 2.453e+02 1.035e+03, threshold=4.046e+02, percent-clipped=2.0 +2023-03-26 05:01:59,786 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=27260.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 05:02:00,300 INFO [finetune.py:976] (4/7) Epoch 5, batch 4350, loss[loss=0.2442, simple_loss=0.2921, pruned_loss=0.09812, over 4908.00 frames. ], tot_loss[loss=0.2159, simple_loss=0.274, pruned_loss=0.07889, over 955800.30 frames. ], batch size: 35, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:02:30,557 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=27306.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:02:33,996 INFO [finetune.py:976] (4/7) Epoch 5, batch 4400, loss[loss=0.1717, simple_loss=0.2388, pruned_loss=0.05228, over 4745.00 frames. ], tot_loss[loss=0.2173, simple_loss=0.2748, pruned_loss=0.07988, over 956627.35 frames. ], batch size: 27, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:02:41,215 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.108e+02 1.601e+02 1.888e+02 2.389e+02 3.644e+02, threshold=3.775e+02, percent-clipped=0.0 +2023-03-26 05:02:51,660 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6501, 3.5530, 3.4062, 1.5641, 3.6160, 2.6674, 0.8414, 2.4982], + device='cuda:4'), covar=tensor([0.2542, 0.2019, 0.1580, 0.3594, 0.1138, 0.1033, 0.4803, 0.1570], + device='cuda:4'), in_proj_covar=tensor([0.0156, 0.0172, 0.0164, 0.0128, 0.0156, 0.0122, 0.0146, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 05:03:07,515 INFO [finetune.py:976] (4/7) Epoch 5, batch 4450, loss[loss=0.2022, simple_loss=0.274, pruned_loss=0.06523, over 4722.00 frames. ], tot_loss[loss=0.2216, simple_loss=0.2795, pruned_loss=0.08183, over 956266.80 frames. ], batch size: 59, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:03:11,366 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 05:03:40,733 INFO [finetune.py:976] (4/7) Epoch 5, batch 4500, loss[loss=0.2409, simple_loss=0.3015, pruned_loss=0.09016, over 4897.00 frames. ], tot_loss[loss=0.2212, simple_loss=0.28, pruned_loss=0.08124, over 957071.84 frames. ], batch size: 36, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:03:48,443 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.291e+02 1.723e+02 2.077e+02 2.543e+02 6.449e+02, threshold=4.154e+02, percent-clipped=4.0 +2023-03-26 05:04:14,226 INFO [finetune.py:976] (4/7) Epoch 5, batch 4550, loss[loss=0.2455, simple_loss=0.2922, pruned_loss=0.09943, over 4737.00 frames. ], tot_loss[loss=0.2241, simple_loss=0.2828, pruned_loss=0.08266, over 956797.50 frames. ], batch size: 23, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:04:39,362 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 05:04:42,683 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=27504.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:04:47,379 INFO [finetune.py:976] (4/7) Epoch 5, batch 4600, loss[loss=0.219, simple_loss=0.2848, pruned_loss=0.07661, over 4913.00 frames. ], tot_loss[loss=0.2232, simple_loss=0.2821, pruned_loss=0.08219, over 957344.91 frames. ], batch size: 37, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:04:55,107 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.119e+02 1.596e+02 2.009e+02 2.524e+02 8.514e+02, threshold=4.018e+02, percent-clipped=5.0 +2023-03-26 05:05:05,407 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.86 vs. limit=5.0 +2023-03-26 05:05:09,394 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.3406, 3.6679, 3.8893, 4.1443, 4.1031, 3.8587, 4.4221, 1.3930], + device='cuda:4'), covar=tensor([0.0722, 0.0939, 0.0853, 0.0933, 0.1173, 0.1500, 0.0658, 0.5012], + device='cuda:4'), in_proj_covar=tensor([0.0359, 0.0246, 0.0278, 0.0295, 0.0340, 0.0288, 0.0308, 0.0301], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:05:11,110 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4136, 1.2825, 1.1928, 1.3521, 1.6276, 1.5032, 1.3818, 1.2056], + device='cuda:4'), covar=tensor([0.0300, 0.0283, 0.0639, 0.0306, 0.0236, 0.0434, 0.0301, 0.0369], + device='cuda:4'), in_proj_covar=tensor([0.0086, 0.0111, 0.0137, 0.0116, 0.0104, 0.0100, 0.0090, 0.0108], + device='cuda:4'), out_proj_covar=tensor([6.6996e-05, 8.7301e-05, 1.0992e-04, 9.2182e-05, 8.1912e-05, 7.4333e-05, + 6.8954e-05, 8.4541e-05], device='cuda:4') +2023-03-26 05:05:20,049 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=27560.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 05:05:20,543 INFO [finetune.py:976] (4/7) Epoch 5, batch 4650, loss[loss=0.205, simple_loss=0.2642, pruned_loss=0.07295, over 4864.00 frames. ], tot_loss[loss=0.2201, simple_loss=0.2782, pruned_loss=0.08099, over 956752.94 frames. ], batch size: 31, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:05:39,074 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0161, 1.8763, 1.7517, 2.0310, 1.2685, 4.3047, 1.6577, 2.1905], + device='cuda:4'), covar=tensor([0.3221, 0.2202, 0.1913, 0.2054, 0.1737, 0.0107, 0.2471, 0.1315], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0114, 0.0117, 0.0122, 0.0117, 0.0098, 0.0101, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 05:06:07,101 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=27606.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:06:13,638 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=27608.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 05:06:15,420 INFO [finetune.py:976] (4/7) Epoch 5, batch 4700, loss[loss=0.1907, simple_loss=0.2276, pruned_loss=0.07688, over 4163.00 frames. ], tot_loss[loss=0.2161, simple_loss=0.274, pruned_loss=0.07908, over 956749.60 frames. ], batch size: 18, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:06:27,911 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.549e+02 1.903e+02 2.293e+02 3.137e+02, threshold=3.806e+02, percent-clipped=0.0 +2023-03-26 05:07:05,707 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=27654.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:07:14,262 INFO [finetune.py:976] (4/7) Epoch 5, batch 4750, loss[loss=0.2529, simple_loss=0.3037, pruned_loss=0.101, over 4854.00 frames. ], tot_loss[loss=0.2157, simple_loss=0.2731, pruned_loss=0.07911, over 956129.87 frames. ], batch size: 49, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:07:56,928 INFO [finetune.py:976] (4/7) Epoch 5, batch 4800, loss[loss=0.3153, simple_loss=0.3604, pruned_loss=0.1351, over 4866.00 frames. ], tot_loss[loss=0.2181, simple_loss=0.2759, pruned_loss=0.08009, over 956786.56 frames. ], batch size: 44, lr: 3.93e-03, grad_scale: 16.0 +2023-03-26 05:07:59,317 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=27714.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 05:08:04,701 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.165e+02 1.708e+02 2.066e+02 2.363e+02 4.852e+02, threshold=4.133e+02, percent-clipped=3.0 +2023-03-26 05:08:30,295 INFO [finetune.py:976] (4/7) Epoch 5, batch 4850, loss[loss=0.2355, simple_loss=0.3007, pruned_loss=0.08517, over 4869.00 frames. ], tot_loss[loss=0.2188, simple_loss=0.2778, pruned_loss=0.07991, over 957223.80 frames. ], batch size: 34, lr: 3.92e-03, grad_scale: 16.0 +2023-03-26 05:08:39,476 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=27775.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 05:08:59,122 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=27804.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:09:04,311 INFO [finetune.py:976] (4/7) Epoch 5, batch 4900, loss[loss=0.1772, simple_loss=0.2453, pruned_loss=0.0545, over 4672.00 frames. ], tot_loss[loss=0.2202, simple_loss=0.2796, pruned_loss=0.08037, over 957845.57 frames. ], batch size: 23, lr: 3.92e-03, grad_scale: 16.0 +2023-03-26 05:09:10,302 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.4118, 3.7533, 3.9627, 4.2485, 4.1120, 3.8801, 4.4895, 1.2710], + device='cuda:4'), covar=tensor([0.0684, 0.0821, 0.0743, 0.0810, 0.1150, 0.1528, 0.0605, 0.5268], + device='cuda:4'), in_proj_covar=tensor([0.0358, 0.0245, 0.0278, 0.0294, 0.0338, 0.0286, 0.0306, 0.0300], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:09:12,048 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.126e+02 1.628e+02 1.864e+02 2.335e+02 3.818e+02, threshold=3.728e+02, percent-clipped=0.0 +2023-03-26 05:09:20,494 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5652, 1.4485, 1.3463, 1.5509, 1.1084, 3.5393, 1.4184, 1.9214], + device='cuda:4'), covar=tensor([0.4393, 0.3019, 0.2654, 0.3121, 0.2131, 0.0211, 0.2714, 0.1414], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0114, 0.0118, 0.0122, 0.0117, 0.0098, 0.0101, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 05:09:30,673 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=27852.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:09:37,159 INFO [finetune.py:976] (4/7) Epoch 5, batch 4950, loss[loss=0.2634, simple_loss=0.3124, pruned_loss=0.1072, over 4147.00 frames. ], tot_loss[loss=0.2204, simple_loss=0.2802, pruned_loss=0.0803, over 956458.50 frames. ], batch size: 65, lr: 3.92e-03, grad_scale: 16.0 +2023-03-26 05:10:04,850 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7264, 1.6100, 1.5656, 1.8451, 2.2517, 1.7818, 1.4393, 1.4724], + device='cuda:4'), covar=tensor([0.2296, 0.2261, 0.1937, 0.1701, 0.1879, 0.1310, 0.2696, 0.1950], + device='cuda:4'), in_proj_covar=tensor([0.0237, 0.0211, 0.0203, 0.0186, 0.0238, 0.0178, 0.0217, 0.0191], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:10:10,466 INFO [finetune.py:976] (4/7) Epoch 5, batch 5000, loss[loss=0.2187, simple_loss=0.2889, pruned_loss=0.07419, over 4840.00 frames. ], tot_loss[loss=0.2183, simple_loss=0.2778, pruned_loss=0.07936, over 953544.16 frames. ], batch size: 47, lr: 3.92e-03, grad_scale: 16.0 +2023-03-26 05:10:19,083 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.230e+02 1.618e+02 1.837e+02 2.301e+02 4.829e+02, threshold=3.674e+02, percent-clipped=1.0 +2023-03-26 05:10:43,560 INFO [finetune.py:976] (4/7) Epoch 5, batch 5050, loss[loss=0.2351, simple_loss=0.2936, pruned_loss=0.08828, over 4843.00 frames. ], tot_loss[loss=0.2163, simple_loss=0.2754, pruned_loss=0.07858, over 954465.22 frames. ], batch size: 47, lr: 3.92e-03, grad_scale: 16.0 +2023-03-26 05:11:48,643 INFO [finetune.py:976] (4/7) Epoch 5, batch 5100, loss[loss=0.1619, simple_loss=0.2221, pruned_loss=0.05087, over 4772.00 frames. ], tot_loss[loss=0.2158, simple_loss=0.2741, pruned_loss=0.07873, over 954781.93 frames. ], batch size: 26, lr: 3.92e-03, grad_scale: 16.0 +2023-03-26 05:11:50,612 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6166, 1.5165, 1.5365, 1.5659, 1.0037, 2.8328, 1.0772, 1.6086], + device='cuda:4'), covar=tensor([0.3190, 0.2251, 0.1973, 0.2266, 0.1919, 0.0255, 0.2619, 0.1285], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0114, 0.0118, 0.0121, 0.0117, 0.0098, 0.0101, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 05:12:00,951 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 05:12:02,950 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.177e+02 1.639e+02 1.875e+02 2.408e+02 3.954e+02, threshold=3.749e+02, percent-clipped=2.0 +2023-03-26 05:12:15,622 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1326, 2.0260, 1.5844, 2.3617, 2.1406, 1.7347, 2.8343, 2.1268], + device='cuda:4'), covar=tensor([0.1815, 0.3735, 0.4347, 0.3769, 0.3331, 0.2189, 0.3756, 0.2637], + device='cuda:4'), in_proj_covar=tensor([0.0167, 0.0193, 0.0236, 0.0254, 0.0230, 0.0189, 0.0210, 0.0189], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:12:32,818 INFO [finetune.py:976] (4/7) Epoch 5, batch 5150, loss[loss=0.2586, simple_loss=0.3074, pruned_loss=0.1048, over 4210.00 frames. ], tot_loss[loss=0.2161, simple_loss=0.2741, pruned_loss=0.07902, over 954741.50 frames. ], batch size: 65, lr: 3.92e-03, grad_scale: 16.0 +2023-03-26 05:12:38,915 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28070.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 05:12:54,671 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 05:13:06,310 INFO [finetune.py:976] (4/7) Epoch 5, batch 5200, loss[loss=0.2356, simple_loss=0.2984, pruned_loss=0.08644, over 4806.00 frames. ], tot_loss[loss=0.2191, simple_loss=0.2777, pruned_loss=0.0802, over 952896.95 frames. ], batch size: 45, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:13:14,537 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.308e+02 1.748e+02 1.996e+02 2.342e+02 5.311e+02, threshold=3.992e+02, percent-clipped=1.0 +2023-03-26 05:13:39,349 INFO [finetune.py:976] (4/7) Epoch 5, batch 5250, loss[loss=0.2368, simple_loss=0.3011, pruned_loss=0.08621, over 4812.00 frames. ], tot_loss[loss=0.2226, simple_loss=0.2815, pruned_loss=0.08187, over 953448.23 frames. ], batch size: 30, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:14:01,708 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0268, 1.5760, 1.8508, 1.7977, 1.6031, 1.6833, 1.7915, 1.7333], + device='cuda:4'), covar=tensor([0.5546, 0.7607, 0.5903, 0.7374, 0.8177, 0.6217, 0.9031, 0.5895], + device='cuda:4'), in_proj_covar=tensor([0.0229, 0.0248, 0.0254, 0.0257, 0.0241, 0.0219, 0.0273, 0.0223], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:14:12,315 INFO [finetune.py:976] (4/7) Epoch 5, batch 5300, loss[loss=0.1884, simple_loss=0.2629, pruned_loss=0.05693, over 4818.00 frames. ], tot_loss[loss=0.2239, simple_loss=0.2828, pruned_loss=0.0825, over 954251.04 frames. ], batch size: 39, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:14:12,853 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 05:14:19,563 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.181e+02 1.725e+02 1.957e+02 2.435e+02 6.444e+02, threshold=3.915e+02, percent-clipped=2.0 +2023-03-26 05:14:24,312 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28229.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:14:39,417 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3877, 1.3298, 1.4517, 0.7251, 1.3851, 1.6736, 1.6568, 1.3490], + device='cuda:4'), covar=tensor([0.0977, 0.0823, 0.0468, 0.0628, 0.0465, 0.0463, 0.0385, 0.0704], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0159, 0.0121, 0.0138, 0.0133, 0.0124, 0.0148, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.7988e-05, 1.1759e-04, 8.7535e-05, 1.0033e-04, 9.5878e-05, 9.1481e-05, + 1.0980e-04, 1.0830e-04], device='cuda:4') +2023-03-26 05:14:51,796 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28255.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 05:14:56,276 INFO [finetune.py:976] (4/7) Epoch 5, batch 5350, loss[loss=0.1987, simple_loss=0.2616, pruned_loss=0.06787, over 4878.00 frames. ], tot_loss[loss=0.2234, simple_loss=0.2827, pruned_loss=0.08208, over 954404.69 frames. ], batch size: 31, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:15:13,528 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28287.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:15:15,778 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28290.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:15:28,418 INFO [finetune.py:976] (4/7) Epoch 5, batch 5400, loss[loss=0.1818, simple_loss=0.2502, pruned_loss=0.05673, over 4926.00 frames. ], tot_loss[loss=0.2191, simple_loss=0.2786, pruned_loss=0.07985, over 954453.42 frames. ], batch size: 33, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:15:31,987 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28316.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 05:15:36,015 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.900e+01 1.568e+02 1.878e+02 2.260e+02 3.573e+02, threshold=3.756e+02, percent-clipped=0.0 +2023-03-26 05:15:51,318 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9967, 1.7404, 2.1847, 3.6228, 2.5992, 2.6830, 0.8381, 2.8358], + device='cuda:4'), covar=tensor([0.1623, 0.1393, 0.1356, 0.0483, 0.0734, 0.1811, 0.1943, 0.0595], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0119, 0.0137, 0.0167, 0.0103, 0.0143, 0.0129, 0.0104], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 05:15:53,750 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28348.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:16:01,447 INFO [finetune.py:976] (4/7) Epoch 5, batch 5450, loss[loss=0.2092, simple_loss=0.2663, pruned_loss=0.07602, over 4861.00 frames. ], tot_loss[loss=0.2179, simple_loss=0.2758, pruned_loss=0.08, over 955063.84 frames. ], batch size: 31, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:16:18,310 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28370.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 05:16:21,444 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.75 vs. limit=5.0 +2023-03-26 05:16:46,829 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.2254, 1.2999, 1.3994, 0.6534, 1.1840, 1.5626, 1.6007, 1.2859], + device='cuda:4'), covar=tensor([0.0880, 0.0488, 0.0448, 0.0509, 0.0437, 0.0393, 0.0254, 0.0621], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0159, 0.0121, 0.0138, 0.0134, 0.0124, 0.0148, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.8104e-05, 1.1789e-04, 8.7888e-05, 1.0040e-04, 9.6370e-05, 9.1654e-05, + 1.0953e-04, 1.0856e-04], device='cuda:4') +2023-03-26 05:17:02,777 INFO [finetune.py:976] (4/7) Epoch 5, batch 5500, loss[loss=0.2178, simple_loss=0.2848, pruned_loss=0.07538, over 4804.00 frames. ], tot_loss[loss=0.2144, simple_loss=0.2721, pruned_loss=0.07833, over 955972.98 frames. ], batch size: 51, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:17:12,778 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=28418.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 05:17:21,151 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.059e+02 1.651e+02 2.036e+02 2.478e+02 5.642e+02, threshold=4.072e+02, percent-clipped=3.0 +2023-03-26 05:17:56,539 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2850, 3.6849, 3.8626, 4.1415, 4.0197, 3.7957, 4.3220, 1.3008], + device='cuda:4'), covar=tensor([0.0684, 0.0793, 0.0754, 0.0802, 0.1185, 0.1453, 0.0647, 0.5365], + device='cuda:4'), in_proj_covar=tensor([0.0355, 0.0244, 0.0275, 0.0293, 0.0337, 0.0285, 0.0304, 0.0298], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:18:07,650 INFO [finetune.py:976] (4/7) Epoch 5, batch 5550, loss[loss=0.3022, simple_loss=0.3484, pruned_loss=0.1281, over 4740.00 frames. ], tot_loss[loss=0.2172, simple_loss=0.275, pruned_loss=0.07974, over 956530.67 frames. ], batch size: 54, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:18:36,690 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28486.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:18:45,456 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28492.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:19:02,060 INFO [finetune.py:976] (4/7) Epoch 5, batch 5600, loss[loss=0.2741, simple_loss=0.3202, pruned_loss=0.114, over 4794.00 frames. ], tot_loss[loss=0.2202, simple_loss=0.2786, pruned_loss=0.08093, over 954863.35 frames. ], batch size: 25, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:19:14,572 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.753e+02 2.098e+02 2.591e+02 4.684e+02, threshold=4.196e+02, percent-clipped=2.0 +2023-03-26 05:19:40,189 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28547.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:19:47,580 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28553.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:19:52,218 INFO [finetune.py:976] (4/7) Epoch 5, batch 5650, loss[loss=0.271, simple_loss=0.3259, pruned_loss=0.108, over 4729.00 frames. ], tot_loss[loss=0.2224, simple_loss=0.2813, pruned_loss=0.08168, over 954137.84 frames. ], batch size: 59, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:19:54,622 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7765, 1.6132, 1.6570, 1.6792, 1.3769, 4.4195, 1.6142, 2.2170], + device='cuda:4'), covar=tensor([0.3518, 0.2499, 0.2017, 0.2221, 0.1681, 0.0101, 0.2425, 0.1269], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0114, 0.0117, 0.0121, 0.0117, 0.0097, 0.0100, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 05:19:58,759 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2442, 1.4847, 0.7402, 2.1639, 2.5659, 1.8319, 1.8485, 2.2043], + device='cuda:4'), covar=tensor([0.1459, 0.2083, 0.2211, 0.1091, 0.1783, 0.1860, 0.1381, 0.1900], + device='cuda:4'), in_proj_covar=tensor([0.0092, 0.0098, 0.0116, 0.0094, 0.0124, 0.0097, 0.0100, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 05:20:06,328 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28585.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:20:21,893 INFO [finetune.py:976] (4/7) Epoch 5, batch 5700, loss[loss=0.1639, simple_loss=0.2229, pruned_loss=0.05246, over 4250.00 frames. ], tot_loss[loss=0.2186, simple_loss=0.2769, pruned_loss=0.08013, over 940268.06 frames. ], batch size: 18, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:20:21,937 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28611.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 05:20:23,324 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 05:20:29,015 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.580e+02 1.894e+02 2.210e+02 5.665e+02, threshold=3.789e+02, percent-clipped=1.0 +2023-03-26 05:20:53,189 INFO [finetune.py:976] (4/7) Epoch 6, batch 0, loss[loss=0.1831, simple_loss=0.2479, pruned_loss=0.05916, over 4748.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2479, pruned_loss=0.05916, over 4748.00 frames. ], batch size: 27, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:20:53,190 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 05:20:56,586 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5191, 1.6129, 1.5723, 1.7049, 1.6999, 2.7905, 1.5141, 1.6770], + device='cuda:4'), covar=tensor([0.0892, 0.1515, 0.1055, 0.0860, 0.1313, 0.0334, 0.1218, 0.1454], + device='cuda:4'), in_proj_covar=tensor([0.0079, 0.0083, 0.0078, 0.0081, 0.0094, 0.0084, 0.0087, 0.0081], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 05:21:08,971 INFO [finetune.py:1010] (4/7) Epoch 6, validation: loss=0.1659, simple_loss=0.2379, pruned_loss=0.04693, over 2265189.00 frames. +2023-03-26 05:21:08,972 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6301MB +2023-03-26 05:21:15,235 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28643.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:21:15,281 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7694, 1.7311, 1.8062, 1.1440, 1.9183, 1.8333, 1.7860, 1.4996], + device='cuda:4'), covar=tensor([0.0686, 0.0699, 0.0739, 0.1038, 0.0610, 0.0754, 0.0724, 0.1289], + device='cuda:4'), in_proj_covar=tensor([0.0141, 0.0136, 0.0146, 0.0131, 0.0114, 0.0145, 0.0149, 0.0166], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:21:18,006 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0281, 2.1122, 1.9567, 1.3605, 2.2372, 2.2296, 2.1244, 1.7078], + device='cuda:4'), covar=tensor([0.0740, 0.0669, 0.0884, 0.1086, 0.0548, 0.0722, 0.0732, 0.1248], + device='cuda:4'), in_proj_covar=tensor([0.0141, 0.0136, 0.0146, 0.0131, 0.0114, 0.0145, 0.0148, 0.0166], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:21:23,705 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3761, 1.4472, 1.5467, 0.6331, 1.4346, 1.7234, 1.8080, 1.4103], + device='cuda:4'), covar=tensor([0.1075, 0.0687, 0.0497, 0.0746, 0.0440, 0.0522, 0.0351, 0.0745], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0159, 0.0121, 0.0138, 0.0134, 0.0124, 0.0148, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.7884e-05, 1.1752e-04, 8.7878e-05, 1.0043e-04, 9.6153e-05, 9.1878e-05, + 1.0956e-04, 1.0867e-04], device='cuda:4') +2023-03-26 05:21:59,715 INFO [finetune.py:976] (4/7) Epoch 6, batch 50, loss[loss=0.2434, simple_loss=0.3025, pruned_loss=0.09214, over 4905.00 frames. ], tot_loss[loss=0.2258, simple_loss=0.2854, pruned_loss=0.08308, over 217362.07 frames. ], batch size: 46, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:22:23,304 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0996, 0.9255, 0.9235, 1.1870, 1.2051, 1.1518, 1.0176, 0.9686], + device='cuda:4'), covar=tensor([0.0297, 0.0287, 0.0619, 0.0245, 0.0273, 0.0403, 0.0306, 0.0377], + device='cuda:4'), in_proj_covar=tensor([0.0087, 0.0111, 0.0138, 0.0118, 0.0104, 0.0101, 0.0091, 0.0109], + device='cuda:4'), out_proj_covar=tensor([6.8117e-05, 8.7615e-05, 1.1100e-04, 9.3341e-05, 8.2424e-05, 7.5017e-05, + 6.9768e-05, 8.5209e-05], device='cuda:4') +2023-03-26 05:22:30,502 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.152e+02 1.667e+02 1.978e+02 2.446e+02 6.098e+02, threshold=3.955e+02, percent-clipped=3.0 +2023-03-26 05:22:41,771 INFO [finetune.py:976] (4/7) Epoch 6, batch 100, loss[loss=0.2378, simple_loss=0.2962, pruned_loss=0.08969, over 4848.00 frames. ], tot_loss[loss=0.2155, simple_loss=0.2753, pruned_loss=0.07786, over 382523.45 frames. ], batch size: 44, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:22:52,799 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7091, 1.5794, 1.5342, 1.7150, 1.1202, 3.5806, 1.4415, 2.0363], + device='cuda:4'), covar=tensor([0.3234, 0.2396, 0.2103, 0.2140, 0.1874, 0.0145, 0.2659, 0.1231], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0113, 0.0117, 0.0121, 0.0117, 0.0097, 0.0101, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0003, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 05:23:14,512 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.96 vs. limit=5.0 +2023-03-26 05:23:15,381 INFO [finetune.py:976] (4/7) Epoch 6, batch 150, loss[loss=0.2071, simple_loss=0.2659, pruned_loss=0.07415, over 4905.00 frames. ], tot_loss[loss=0.2138, simple_loss=0.2718, pruned_loss=0.07794, over 509449.48 frames. ], batch size: 37, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:23:18,997 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3021, 2.2590, 1.8767, 1.7518, 2.5414, 2.4957, 2.3983, 2.1224], + device='cuda:4'), covar=tensor([0.0274, 0.0363, 0.0487, 0.0364, 0.0217, 0.0578, 0.0288, 0.0351], + device='cuda:4'), in_proj_covar=tensor([0.0087, 0.0111, 0.0137, 0.0117, 0.0104, 0.0100, 0.0091, 0.0109], + device='cuda:4'), out_proj_covar=tensor([6.7911e-05, 8.7281e-05, 1.1035e-04, 9.2864e-05, 8.1920e-05, 7.4570e-05, + 6.9356e-05, 8.4710e-05], device='cuda:4') +2023-03-26 05:23:37,612 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.084e+02 1.593e+02 1.877e+02 2.260e+02 4.734e+02, threshold=3.755e+02, percent-clipped=1.0 +2023-03-26 05:23:48,134 INFO [finetune.py:976] (4/7) Epoch 6, batch 200, loss[loss=0.2039, simple_loss=0.267, pruned_loss=0.07037, over 4827.00 frames. ], tot_loss[loss=0.2165, simple_loss=0.2731, pruned_loss=0.0799, over 608010.82 frames. ], batch size: 38, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:23:50,546 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28842.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:23:55,132 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28848.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:24:03,991 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28861.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:24:19,121 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28885.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:24:26,390 INFO [finetune.py:976] (4/7) Epoch 6, batch 250, loss[loss=0.2412, simple_loss=0.3068, pruned_loss=0.08778, over 4808.00 frames. ], tot_loss[loss=0.2167, simple_loss=0.2742, pruned_loss=0.07964, over 684594.01 frames. ], batch size: 39, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:24:37,339 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8439, 1.6734, 2.2777, 1.5725, 1.9969, 2.0796, 1.5570, 2.2791], + device='cuda:4'), covar=tensor([0.1523, 0.2041, 0.1546, 0.2366, 0.1004, 0.1618, 0.2725, 0.0964], + device='cuda:4'), in_proj_covar=tensor([0.0207, 0.0205, 0.0201, 0.0197, 0.0187, 0.0224, 0.0218, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:24:43,394 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 05:24:47,334 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28911.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 05:25:02,683 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28922.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:25:03,153 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.192e+02 1.635e+02 1.950e+02 2.422e+02 4.878e+02, threshold=3.900e+02, percent-clipped=5.0 +2023-03-26 05:25:09,280 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=28933.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:25:13,899 INFO [finetune.py:976] (4/7) Epoch 6, batch 300, loss[loss=0.2153, simple_loss=0.2818, pruned_loss=0.07442, over 4825.00 frames. ], tot_loss[loss=0.2211, simple_loss=0.2793, pruned_loss=0.08143, over 746392.80 frames. ], batch size: 39, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:25:16,435 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28943.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:25:28,663 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=28959.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 05:25:47,451 INFO [finetune.py:976] (4/7) Epoch 6, batch 350, loss[loss=0.2499, simple_loss=0.3165, pruned_loss=0.09166, over 4814.00 frames. ], tot_loss[loss=0.2216, simple_loss=0.2807, pruned_loss=0.08131, over 793804.81 frames. ], batch size: 39, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:25:49,414 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=28991.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:26:03,173 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29002.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:26:23,528 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2399, 2.1218, 2.7521, 1.6736, 2.3563, 2.5283, 1.8653, 2.6329], + device='cuda:4'), covar=tensor([0.1763, 0.1979, 0.1766, 0.2712, 0.1172, 0.1953, 0.2751, 0.1058], + device='cuda:4'), in_proj_covar=tensor([0.0206, 0.0205, 0.0200, 0.0196, 0.0186, 0.0222, 0.0217, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:26:28,283 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.301e+02 1.836e+02 2.201e+02 2.620e+02 4.241e+02, threshold=4.402e+02, percent-clipped=2.0 +2023-03-26 05:26:38,005 INFO [finetune.py:976] (4/7) Epoch 6, batch 400, loss[loss=0.1833, simple_loss=0.2494, pruned_loss=0.05858, over 4789.00 frames. ], tot_loss[loss=0.222, simple_loss=0.2815, pruned_loss=0.08121, over 830397.55 frames. ], batch size: 29, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:27:01,615 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29063.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:27:17,310 INFO [finetune.py:976] (4/7) Epoch 6, batch 450, loss[loss=0.2366, simple_loss=0.293, pruned_loss=0.09012, over 4821.00 frames. ], tot_loss[loss=0.2196, simple_loss=0.2795, pruned_loss=0.07986, over 858368.41 frames. ], batch size: 33, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:27:45,347 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.708e+02 1.996e+02 2.284e+02 5.200e+02, threshold=3.993e+02, percent-clipped=1.0 +2023-03-26 05:27:49,747 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 05:27:55,105 INFO [finetune.py:976] (4/7) Epoch 6, batch 500, loss[loss=0.2324, simple_loss=0.2943, pruned_loss=0.08531, over 4829.00 frames. ], tot_loss[loss=0.2165, simple_loss=0.2758, pruned_loss=0.07855, over 879146.13 frames. ], batch size: 39, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:27:57,523 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29142.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:28:01,679 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29148.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:28:20,072 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0985, 1.9160, 1.5564, 2.1189, 2.1521, 1.7276, 2.4694, 2.0882], + device='cuda:4'), covar=tensor([0.1567, 0.3346, 0.3856, 0.3491, 0.2715, 0.1860, 0.3651, 0.2204], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0192, 0.0235, 0.0253, 0.0230, 0.0189, 0.0210, 0.0190], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:28:28,350 INFO [finetune.py:976] (4/7) Epoch 6, batch 550, loss[loss=0.1665, simple_loss=0.2293, pruned_loss=0.05179, over 4823.00 frames. ], tot_loss[loss=0.2137, simple_loss=0.2725, pruned_loss=0.07749, over 897648.75 frames. ], batch size: 25, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:28:28,996 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29190.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:28:29,061 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5871, 1.5526, 1.9409, 1.2711, 1.7148, 1.7730, 1.4630, 1.9885], + device='cuda:4'), covar=tensor([0.1385, 0.2211, 0.1226, 0.1907, 0.0928, 0.1523, 0.2779, 0.0799], + device='cuda:4'), in_proj_covar=tensor([0.0206, 0.0205, 0.0200, 0.0196, 0.0186, 0.0222, 0.0218, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:28:34,887 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29196.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:29:00,100 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29217.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:29:04,679 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.109e+02 1.638e+02 2.076e+02 2.525e+02 4.090e+02, threshold=4.153e+02, percent-clipped=1.0 +2023-03-26 05:29:18,496 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29236.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:29:20,237 INFO [finetune.py:976] (4/7) Epoch 6, batch 600, loss[loss=0.2267, simple_loss=0.2767, pruned_loss=0.08837, over 4707.00 frames. ], tot_loss[loss=0.2154, simple_loss=0.2741, pruned_loss=0.07838, over 909262.98 frames. ], batch size: 23, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:29:26,813 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 05:29:28,003 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8769, 1.1257, 1.7194, 1.6103, 1.5022, 1.4646, 1.5432, 1.5118], + device='cuda:4'), covar=tensor([0.4805, 0.6999, 0.5476, 0.6195, 0.7220, 0.5299, 0.7700, 0.5317], + device='cuda:4'), in_proj_covar=tensor([0.0229, 0.0247, 0.0254, 0.0256, 0.0241, 0.0218, 0.0273, 0.0223], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:29:58,012 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5404, 2.2427, 1.8228, 2.5808, 2.5260, 2.0216, 3.0040, 2.3585], + device='cuda:4'), covar=tensor([0.1575, 0.3567, 0.4273, 0.3679, 0.2804, 0.1970, 0.3708, 0.2378], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0193, 0.0236, 0.0254, 0.0231, 0.0190, 0.0211, 0.0191], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:30:09,634 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9524, 1.6206, 2.2912, 1.6018, 2.1129, 2.0899, 1.6141, 2.2977], + device='cuda:4'), covar=tensor([0.1271, 0.1875, 0.1382, 0.1959, 0.0798, 0.1431, 0.2375, 0.0751], + device='cuda:4'), in_proj_covar=tensor([0.0208, 0.0207, 0.0202, 0.0198, 0.0188, 0.0223, 0.0219, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:30:21,080 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5256, 1.0831, 0.7536, 1.3810, 1.8971, 0.6961, 1.2537, 1.4517], + device='cuda:4'), covar=tensor([0.1586, 0.2241, 0.1936, 0.1287, 0.2164, 0.2178, 0.1608, 0.2070], + device='cuda:4'), in_proj_covar=tensor([0.0092, 0.0099, 0.0116, 0.0093, 0.0125, 0.0097, 0.0101, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 05:30:24,098 INFO [finetune.py:976] (4/7) Epoch 6, batch 650, loss[loss=0.2104, simple_loss=0.2752, pruned_loss=0.07276, over 4834.00 frames. ], tot_loss[loss=0.2176, simple_loss=0.2772, pruned_loss=0.07898, over 920040.93 frames. ], batch size: 33, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:30:34,121 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29297.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:30:45,653 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8613, 4.2325, 3.9121, 2.2207, 4.2209, 3.1581, 1.1001, 3.0399], + device='cuda:4'), covar=tensor([0.2360, 0.2111, 0.1510, 0.3273, 0.1021, 0.0995, 0.4703, 0.1442], + device='cuda:4'), in_proj_covar=tensor([0.0155, 0.0172, 0.0163, 0.0128, 0.0156, 0.0122, 0.0145, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 05:30:52,832 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9568, 1.6603, 2.1760, 1.5456, 2.0688, 2.1733, 1.6302, 2.2642], + device='cuda:4'), covar=tensor([0.1512, 0.2299, 0.1456, 0.2139, 0.1013, 0.1503, 0.2695, 0.0988], + device='cuda:4'), in_proj_covar=tensor([0.0208, 0.0207, 0.0203, 0.0198, 0.0188, 0.0224, 0.0220, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:31:13,222 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.225e+02 1.735e+02 1.982e+02 2.438e+02 4.583e+02, threshold=3.965e+02, percent-clipped=2.0 +2023-03-26 05:31:33,681 INFO [finetune.py:976] (4/7) Epoch 6, batch 700, loss[loss=0.2183, simple_loss=0.2835, pruned_loss=0.07658, over 4913.00 frames. ], tot_loss[loss=0.2182, simple_loss=0.2784, pruned_loss=0.07898, over 929152.18 frames. ], batch size: 42, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:31:46,289 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29358.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:32:17,028 INFO [finetune.py:976] (4/7) Epoch 6, batch 750, loss[loss=0.2622, simple_loss=0.3066, pruned_loss=0.1089, over 4885.00 frames. ], tot_loss[loss=0.2189, simple_loss=0.2791, pruned_loss=0.07937, over 934754.31 frames. ], batch size: 32, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:32:56,872 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6588, 0.6441, 1.5926, 1.4004, 1.3207, 1.3073, 1.2942, 1.4113], + device='cuda:4'), covar=tensor([0.4511, 0.6571, 0.5452, 0.5898, 0.6688, 0.5313, 0.7127, 0.4984], + device='cuda:4'), in_proj_covar=tensor([0.0229, 0.0247, 0.0254, 0.0256, 0.0242, 0.0219, 0.0274, 0.0223], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:32:57,864 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 05:32:59,102 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.656e+01 1.814e+02 2.109e+02 2.501e+02 5.044e+02, threshold=4.217e+02, percent-clipped=3.0 +2023-03-26 05:33:26,484 INFO [finetune.py:976] (4/7) Epoch 6, batch 800, loss[loss=0.1943, simple_loss=0.256, pruned_loss=0.06633, over 4793.00 frames. ], tot_loss[loss=0.219, simple_loss=0.2794, pruned_loss=0.07933, over 941275.44 frames. ], batch size: 29, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:34:10,553 INFO [finetune.py:976] (4/7) Epoch 6, batch 850, loss[loss=0.1943, simple_loss=0.2661, pruned_loss=0.06131, over 4785.00 frames. ], tot_loss[loss=0.2154, simple_loss=0.2759, pruned_loss=0.07749, over 943441.86 frames. ], batch size: 28, lr: 3.92e-03, grad_scale: 32.0 +2023-03-26 05:34:19,724 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6923, 1.3997, 1.0380, 0.1977, 1.2126, 1.4679, 1.3557, 1.4297], + device='cuda:4'), covar=tensor([0.0926, 0.0934, 0.1363, 0.2261, 0.1503, 0.2445, 0.2332, 0.0902], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0200, 0.0202, 0.0189, 0.0217, 0.0209, 0.0220, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:34:32,901 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.69 vs. limit=5.0 +2023-03-26 05:34:43,538 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29517.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:34:52,273 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.241e+02 1.650e+02 2.021e+02 2.394e+02 5.702e+02, threshold=4.042e+02, percent-clipped=1.0 +2023-03-26 05:35:14,735 INFO [finetune.py:976] (4/7) Epoch 6, batch 900, loss[loss=0.1538, simple_loss=0.2213, pruned_loss=0.04311, over 4827.00 frames. ], tot_loss[loss=0.2132, simple_loss=0.2728, pruned_loss=0.07683, over 946629.10 frames. ], batch size: 30, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:35:32,187 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29551.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:35:46,203 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29565.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:36:14,403 INFO [finetune.py:976] (4/7) Epoch 6, batch 950, loss[loss=0.2061, simple_loss=0.2682, pruned_loss=0.07199, over 4905.00 frames. ], tot_loss[loss=0.212, simple_loss=0.2711, pruned_loss=0.07648, over 948432.99 frames. ], batch size: 35, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:36:21,377 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29592.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:36:22,673 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5134, 1.3776, 1.4325, 1.4281, 1.0509, 3.2010, 1.2803, 1.7915], + device='cuda:4'), covar=tensor([0.3482, 0.2426, 0.2072, 0.2428, 0.1975, 0.0206, 0.2804, 0.1346], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0115, 0.0119, 0.0123, 0.0118, 0.0099, 0.0102, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 05:36:44,005 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29612.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:36:56,224 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.131e+02 1.661e+02 1.977e+02 2.359e+02 4.237e+02, threshold=3.954e+02, percent-clipped=2.0 +2023-03-26 05:37:17,830 INFO [finetune.py:976] (4/7) Epoch 6, batch 1000, loss[loss=0.2571, simple_loss=0.3273, pruned_loss=0.09341, over 4847.00 frames. ], tot_loss[loss=0.2149, simple_loss=0.2742, pruned_loss=0.07777, over 949978.52 frames. ], batch size: 49, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:37:45,089 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29658.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:37:47,501 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29662.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:37:58,424 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-03-26 05:38:20,623 INFO [finetune.py:976] (4/7) Epoch 6, batch 1050, loss[loss=0.2111, simple_loss=0.2831, pruned_loss=0.06954, over 4803.00 frames. ], tot_loss[loss=0.2168, simple_loss=0.277, pruned_loss=0.07827, over 953076.24 frames. ], batch size: 45, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:38:41,610 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29706.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:38:49,786 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-03-26 05:38:59,044 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3572, 1.5007, 1.9718, 1.7664, 1.7195, 3.6483, 1.3340, 1.7677], + device='cuda:4'), covar=tensor([0.1075, 0.1624, 0.1299, 0.1054, 0.1422, 0.0235, 0.1415, 0.1538], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0082, 0.0078, 0.0080, 0.0093, 0.0084, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 05:39:02,907 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.795e+02 2.095e+02 2.533e+02 7.754e+02, threshold=4.191e+02, percent-clipped=4.0 +2023-03-26 05:39:03,051 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29723.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 05:39:23,374 INFO [finetune.py:976] (4/7) Epoch 6, batch 1100, loss[loss=0.2087, simple_loss=0.2843, pruned_loss=0.0665, over 4905.00 frames. ], tot_loss[loss=0.2182, simple_loss=0.2787, pruned_loss=0.07881, over 954775.32 frames. ], batch size: 36, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:39:29,808 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7370, 1.5499, 2.1907, 3.5094, 2.5395, 2.3599, 1.3155, 2.7237], + device='cuda:4'), covar=tensor([0.1749, 0.1489, 0.1345, 0.0614, 0.0736, 0.1621, 0.1689, 0.0602], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0117, 0.0135, 0.0164, 0.0102, 0.0141, 0.0127, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 05:39:41,426 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0647, 1.8506, 2.4075, 1.6006, 2.1776, 2.2754, 1.7991, 2.4946], + device='cuda:4'), covar=tensor([0.1503, 0.2108, 0.1360, 0.2179, 0.1022, 0.1574, 0.2562, 0.0898], + device='cuda:4'), in_proj_covar=tensor([0.0205, 0.0203, 0.0199, 0.0195, 0.0184, 0.0220, 0.0216, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:40:24,914 INFO [finetune.py:976] (4/7) Epoch 6, batch 1150, loss[loss=0.2424, simple_loss=0.2907, pruned_loss=0.09709, over 4840.00 frames. ], tot_loss[loss=0.2178, simple_loss=0.2786, pruned_loss=0.07852, over 953867.16 frames. ], batch size: 30, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:40:37,953 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29802.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:40:47,268 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 05:41:01,073 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.060e+02 1.742e+02 2.057e+02 2.377e+02 6.600e+02, threshold=4.115e+02, percent-clipped=1.0 +2023-03-26 05:41:11,693 INFO [finetune.py:976] (4/7) Epoch 6, batch 1200, loss[loss=0.1909, simple_loss=0.2538, pruned_loss=0.06396, over 4795.00 frames. ], tot_loss[loss=0.2147, simple_loss=0.2753, pruned_loss=0.07705, over 954748.41 frames. ], batch size: 45, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:41:11,803 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0873, 1.3410, 1.0540, 1.3969, 1.4616, 2.3723, 1.2385, 1.4867], + device='cuda:4'), covar=tensor([0.1078, 0.1810, 0.1271, 0.0989, 0.1578, 0.0398, 0.1484, 0.1725], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0082, 0.0077, 0.0080, 0.0093, 0.0083, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 05:41:28,784 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29863.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:41:44,760 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8243, 4.5269, 4.2775, 2.2494, 4.5756, 3.3787, 0.9854, 3.1029], + device='cuda:4'), covar=tensor([0.2606, 0.1740, 0.1333, 0.3043, 0.0661, 0.0936, 0.4371, 0.1402], + device='cuda:4'), in_proj_covar=tensor([0.0156, 0.0172, 0.0165, 0.0129, 0.0156, 0.0123, 0.0146, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 05:41:45,291 INFO [finetune.py:976] (4/7) Epoch 6, batch 1250, loss[loss=0.1903, simple_loss=0.2549, pruned_loss=0.06284, over 4821.00 frames. ], tot_loss[loss=0.2125, simple_loss=0.2723, pruned_loss=0.07637, over 954380.14 frames. ], batch size: 41, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:41:47,163 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29892.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:41:57,710 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29907.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:42:07,774 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.170e+02 1.638e+02 1.953e+02 2.201e+02 4.150e+02, threshold=3.906e+02, percent-clipped=1.0 +2023-03-26 05:42:07,911 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8188, 1.8773, 1.5475, 1.4436, 2.0630, 2.0973, 1.9039, 1.8276], + device='cuda:4'), covar=tensor([0.0344, 0.0372, 0.0562, 0.0461, 0.0311, 0.0546, 0.0355, 0.0420], + device='cuda:4'), in_proj_covar=tensor([0.0088, 0.0112, 0.0139, 0.0118, 0.0104, 0.0101, 0.0091, 0.0109], + device='cuda:4'), out_proj_covar=tensor([6.8802e-05, 8.8021e-05, 1.1113e-04, 9.3649e-05, 8.2152e-05, 7.4853e-05, + 6.9935e-05, 8.5309e-05], device='cuda:4') +2023-03-26 05:42:18,525 INFO [finetune.py:976] (4/7) Epoch 6, batch 1300, loss[loss=0.1961, simple_loss=0.258, pruned_loss=0.06715, over 4747.00 frames. ], tot_loss[loss=0.2106, simple_loss=0.2698, pruned_loss=0.0757, over 954671.00 frames. ], batch size: 27, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:42:19,141 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=29940.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:42:53,740 INFO [finetune.py:976] (4/7) Epoch 6, batch 1350, loss[loss=0.2103, simple_loss=0.2664, pruned_loss=0.07707, over 4708.00 frames. ], tot_loss[loss=0.2109, simple_loss=0.2699, pruned_loss=0.07591, over 955892.30 frames. ], batch size: 23, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:43:11,740 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3229, 2.5183, 2.1558, 1.9858, 2.6457, 2.9660, 2.6893, 2.3647], + device='cuda:4'), covar=tensor([0.0349, 0.0329, 0.0530, 0.0381, 0.0268, 0.0320, 0.0249, 0.0337], + device='cuda:4'), in_proj_covar=tensor([0.0088, 0.0112, 0.0138, 0.0118, 0.0104, 0.0100, 0.0091, 0.0109], + device='cuda:4'), out_proj_covar=tensor([6.8544e-05, 8.7849e-05, 1.1045e-04, 9.3178e-05, 8.1821e-05, 7.4603e-05, + 6.9649e-05, 8.5208e-05], device='cuda:4') +2023-03-26 05:43:22,435 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30018.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 05:43:25,358 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.660e+02 2.040e+02 2.486e+02 4.804e+02, threshold=4.081e+02, percent-clipped=2.0 +2023-03-26 05:43:35,491 INFO [finetune.py:976] (4/7) Epoch 6, batch 1400, loss[loss=0.2092, simple_loss=0.2808, pruned_loss=0.06874, over 4903.00 frames. ], tot_loss[loss=0.2135, simple_loss=0.2731, pruned_loss=0.07695, over 953181.02 frames. ], batch size: 37, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:43:57,177 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-26 05:43:58,005 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2674, 1.2977, 1.5571, 1.1212, 1.2753, 1.4477, 1.2671, 1.5704], + device='cuda:4'), covar=tensor([0.1296, 0.2147, 0.1413, 0.1597, 0.1047, 0.1306, 0.2969, 0.1001], + device='cuda:4'), in_proj_covar=tensor([0.0206, 0.0205, 0.0200, 0.0196, 0.0184, 0.0221, 0.0217, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:44:14,310 INFO [finetune.py:976] (4/7) Epoch 6, batch 1450, loss[loss=0.1775, simple_loss=0.2375, pruned_loss=0.05873, over 4790.00 frames. ], tot_loss[loss=0.2154, simple_loss=0.2754, pruned_loss=0.07768, over 954972.82 frames. ], batch size: 26, lr: 3.91e-03, grad_scale: 64.0 +2023-03-26 05:44:16,140 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6291, 2.2699, 2.8249, 1.7764, 2.5745, 2.8490, 2.2276, 2.9197], + device='cuda:4'), covar=tensor([0.1453, 0.2054, 0.1792, 0.2671, 0.1129, 0.1800, 0.2410, 0.0983], + device='cuda:4'), in_proj_covar=tensor([0.0207, 0.0205, 0.0200, 0.0196, 0.0184, 0.0221, 0.0217, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:44:58,653 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.184e+02 1.825e+02 2.216e+02 2.642e+02 7.386e+02, threshold=4.431e+02, percent-clipped=2.0 +2023-03-26 05:44:59,412 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30125.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:45:16,735 INFO [finetune.py:976] (4/7) Epoch 6, batch 1500, loss[loss=0.227, simple_loss=0.289, pruned_loss=0.08246, over 4910.00 frames. ], tot_loss[loss=0.2184, simple_loss=0.2789, pruned_loss=0.07893, over 956389.28 frames. ], batch size: 37, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:45:16,896 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8485, 1.7228, 1.4675, 1.5270, 1.6416, 1.5466, 1.6343, 2.3525], + device='cuda:4'), covar=tensor([0.6461, 0.6395, 0.5030, 0.6250, 0.5508, 0.3677, 0.6070, 0.2305], + device='cuda:4'), in_proj_covar=tensor([0.0282, 0.0257, 0.0219, 0.0284, 0.0239, 0.0201, 0.0245, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:45:36,531 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 05:45:38,727 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30158.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:45:59,650 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30186.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:46:01,369 INFO [finetune.py:976] (4/7) Epoch 6, batch 1550, loss[loss=0.2067, simple_loss=0.2614, pruned_loss=0.076, over 4863.00 frames. ], tot_loss[loss=0.2191, simple_loss=0.2797, pruned_loss=0.07928, over 954924.81 frames. ], batch size: 34, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:46:01,444 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9179, 4.1285, 3.9391, 2.1198, 4.1304, 3.2294, 0.9291, 2.9496], + device='cuda:4'), covar=tensor([0.3018, 0.1811, 0.1470, 0.3201, 0.0975, 0.0921, 0.4391, 0.1479], + device='cuda:4'), in_proj_covar=tensor([0.0157, 0.0173, 0.0165, 0.0130, 0.0158, 0.0124, 0.0146, 0.0125], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 05:46:12,979 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30207.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:46:14,234 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30209.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:46:25,109 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.142e+02 1.698e+02 2.038e+02 2.458e+02 3.858e+02, threshold=4.076e+02, percent-clipped=0.0 +2023-03-26 05:46:34,721 INFO [finetune.py:976] (4/7) Epoch 6, batch 1600, loss[loss=0.1755, simple_loss=0.2478, pruned_loss=0.05159, over 4782.00 frames. ], tot_loss[loss=0.2159, simple_loss=0.2762, pruned_loss=0.07783, over 955512.42 frames. ], batch size: 29, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:46:43,569 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.48 vs. limit=5.0 +2023-03-26 05:46:44,947 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=30255.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:46:56,013 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30270.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:47:08,006 INFO [finetune.py:976] (4/7) Epoch 6, batch 1650, loss[loss=0.2079, simple_loss=0.2688, pruned_loss=0.07349, over 4896.00 frames. ], tot_loss[loss=0.2149, simple_loss=0.2744, pruned_loss=0.07773, over 955394.17 frames. ], batch size: 46, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:47:08,147 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0440, 1.9213, 1.5959, 1.7771, 1.8229, 1.7838, 1.7501, 2.6416], + device='cuda:4'), covar=tensor([0.6393, 0.6912, 0.5089, 0.6666, 0.5806, 0.3693, 0.6129, 0.2346], + device='cuda:4'), in_proj_covar=tensor([0.0283, 0.0259, 0.0221, 0.0285, 0.0241, 0.0203, 0.0246, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:47:27,132 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30318.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 05:47:31,617 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.138e+02 1.710e+02 1.999e+02 2.408e+02 3.997e+02, threshold=3.998e+02, percent-clipped=0.0 +2023-03-26 05:47:41,358 INFO [finetune.py:976] (4/7) Epoch 6, batch 1700, loss[loss=0.1743, simple_loss=0.2474, pruned_loss=0.05056, over 4781.00 frames. ], tot_loss[loss=0.2126, simple_loss=0.2714, pruned_loss=0.07694, over 955405.72 frames. ], batch size: 29, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:47:55,184 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30360.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:47:58,775 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=30366.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:48:00,067 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.2178, 1.4252, 1.5115, 0.8122, 1.3205, 1.6254, 1.7027, 1.4657], + device='cuda:4'), covar=tensor([0.0934, 0.0649, 0.0468, 0.0517, 0.0533, 0.0633, 0.0332, 0.0599], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0159, 0.0122, 0.0138, 0.0134, 0.0124, 0.0148, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.8896e-05, 1.1782e-04, 8.8427e-05, 1.0118e-04, 9.6053e-05, 9.1774e-05, + 1.0956e-04, 1.0856e-04], device='cuda:4') +2023-03-26 05:48:27,034 INFO [finetune.py:976] (4/7) Epoch 6, batch 1750, loss[loss=0.2458, simple_loss=0.3102, pruned_loss=0.09069, over 4825.00 frames. ], tot_loss[loss=0.2148, simple_loss=0.2737, pruned_loss=0.07789, over 956808.03 frames. ], batch size: 39, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:48:48,758 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30421.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:48:50,961 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.220e+01 1.748e+02 2.265e+02 2.778e+02 4.150e+02, threshold=4.530e+02, percent-clipped=1.0 +2023-03-26 05:49:00,664 INFO [finetune.py:976] (4/7) Epoch 6, batch 1800, loss[loss=0.2151, simple_loss=0.2787, pruned_loss=0.07581, over 4859.00 frames. ], tot_loss[loss=0.2168, simple_loss=0.2765, pruned_loss=0.07852, over 956823.37 frames. ], batch size: 31, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:49:13,263 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30458.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:49:28,997 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30481.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:49:33,834 INFO [finetune.py:976] (4/7) Epoch 6, batch 1850, loss[loss=0.2294, simple_loss=0.277, pruned_loss=0.09088, over 4796.00 frames. ], tot_loss[loss=0.2183, simple_loss=0.2779, pruned_loss=0.07935, over 954553.81 frames. ], batch size: 26, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:49:39,911 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2889, 2.0647, 2.7080, 1.6611, 2.6604, 2.6903, 1.8835, 2.7809], + device='cuda:4'), covar=tensor([0.1861, 0.2197, 0.1581, 0.2658, 0.0980, 0.1720, 0.2714, 0.0912], + device='cuda:4'), in_proj_covar=tensor([0.0205, 0.0204, 0.0198, 0.0195, 0.0183, 0.0220, 0.0216, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:49:44,724 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=30506.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:50:03,765 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.671e+02 2.010e+02 2.577e+02 4.739e+02, threshold=4.020e+02, percent-clipped=1.0 +2023-03-26 05:50:22,792 INFO [finetune.py:976] (4/7) Epoch 6, batch 1900, loss[loss=0.1953, simple_loss=0.2757, pruned_loss=0.05744, over 4799.00 frames. ], tot_loss[loss=0.2176, simple_loss=0.2777, pruned_loss=0.07877, over 953226.73 frames. ], batch size: 41, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:50:52,243 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30565.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:51:20,334 INFO [finetune.py:976] (4/7) Epoch 6, batch 1950, loss[loss=0.187, simple_loss=0.2497, pruned_loss=0.06216, over 4870.00 frames. ], tot_loss[loss=0.2138, simple_loss=0.2741, pruned_loss=0.07674, over 952746.68 frames. ], batch size: 31, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:51:58,194 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.556e+02 1.883e+02 2.215e+02 4.222e+02, threshold=3.767e+02, percent-clipped=2.0 +2023-03-26 05:52:07,042 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6845, 3.3988, 3.2425, 1.4397, 3.4489, 2.5051, 0.7244, 2.2207], + device='cuda:4'), covar=tensor([0.2436, 0.1657, 0.1940, 0.3608, 0.1146, 0.1123, 0.4777, 0.1725], + device='cuda:4'), in_proj_covar=tensor([0.0155, 0.0171, 0.0163, 0.0128, 0.0156, 0.0122, 0.0145, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 05:52:07,152 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 05:52:09,310 INFO [finetune.py:976] (4/7) Epoch 6, batch 2000, loss[loss=0.1998, simple_loss=0.2524, pruned_loss=0.07363, over 4865.00 frames. ], tot_loss[loss=0.2134, simple_loss=0.2729, pruned_loss=0.07689, over 952642.16 frames. ], batch size: 31, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:52:16,074 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6839, 1.2797, 0.9022, 1.6355, 2.1750, 1.2563, 1.5702, 1.6741], + device='cuda:4'), covar=tensor([0.1440, 0.1992, 0.1964, 0.1113, 0.1783, 0.1859, 0.1323, 0.1806], + device='cuda:4'), in_proj_covar=tensor([0.0092, 0.0099, 0.0115, 0.0094, 0.0125, 0.0097, 0.0100, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 05:52:19,631 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30655.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:52:21,425 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0904, 1.7283, 2.5846, 4.0229, 2.9121, 2.7510, 0.6029, 3.2037], + device='cuda:4'), covar=tensor([0.1866, 0.1629, 0.1374, 0.0525, 0.0754, 0.1409, 0.2405, 0.0520], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0118, 0.0135, 0.0165, 0.0102, 0.0141, 0.0128, 0.0103], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 05:52:56,172 INFO [finetune.py:976] (4/7) Epoch 6, batch 2050, loss[loss=0.2009, simple_loss=0.2551, pruned_loss=0.07331, over 4913.00 frames. ], tot_loss[loss=0.2091, simple_loss=0.2688, pruned_loss=0.07473, over 955318.40 frames. ], batch size: 37, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:53:12,586 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7067, 1.5487, 1.4192, 1.6798, 2.1581, 1.7561, 1.2879, 1.3885], + device='cuda:4'), covar=tensor([0.2047, 0.2061, 0.1828, 0.1617, 0.1756, 0.1184, 0.2736, 0.1791], + device='cuda:4'), in_proj_covar=tensor([0.0233, 0.0208, 0.0202, 0.0186, 0.0236, 0.0175, 0.0213, 0.0189], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:53:14,343 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30716.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:53:14,390 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30716.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 05:53:20,052 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.195e+02 1.687e+02 1.889e+02 2.318e+02 4.112e+02, threshold=3.779e+02, percent-clipped=3.0 +2023-03-26 05:53:30,033 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2332, 3.7998, 3.9338, 3.9430, 3.7566, 3.5363, 4.3885, 1.4979], + device='cuda:4'), covar=tensor([0.1314, 0.1540, 0.1546, 0.2230, 0.2247, 0.2469, 0.1193, 0.7349], + device='cuda:4'), in_proj_covar=tensor([0.0354, 0.0242, 0.0276, 0.0292, 0.0331, 0.0282, 0.0302, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:53:40,714 INFO [finetune.py:976] (4/7) Epoch 6, batch 2100, loss[loss=0.2226, simple_loss=0.2957, pruned_loss=0.07474, over 4817.00 frames. ], tot_loss[loss=0.2107, simple_loss=0.2705, pruned_loss=0.07544, over 956379.44 frames. ], batch size: 41, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:54:13,638 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30781.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:54:18,839 INFO [finetune.py:976] (4/7) Epoch 6, batch 2150, loss[loss=0.2435, simple_loss=0.3057, pruned_loss=0.09069, over 4814.00 frames. ], tot_loss[loss=0.2124, simple_loss=0.2728, pruned_loss=0.07604, over 956800.49 frames. ], batch size: 51, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:54:42,038 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.093e+02 1.708e+02 2.011e+02 2.625e+02 4.679e+02, threshold=4.022e+02, percent-clipped=7.0 +2023-03-26 05:54:46,131 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=30829.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:54:52,228 INFO [finetune.py:976] (4/7) Epoch 6, batch 2200, loss[loss=0.2263, simple_loss=0.2939, pruned_loss=0.07935, over 4890.00 frames. ], tot_loss[loss=0.216, simple_loss=0.2766, pruned_loss=0.07769, over 957485.72 frames. ], batch size: 35, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:54:55,385 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.72 vs. limit=5.0 +2023-03-26 05:55:16,404 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30865.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:55:26,537 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6830, 1.2115, 0.9183, 1.5716, 1.9964, 1.1796, 1.4839, 1.6587], + device='cuda:4'), covar=tensor([0.1536, 0.2183, 0.2113, 0.1193, 0.2137, 0.2421, 0.1358, 0.1942], + device='cuda:4'), in_proj_covar=tensor([0.0092, 0.0099, 0.0116, 0.0094, 0.0125, 0.0097, 0.0100, 0.0093], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 05:55:37,472 INFO [finetune.py:976] (4/7) Epoch 6, batch 2250, loss[loss=0.1929, simple_loss=0.2473, pruned_loss=0.06927, over 4714.00 frames. ], tot_loss[loss=0.2166, simple_loss=0.277, pruned_loss=0.0781, over 956594.28 frames. ], batch size: 23, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:55:58,592 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30903.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:56:10,382 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=30913.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:56:22,249 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.065e+02 1.611e+02 1.958e+02 2.302e+02 5.232e+02, threshold=3.915e+02, percent-clipped=2.0 +2023-03-26 05:56:22,992 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30925.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:56:34,322 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30934.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:56:42,906 INFO [finetune.py:976] (4/7) Epoch 6, batch 2300, loss[loss=0.1752, simple_loss=0.233, pruned_loss=0.05871, over 4738.00 frames. ], tot_loss[loss=0.2158, simple_loss=0.2764, pruned_loss=0.07767, over 956790.24 frames. ], batch size: 23, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:57:16,237 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30964.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:57:34,576 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1198, 1.7565, 1.7615, 0.9550, 1.8740, 2.1239, 1.8649, 1.7641], + device='cuda:4'), covar=tensor([0.0867, 0.0652, 0.0474, 0.0650, 0.0617, 0.0529, 0.0493, 0.0507], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0157, 0.0119, 0.0137, 0.0132, 0.0123, 0.0146, 0.0144], + device='cuda:4'), out_proj_covar=tensor([9.7596e-05, 1.1587e-04, 8.6636e-05, 9.9623e-05, 9.4956e-05, 9.0943e-05, + 1.0826e-04, 1.0673e-04], device='cuda:4') +2023-03-26 05:57:46,794 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30986.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:57:49,015 INFO [finetune.py:976] (4/7) Epoch 6, batch 2350, loss[loss=0.2401, simple_loss=0.3016, pruned_loss=0.08927, over 4380.00 frames. ], tot_loss[loss=0.2134, simple_loss=0.2739, pruned_loss=0.07646, over 956677.84 frames. ], batch size: 66, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:57:57,489 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30995.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:58:06,677 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1904, 2.3310, 2.1082, 1.4916, 2.4400, 2.4112, 2.3943, 2.1041], + device='cuda:4'), covar=tensor([0.0634, 0.0565, 0.0750, 0.0962, 0.0429, 0.0681, 0.0585, 0.0828], + device='cuda:4'), in_proj_covar=tensor([0.0139, 0.0135, 0.0144, 0.0128, 0.0113, 0.0144, 0.0146, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:58:07,210 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7420, 3.8447, 3.6812, 1.9559, 3.8836, 2.9516, 0.8912, 2.7393], + device='cuda:4'), covar=tensor([0.2554, 0.1610, 0.1337, 0.2817, 0.0898, 0.0965, 0.4087, 0.1374], + device='cuda:4'), in_proj_covar=tensor([0.0155, 0.0170, 0.0163, 0.0128, 0.0156, 0.0123, 0.0145, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 05:58:19,821 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31011.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 05:58:19,871 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31011.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 05:58:28,249 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31016.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:58:33,048 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.621e+02 1.904e+02 2.250e+02 3.149e+02, threshold=3.808e+02, percent-clipped=0.0 +2023-03-26 05:58:41,923 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9974, 1.7949, 1.5238, 1.7421, 1.8246, 1.7352, 1.6947, 2.5325], + device='cuda:4'), covar=tensor([0.6275, 0.6907, 0.5065, 0.6750, 0.5715, 0.3464, 0.6659, 0.2393], + device='cuda:4'), in_proj_covar=tensor([0.0284, 0.0259, 0.0220, 0.0285, 0.0241, 0.0203, 0.0247, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:58:53,279 INFO [finetune.py:976] (4/7) Epoch 6, batch 2400, loss[loss=0.1903, simple_loss=0.2533, pruned_loss=0.06364, over 4857.00 frames. ], tot_loss[loss=0.2121, simple_loss=0.2719, pruned_loss=0.07612, over 958730.39 frames. ], batch size: 49, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 05:59:17,403 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31064.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 05:59:18,657 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9934, 1.5908, 2.3065, 4.0304, 2.8212, 2.8121, 0.6770, 3.2153], + device='cuda:4'), covar=tensor([0.1839, 0.1700, 0.1580, 0.0548, 0.0803, 0.1573, 0.2367, 0.0485], + device='cuda:4'), in_proj_covar=tensor([0.0104, 0.0121, 0.0139, 0.0169, 0.0104, 0.0145, 0.0131, 0.0105], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 05:59:24,031 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31072.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 05:59:37,284 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6921, 1.5786, 1.3787, 1.7109, 2.0941, 1.7262, 1.2728, 1.3802], + device='cuda:4'), covar=tensor([0.2218, 0.2108, 0.1943, 0.1745, 0.1601, 0.1200, 0.2654, 0.1948], + device='cuda:4'), in_proj_covar=tensor([0.0235, 0.0209, 0.0204, 0.0186, 0.0237, 0.0176, 0.0214, 0.0190], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 05:59:44,206 INFO [finetune.py:976] (4/7) Epoch 6, batch 2450, loss[loss=0.1677, simple_loss=0.2295, pruned_loss=0.05292, over 4782.00 frames. ], tot_loss[loss=0.2091, simple_loss=0.2685, pruned_loss=0.07485, over 955565.23 frames. ], batch size: 26, lr: 3.91e-03, grad_scale: 32.0 +2023-03-26 06:00:16,635 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4381, 1.4764, 1.1990, 1.4147, 1.7418, 1.6424, 1.4749, 1.2310], + device='cuda:4'), covar=tensor([0.0345, 0.0298, 0.0640, 0.0312, 0.0251, 0.0496, 0.0297, 0.0412], + device='cuda:4'), in_proj_covar=tensor([0.0086, 0.0110, 0.0135, 0.0116, 0.0102, 0.0099, 0.0089, 0.0107], + device='cuda:4'), out_proj_covar=tensor([6.6860e-05, 8.6685e-05, 1.0852e-04, 9.1318e-05, 8.0133e-05, 7.3434e-05, + 6.8094e-05, 8.3780e-05], device='cuda:4') +2023-03-26 06:00:21,949 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.316e+02 1.762e+02 2.221e+02 2.552e+02 5.044e+02, threshold=4.442e+02, percent-clipped=4.0 +2023-03-26 06:00:30,922 INFO [finetune.py:976] (4/7) Epoch 6, batch 2500, loss[loss=0.2277, simple_loss=0.299, pruned_loss=0.07819, over 4721.00 frames. ], tot_loss[loss=0.2133, simple_loss=0.2725, pruned_loss=0.07704, over 957248.10 frames. ], batch size: 54, lr: 3.91e-03, grad_scale: 16.0 +2023-03-26 06:01:06,652 INFO [finetune.py:976] (4/7) Epoch 6, batch 2550, loss[loss=0.2844, simple_loss=0.3384, pruned_loss=0.1152, over 4867.00 frames. ], tot_loss[loss=0.2163, simple_loss=0.2764, pruned_loss=0.07807, over 957486.89 frames. ], batch size: 34, lr: 3.91e-03, grad_scale: 16.0 +2023-03-26 06:01:50,623 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.680e+02 2.034e+02 2.315e+02 3.655e+02, threshold=4.067e+02, percent-clipped=0.0 +2023-03-26 06:02:04,701 INFO [finetune.py:976] (4/7) Epoch 6, batch 2600, loss[loss=0.2295, simple_loss=0.288, pruned_loss=0.08551, over 4825.00 frames. ], tot_loss[loss=0.2171, simple_loss=0.2772, pruned_loss=0.07854, over 956663.21 frames. ], batch size: 38, lr: 3.91e-03, grad_scale: 16.0 +2023-03-26 06:02:23,273 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.66 vs. limit=5.0 +2023-03-26 06:02:33,431 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31259.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:02:48,069 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4942, 2.2091, 2.0881, 2.4377, 2.4634, 2.2423, 2.7514, 2.3491], + device='cuda:4'), covar=tensor([0.1301, 0.2507, 0.3084, 0.2622, 0.2185, 0.1486, 0.2553, 0.2002], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0192, 0.0235, 0.0253, 0.0230, 0.0190, 0.0211, 0.0191], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:03:04,533 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31281.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:03:09,263 INFO [finetune.py:976] (4/7) Epoch 6, batch 2650, loss[loss=0.1833, simple_loss=0.2454, pruned_loss=0.06059, over 4873.00 frames. ], tot_loss[loss=0.218, simple_loss=0.2784, pruned_loss=0.07881, over 957058.21 frames. ], batch size: 34, lr: 3.91e-03, grad_scale: 16.0 +2023-03-26 06:03:15,217 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31290.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:03:38,131 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31311.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:03:42,103 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.0194, 4.3772, 4.5552, 4.8132, 4.7623, 4.5007, 5.1341, 1.5966], + device='cuda:4'), covar=tensor([0.0675, 0.0714, 0.0704, 0.0814, 0.1075, 0.1480, 0.0504, 0.5428], + device='cuda:4'), in_proj_covar=tensor([0.0356, 0.0243, 0.0276, 0.0293, 0.0333, 0.0284, 0.0302, 0.0298], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:03:47,924 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.761e+02 2.106e+02 2.462e+02 3.966e+02, threshold=4.213e+02, percent-clipped=0.0 +2023-03-26 06:03:59,091 INFO [finetune.py:976] (4/7) Epoch 6, batch 2700, loss[loss=0.1533, simple_loss=0.2193, pruned_loss=0.04369, over 4794.00 frames. ], tot_loss[loss=0.2163, simple_loss=0.2771, pruned_loss=0.07774, over 957855.02 frames. ], batch size: 51, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:04:20,541 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 06:04:23,338 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31359.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:04:33,879 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31367.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 06:05:03,592 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1809, 2.0476, 1.6378, 2.1042, 2.1951, 1.7580, 2.4498, 2.0742], + device='cuda:4'), covar=tensor([0.1735, 0.3120, 0.4162, 0.3267, 0.2902, 0.2131, 0.3351, 0.2433], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0192, 0.0235, 0.0252, 0.0230, 0.0190, 0.0211, 0.0191], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:05:04,060 INFO [finetune.py:976] (4/7) Epoch 6, batch 2750, loss[loss=0.1916, simple_loss=0.2584, pruned_loss=0.06243, over 4754.00 frames. ], tot_loss[loss=0.2129, simple_loss=0.2735, pruned_loss=0.07618, over 955954.50 frames. ], batch size: 26, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:05:06,613 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7379, 3.0278, 2.4376, 1.9684, 3.1859, 3.0696, 2.9298, 2.6380], + device='cuda:4'), covar=tensor([0.0610, 0.0548, 0.0890, 0.0935, 0.0346, 0.0662, 0.0645, 0.0792], + device='cuda:4'), in_proj_covar=tensor([0.0140, 0.0135, 0.0145, 0.0129, 0.0114, 0.0145, 0.0146, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:05:28,059 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31409.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:05:46,890 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.163e+02 1.636e+02 1.989e+02 2.265e+02 3.886e+02, threshold=3.977e+02, percent-clipped=0.0 +2023-03-26 06:05:56,331 INFO [finetune.py:976] (4/7) Epoch 6, batch 2800, loss[loss=0.1647, simple_loss=0.2447, pruned_loss=0.04232, over 4905.00 frames. ], tot_loss[loss=0.2086, simple_loss=0.2691, pruned_loss=0.07404, over 957767.00 frames. ], batch size: 32, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:06:16,540 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31470.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:06:42,529 INFO [finetune.py:976] (4/7) Epoch 6, batch 2850, loss[loss=0.2296, simple_loss=0.2878, pruned_loss=0.0857, over 4821.00 frames. ], tot_loss[loss=0.2078, simple_loss=0.2682, pruned_loss=0.0737, over 958771.48 frames. ], batch size: 38, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:06:43,857 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31491.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 06:07:03,164 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7434, 1.7119, 1.7683, 1.1145, 1.8921, 1.8106, 1.7661, 1.4993], + device='cuda:4'), covar=tensor([0.0608, 0.0641, 0.0673, 0.0931, 0.0550, 0.0765, 0.0667, 0.1085], + device='cuda:4'), in_proj_covar=tensor([0.0141, 0.0135, 0.0146, 0.0130, 0.0114, 0.0145, 0.0147, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:07:26,132 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.153e+02 1.576e+02 1.974e+02 2.520e+02 5.037e+02, threshold=3.948e+02, percent-clipped=2.0 +2023-03-26 06:07:42,541 INFO [finetune.py:976] (4/7) Epoch 6, batch 2900, loss[loss=0.2399, simple_loss=0.3124, pruned_loss=0.08374, over 4942.00 frames. ], tot_loss[loss=0.2093, simple_loss=0.2699, pruned_loss=0.07436, over 953608.78 frames. ], batch size: 39, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:07:51,210 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=5.23 vs. limit=5.0 +2023-03-26 06:07:51,402 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7370, 3.6923, 3.4751, 1.6205, 3.7473, 2.9358, 0.7790, 2.5914], + device='cuda:4'), covar=tensor([0.2664, 0.1995, 0.1748, 0.3470, 0.1046, 0.0958, 0.4391, 0.1486], + device='cuda:4'), in_proj_covar=tensor([0.0157, 0.0172, 0.0164, 0.0129, 0.0156, 0.0123, 0.0146, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 06:07:59,543 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31552.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 06:08:04,308 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31559.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:08:18,610 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7233, 1.4672, 1.0650, 0.2953, 1.3069, 1.4761, 1.3743, 1.4236], + device='cuda:4'), covar=tensor([0.1048, 0.0879, 0.1499, 0.2184, 0.1435, 0.2849, 0.2524, 0.0987], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0202, 0.0201, 0.0190, 0.0217, 0.0209, 0.0221, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:08:19,183 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31581.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:08:25,535 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 06:08:27,345 INFO [finetune.py:976] (4/7) Epoch 6, batch 2950, loss[loss=0.2218, simple_loss=0.2817, pruned_loss=0.08091, over 4921.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.2741, pruned_loss=0.0758, over 954617.76 frames. ], batch size: 42, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:08:33,450 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31590.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:08:47,781 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31607.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:08:52,967 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8501, 1.7307, 1.5695, 1.9124, 2.3870, 1.8526, 1.5316, 1.4862], + device='cuda:4'), covar=tensor([0.2210, 0.2221, 0.1952, 0.1778, 0.1870, 0.1195, 0.2682, 0.1944], + device='cuda:4'), in_proj_covar=tensor([0.0236, 0.0210, 0.0203, 0.0186, 0.0238, 0.0176, 0.0213, 0.0191], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:08:59,953 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.780e+02 2.105e+02 2.565e+02 4.082e+02, threshold=4.210e+02, percent-clipped=1.0 +2023-03-26 06:09:02,950 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31629.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:09:09,349 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31638.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:09:09,898 INFO [finetune.py:976] (4/7) Epoch 6, batch 3000, loss[loss=0.137, simple_loss=0.2109, pruned_loss=0.03156, over 4768.00 frames. ], tot_loss[loss=0.2148, simple_loss=0.2758, pruned_loss=0.0769, over 954455.70 frames. ], batch size: 26, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:09:09,898 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 06:09:14,470 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5637, 1.2627, 1.3695, 1.3356, 1.6338, 1.6429, 1.5247, 1.2950], + device='cuda:4'), covar=tensor([0.0305, 0.0302, 0.0523, 0.0329, 0.0318, 0.0390, 0.0247, 0.0431], + device='cuda:4'), in_proj_covar=tensor([0.0086, 0.0111, 0.0137, 0.0116, 0.0103, 0.0100, 0.0090, 0.0108], + device='cuda:4'), out_proj_covar=tensor([6.7290e-05, 8.7190e-05, 1.0945e-04, 9.1587e-05, 8.1257e-05, 7.4104e-05, + 6.8327e-05, 8.4034e-05], device='cuda:4') +2023-03-26 06:09:23,095 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8566, 1.8789, 1.9378, 1.2227, 2.0023, 2.0286, 1.9464, 1.6386], + device='cuda:4'), covar=tensor([0.0668, 0.0592, 0.0688, 0.0958, 0.0638, 0.0741, 0.0591, 0.1132], + device='cuda:4'), in_proj_covar=tensor([0.0139, 0.0134, 0.0144, 0.0129, 0.0114, 0.0144, 0.0146, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:09:23,489 INFO [finetune.py:1010] (4/7) Epoch 6, validation: loss=0.1625, simple_loss=0.2344, pruned_loss=0.04534, over 2265189.00 frames. +2023-03-26 06:09:23,490 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6301MB +2023-03-26 06:09:55,808 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31667.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 06:10:19,289 INFO [finetune.py:976] (4/7) Epoch 6, batch 3050, loss[loss=0.2175, simple_loss=0.2842, pruned_loss=0.0754, over 4796.00 frames. ], tot_loss[loss=0.2142, simple_loss=0.2759, pruned_loss=0.0763, over 955251.45 frames. ], batch size: 29, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:10:27,505 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9884, 2.7523, 2.4174, 1.3956, 2.5648, 2.1501, 1.9414, 2.2580], + device='cuda:4'), covar=tensor([0.0932, 0.0963, 0.1986, 0.2507, 0.1843, 0.2136, 0.2288, 0.1317], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0202, 0.0202, 0.0190, 0.0217, 0.0209, 0.0221, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:10:29,944 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31704.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:10:31,215 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8801, 1.7235, 2.3430, 1.5542, 2.0791, 2.1222, 1.6545, 2.3115], + device='cuda:4'), covar=tensor([0.1582, 0.2063, 0.1462, 0.2209, 0.0925, 0.1627, 0.2812, 0.0916], + device='cuda:4'), in_proj_covar=tensor([0.0204, 0.0205, 0.0197, 0.0195, 0.0183, 0.0220, 0.0217, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:10:36,654 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=31715.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 06:10:43,058 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.708e+02 2.072e+02 2.420e+02 4.871e+02, threshold=4.144e+02, percent-clipped=1.0 +2023-03-26 06:10:59,684 INFO [finetune.py:976] (4/7) Epoch 6, batch 3100, loss[loss=0.169, simple_loss=0.2353, pruned_loss=0.0513, over 4806.00 frames. ], tot_loss[loss=0.2122, simple_loss=0.2737, pruned_loss=0.07537, over 957110.06 frames. ], batch size: 40, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:11:20,231 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31765.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:11:20,295 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31765.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:11:35,796 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31788.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:11:36,281 INFO [finetune.py:976] (4/7) Epoch 6, batch 3150, loss[loss=0.1814, simple_loss=0.2363, pruned_loss=0.0632, over 4852.00 frames. ], tot_loss[loss=0.2107, simple_loss=0.2715, pruned_loss=0.07497, over 956020.24 frames. ], batch size: 44, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:11:43,464 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7638, 1.5863, 2.3214, 3.5220, 2.5018, 2.6040, 0.9303, 2.7108], + device='cuda:4'), covar=tensor([0.1720, 0.1470, 0.1247, 0.0555, 0.0759, 0.1400, 0.2010, 0.0574], + device='cuda:4'), in_proj_covar=tensor([0.0104, 0.0120, 0.0138, 0.0169, 0.0103, 0.0144, 0.0131, 0.0104], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 06:11:44,299 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0989, 2.0818, 2.4135, 1.1021, 2.5147, 2.6849, 2.3163, 2.0880], + device='cuda:4'), covar=tensor([0.0991, 0.0776, 0.0400, 0.0717, 0.0424, 0.0651, 0.0430, 0.0602], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0159, 0.0121, 0.0138, 0.0134, 0.0125, 0.0148, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.8726e-05, 1.1750e-04, 8.7876e-05, 1.0097e-04, 9.6662e-05, 9.2555e-05, + 1.0968e-04, 1.0755e-04], device='cuda:4') +2023-03-26 06:12:00,909 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31823.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:12:01,993 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.748e+01 1.682e+02 2.020e+02 2.535e+02 4.605e+02, threshold=4.040e+02, percent-clipped=1.0 +2023-03-26 06:12:15,951 INFO [finetune.py:976] (4/7) Epoch 6, batch 3200, loss[loss=0.2093, simple_loss=0.269, pruned_loss=0.07477, over 4838.00 frames. ], tot_loss[loss=0.209, simple_loss=0.2686, pruned_loss=0.07467, over 956414.65 frames. ], batch size: 47, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:12:25,650 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31847.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 06:12:32,397 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31849.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:13:06,133 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31878.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:13:11,201 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31884.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:13:14,090 INFO [finetune.py:976] (4/7) Epoch 6, batch 3250, loss[loss=0.2506, simple_loss=0.3113, pruned_loss=0.09495, over 4738.00 frames. ], tot_loss[loss=0.2079, simple_loss=0.2681, pruned_loss=0.0739, over 955906.63 frames. ], batch size: 59, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:13:43,135 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6736, 3.6717, 3.4585, 1.6557, 3.6754, 2.6692, 0.8145, 2.4327], + device='cuda:4'), covar=tensor([0.2472, 0.1622, 0.1583, 0.3392, 0.0999, 0.1089, 0.4510, 0.1564], + device='cuda:4'), in_proj_covar=tensor([0.0155, 0.0170, 0.0162, 0.0127, 0.0155, 0.0122, 0.0144, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 06:13:51,967 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31915.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:14:01,796 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.169e+02 1.695e+02 2.048e+02 2.376e+02 4.231e+02, threshold=4.096e+02, percent-clipped=1.0 +2023-03-26 06:14:17,635 INFO [finetune.py:976] (4/7) Epoch 6, batch 3300, loss[loss=0.1826, simple_loss=0.2518, pruned_loss=0.05673, over 4894.00 frames. ], tot_loss[loss=0.2091, simple_loss=0.2701, pruned_loss=0.07409, over 954862.76 frames. ], batch size: 35, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:14:17,769 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31939.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:14:29,143 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31950.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:14:46,597 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31976.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:14:54,838 INFO [finetune.py:976] (4/7) Epoch 6, batch 3350, loss[loss=0.2235, simple_loss=0.2706, pruned_loss=0.08818, over 4154.00 frames. ], tot_loss[loss=0.2108, simple_loss=0.2721, pruned_loss=0.07477, over 953003.61 frames. ], batch size: 65, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:14:57,969 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.0281, 4.4060, 4.5077, 4.8806, 4.7084, 4.4644, 5.1390, 1.5003], + device='cuda:4'), covar=tensor([0.0663, 0.0827, 0.0732, 0.0794, 0.1212, 0.1403, 0.0510, 0.5372], + device='cuda:4'), in_proj_covar=tensor([0.0355, 0.0243, 0.0276, 0.0293, 0.0333, 0.0284, 0.0303, 0.0298], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:15:11,491 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=32011.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:15:22,687 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.186e+02 1.743e+02 2.097e+02 2.540e+02 4.089e+02, threshold=4.194e+02, percent-clipped=0.0 +2023-03-26 06:15:41,805 INFO [finetune.py:976] (4/7) Epoch 6, batch 3400, loss[loss=0.1702, simple_loss=0.2268, pruned_loss=0.05678, over 4437.00 frames. ], tot_loss[loss=0.2125, simple_loss=0.2739, pruned_loss=0.07557, over 954213.59 frames. ], batch size: 19, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:16:04,437 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32060.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:16:13,850 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32065.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:16:41,405 INFO [finetune.py:976] (4/7) Epoch 6, batch 3450, loss[loss=0.2847, simple_loss=0.3202, pruned_loss=0.1246, over 4732.00 frames. ], tot_loss[loss=0.2131, simple_loss=0.2749, pruned_loss=0.07563, over 953131.68 frames. ], batch size: 54, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:16:52,470 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7230, 1.5250, 1.3461, 1.0952, 1.4966, 1.4601, 1.4713, 2.0953], + device='cuda:4'), covar=tensor([0.6129, 0.5837, 0.4623, 0.5698, 0.5470, 0.3068, 0.5230, 0.2295], + device='cuda:4'), in_proj_covar=tensor([0.0282, 0.0258, 0.0220, 0.0284, 0.0239, 0.0203, 0.0245, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:17:04,770 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0326, 1.8982, 1.6011, 1.9647, 2.0282, 1.6702, 2.3088, 2.0179], + device='cuda:4'), covar=tensor([0.1491, 0.2796, 0.3415, 0.2971, 0.2669, 0.1906, 0.3772, 0.2059], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0192, 0.0235, 0.0253, 0.0231, 0.0191, 0.0211, 0.0191], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:17:07,736 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32113.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:17:16,936 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.122e+02 1.573e+02 1.926e+02 2.516e+02 4.351e+02, threshold=3.853e+02, percent-clipped=2.0 +2023-03-26 06:17:18,899 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1350, 2.0034, 1.5856, 1.9858, 2.0776, 1.6976, 2.4354, 2.0141], + device='cuda:4'), covar=tensor([0.1459, 0.3055, 0.3722, 0.3345, 0.2976, 0.1931, 0.3462, 0.2129], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0191, 0.0235, 0.0253, 0.0231, 0.0191, 0.0211, 0.0191], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:17:25,471 INFO [finetune.py:976] (4/7) Epoch 6, batch 3500, loss[loss=0.2064, simple_loss=0.2598, pruned_loss=0.07645, over 4906.00 frames. ], tot_loss[loss=0.2102, simple_loss=0.2716, pruned_loss=0.07437, over 954282.61 frames. ], batch size: 36, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:17:29,080 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32144.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:17:30,913 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32147.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 06:17:52,518 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6990, 2.4914, 2.0982, 1.1541, 2.2808, 1.9793, 1.8648, 2.0841], + device='cuda:4'), covar=tensor([0.0781, 0.0788, 0.1479, 0.2074, 0.1529, 0.2067, 0.2188, 0.1144], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0202, 0.0200, 0.0189, 0.0216, 0.0208, 0.0220, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:18:09,966 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32179.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:18:15,992 INFO [finetune.py:976] (4/7) Epoch 6, batch 3550, loss[loss=0.2321, simple_loss=0.2867, pruned_loss=0.08873, over 4805.00 frames. ], tot_loss[loss=0.2078, simple_loss=0.2689, pruned_loss=0.07335, over 956725.64 frames. ], batch size: 25, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:18:16,722 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0553, 1.7610, 2.2107, 1.6164, 1.9948, 2.1739, 1.8710, 2.3770], + device='cuda:4'), covar=tensor([0.1179, 0.1828, 0.1323, 0.1649, 0.0999, 0.1215, 0.2341, 0.0719], + device='cuda:4'), in_proj_covar=tensor([0.0205, 0.0206, 0.0199, 0.0196, 0.0185, 0.0222, 0.0218, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:18:19,694 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32195.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 06:18:21,413 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2157, 1.6038, 0.6188, 2.0875, 2.3523, 1.7652, 1.7657, 1.8711], + device='cuda:4'), covar=tensor([0.2255, 0.2952, 0.3346, 0.1686, 0.2548, 0.2540, 0.2128, 0.3045], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0098, 0.0115, 0.0092, 0.0124, 0.0096, 0.0100, 0.0093], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 06:18:40,465 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.617e+02 1.889e+02 2.318e+02 4.823e+02, threshold=3.777e+02, percent-clipped=2.0 +2023-03-26 06:18:52,027 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32234.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:19:00,026 INFO [finetune.py:976] (4/7) Epoch 6, batch 3600, loss[loss=0.2043, simple_loss=0.2628, pruned_loss=0.07293, over 4927.00 frames. ], tot_loss[loss=0.2082, simple_loss=0.2688, pruned_loss=0.07382, over 957076.23 frames. ], batch size: 33, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:19:19,279 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 06:19:36,023 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32271.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:19:36,285 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 06:19:57,774 INFO [finetune.py:976] (4/7) Epoch 6, batch 3650, loss[loss=0.1949, simple_loss=0.2585, pruned_loss=0.06567, over 4927.00 frames. ], tot_loss[loss=0.2093, simple_loss=0.2699, pruned_loss=0.07431, over 953762.98 frames. ], batch size: 33, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:20:14,269 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32306.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:20:26,743 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.020e+02 1.776e+02 2.129e+02 2.587e+02 5.126e+02, threshold=4.259e+02, percent-clipped=5.0 +2023-03-26 06:20:32,415 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7063, 4.0095, 3.7370, 1.8365, 4.1385, 3.0685, 0.8226, 2.7701], + device='cuda:4'), covar=tensor([0.2994, 0.2219, 0.1739, 0.3848, 0.1067, 0.1028, 0.4858, 0.1726], + device='cuda:4'), in_proj_covar=tensor([0.0156, 0.0172, 0.0163, 0.0129, 0.0156, 0.0123, 0.0146, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 06:20:43,055 INFO [finetune.py:976] (4/7) Epoch 6, batch 3700, loss[loss=0.1483, simple_loss=0.2156, pruned_loss=0.04047, over 4713.00 frames. ], tot_loss[loss=0.2133, simple_loss=0.2744, pruned_loss=0.07609, over 954341.47 frames. ], batch size: 23, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:20:43,177 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1260, 2.0551, 2.0395, 1.4582, 2.2597, 2.2753, 2.2108, 1.8096], + device='cuda:4'), covar=tensor([0.0611, 0.0651, 0.0736, 0.0976, 0.0496, 0.0634, 0.0611, 0.1015], + device='cuda:4'), in_proj_covar=tensor([0.0140, 0.0136, 0.0145, 0.0129, 0.0114, 0.0145, 0.0147, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:20:48,068 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8309, 1.7940, 1.6783, 1.7949, 1.4933, 3.7205, 1.7836, 2.3466], + device='cuda:4'), covar=tensor([0.3360, 0.2380, 0.1963, 0.2178, 0.1698, 0.0215, 0.2655, 0.1221], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0115, 0.0119, 0.0123, 0.0118, 0.0100, 0.0102, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 06:20:56,545 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32360.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:21:16,590 INFO [finetune.py:976] (4/7) Epoch 6, batch 3750, loss[loss=0.2055, simple_loss=0.2676, pruned_loss=0.07175, over 4692.00 frames. ], tot_loss[loss=0.2146, simple_loss=0.2756, pruned_loss=0.07675, over 954508.61 frames. ], batch size: 59, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:21:26,320 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0315, 1.9793, 2.1511, 0.8875, 2.2017, 2.4863, 2.0995, 2.0177], + device='cuda:4'), covar=tensor([0.1006, 0.0759, 0.0485, 0.0813, 0.0681, 0.0612, 0.0530, 0.0654], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0157, 0.0121, 0.0138, 0.0133, 0.0125, 0.0147, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.7452e-05, 1.1588e-04, 8.7807e-05, 1.0036e-04, 9.5856e-05, 9.2104e-05, + 1.0887e-04, 1.0723e-04], device='cuda:4') +2023-03-26 06:21:37,054 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32408.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:21:48,323 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.626e+02 1.929e+02 2.294e+02 3.909e+02, threshold=3.857e+02, percent-clipped=0.0 +2023-03-26 06:21:58,659 INFO [finetune.py:976] (4/7) Epoch 6, batch 3800, loss[loss=0.1866, simple_loss=0.2613, pruned_loss=0.0559, over 4760.00 frames. ], tot_loss[loss=0.2159, simple_loss=0.2775, pruned_loss=0.07719, over 955820.81 frames. ], batch size: 28, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:22:01,797 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32444.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:22:10,514 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-03-26 06:22:24,392 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32479.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:22:31,248 INFO [finetune.py:976] (4/7) Epoch 6, batch 3850, loss[loss=0.1878, simple_loss=0.2564, pruned_loss=0.05961, over 4903.00 frames. ], tot_loss[loss=0.2152, simple_loss=0.2766, pruned_loss=0.07687, over 957732.75 frames. ], batch size: 43, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:22:31,364 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6210, 1.5519, 1.5498, 1.6218, 1.2065, 2.7847, 1.2893, 1.7992], + device='cuda:4'), covar=tensor([0.2860, 0.2077, 0.1768, 0.2005, 0.1695, 0.0311, 0.2788, 0.1171], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0115, 0.0119, 0.0122, 0.0118, 0.0099, 0.0102, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 06:22:33,636 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32492.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:22:53,999 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.617e+01 1.620e+02 2.056e+02 2.694e+02 5.558e+02, threshold=4.113e+02, percent-clipped=4.0 +2023-03-26 06:22:54,225 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-26 06:22:55,288 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32527.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:22:57,083 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7005, 2.4677, 2.0500, 1.0852, 2.2541, 2.0555, 1.7925, 2.0959], + device='cuda:4'), covar=tensor([0.0911, 0.0893, 0.1795, 0.2240, 0.1875, 0.2386, 0.2332, 0.1119], + device='cuda:4'), in_proj_covar=tensor([0.0167, 0.0200, 0.0199, 0.0188, 0.0214, 0.0207, 0.0220, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:23:00,099 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32534.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:23:04,462 INFO [finetune.py:976] (4/7) Epoch 6, batch 3900, loss[loss=0.1759, simple_loss=0.2472, pruned_loss=0.05229, over 4835.00 frames. ], tot_loss[loss=0.2143, simple_loss=0.2751, pruned_loss=0.07677, over 956865.87 frames. ], batch size: 47, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:23:24,750 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32571.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:23:31,335 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32582.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:23:36,052 INFO [finetune.py:976] (4/7) Epoch 6, batch 3950, loss[loss=0.2324, simple_loss=0.2859, pruned_loss=0.08949, over 4829.00 frames. ], tot_loss[loss=0.2111, simple_loss=0.2711, pruned_loss=0.07554, over 957286.19 frames. ], batch size: 30, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:23:59,348 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32606.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:24:07,665 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32619.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:24:11,251 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.084e+02 1.672e+02 2.141e+02 2.509e+02 4.478e+02, threshold=4.281e+02, percent-clipped=2.0 +2023-03-26 06:24:20,842 INFO [finetune.py:976] (4/7) Epoch 6, batch 4000, loss[loss=0.2624, simple_loss=0.308, pruned_loss=0.1084, over 4820.00 frames. ], tot_loss[loss=0.2102, simple_loss=0.2697, pruned_loss=0.07534, over 957123.25 frames. ], batch size: 38, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:24:30,955 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=32654.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:25:04,478 INFO [finetune.py:976] (4/7) Epoch 6, batch 4050, loss[loss=0.2135, simple_loss=0.28, pruned_loss=0.07347, over 4826.00 frames. ], tot_loss[loss=0.2137, simple_loss=0.2734, pruned_loss=0.07697, over 955768.06 frames. ], batch size: 47, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:25:10,359 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7690, 3.7646, 3.7843, 2.2772, 3.8722, 3.0491, 1.2993, 2.8427], + device='cuda:4'), covar=tensor([0.2953, 0.1763, 0.1328, 0.2692, 0.0954, 0.0882, 0.3778, 0.1365], + device='cuda:4'), in_proj_covar=tensor([0.0155, 0.0171, 0.0162, 0.0128, 0.0155, 0.0122, 0.0145, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 06:25:28,436 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.148e+02 1.831e+02 2.133e+02 2.637e+02 5.226e+02, threshold=4.267e+02, percent-clipped=1.0 +2023-03-26 06:25:42,717 INFO [finetune.py:976] (4/7) Epoch 6, batch 4100, loss[loss=0.2344, simple_loss=0.2942, pruned_loss=0.08732, over 4912.00 frames. ], tot_loss[loss=0.2135, simple_loss=0.2737, pruned_loss=0.07666, over 954640.48 frames. ], batch size: 38, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:26:18,423 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-03-26 06:26:34,204 INFO [finetune.py:976] (4/7) Epoch 6, batch 4150, loss[loss=0.2837, simple_loss=0.3355, pruned_loss=0.1159, over 4807.00 frames. ], tot_loss[loss=0.2159, simple_loss=0.2763, pruned_loss=0.07779, over 953693.05 frames. ], batch size: 38, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:26:36,224 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 06:27:07,571 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.72 vs. limit=5.0 +2023-03-26 06:27:16,106 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-26 06:27:18,097 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.203e+02 1.784e+02 2.149e+02 2.523e+02 4.029e+02, threshold=4.299e+02, percent-clipped=0.0 +2023-03-26 06:27:37,293 INFO [finetune.py:976] (4/7) Epoch 6, batch 4200, loss[loss=0.2275, simple_loss=0.2829, pruned_loss=0.08608, over 4829.00 frames. ], tot_loss[loss=0.2151, simple_loss=0.2763, pruned_loss=0.07699, over 954532.11 frames. ], batch size: 44, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:27:40,598 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 06:27:48,563 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7511, 2.3895, 3.1343, 4.5404, 3.2564, 3.2035, 1.3583, 3.5824], + device='cuda:4'), covar=tensor([0.1495, 0.1316, 0.1228, 0.0463, 0.0704, 0.1201, 0.1904, 0.0577], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0118, 0.0135, 0.0166, 0.0102, 0.0141, 0.0129, 0.0103], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 06:28:34,721 INFO [finetune.py:976] (4/7) Epoch 6, batch 4250, loss[loss=0.2007, simple_loss=0.2706, pruned_loss=0.06534, over 4850.00 frames. ], tot_loss[loss=0.2122, simple_loss=0.2735, pruned_loss=0.07547, over 953606.01 frames. ], batch size: 47, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:28:55,409 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 06:29:03,430 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.76 vs. limit=5.0 +2023-03-26 06:29:25,480 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.595e+02 1.920e+02 2.303e+02 3.727e+02, threshold=3.841e+02, percent-clipped=0.0 +2023-03-26 06:29:44,594 INFO [finetune.py:976] (4/7) Epoch 6, batch 4300, loss[loss=0.1803, simple_loss=0.2467, pruned_loss=0.05688, over 4933.00 frames. ], tot_loss[loss=0.2108, simple_loss=0.271, pruned_loss=0.07529, over 955301.81 frames. ], batch size: 38, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:29:57,498 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8498, 4.0272, 3.8239, 1.9819, 4.1631, 3.0181, 0.7868, 2.9232], + device='cuda:4'), covar=tensor([0.2350, 0.1940, 0.1637, 0.3477, 0.0889, 0.1083, 0.4829, 0.1549], + device='cuda:4'), in_proj_covar=tensor([0.0155, 0.0172, 0.0162, 0.0128, 0.0156, 0.0122, 0.0146, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 06:30:17,445 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=32966.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:30:46,856 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 06:30:47,044 INFO [finetune.py:976] (4/7) Epoch 6, batch 4350, loss[loss=0.2335, simple_loss=0.2901, pruned_loss=0.08843, over 4803.00 frames. ], tot_loss[loss=0.2096, simple_loss=0.2693, pruned_loss=0.07495, over 956120.53 frames. ], batch size: 45, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:31:32,014 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.139e+02 1.730e+02 2.041e+02 2.589e+02 3.941e+02, threshold=4.082e+02, percent-clipped=1.0 +2023-03-26 06:31:38,498 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33027.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:31:50,661 INFO [finetune.py:976] (4/7) Epoch 6, batch 4400, loss[loss=0.1849, simple_loss=0.2363, pruned_loss=0.06676, over 4307.00 frames. ], tot_loss[loss=0.2117, simple_loss=0.2717, pruned_loss=0.07585, over 956208.20 frames. ], batch size: 18, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:32:31,749 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33069.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:32:33,604 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3539, 1.9323, 1.5398, 0.7549, 1.7732, 1.8302, 1.4722, 1.7425], + device='cuda:4'), covar=tensor([0.0825, 0.1003, 0.1619, 0.2141, 0.1484, 0.2356, 0.2480, 0.1061], + device='cuda:4'), in_proj_covar=tensor([0.0167, 0.0199, 0.0199, 0.0188, 0.0214, 0.0206, 0.0218, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:32:42,458 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-03-26 06:32:54,584 INFO [finetune.py:976] (4/7) Epoch 6, batch 4450, loss[loss=0.2351, simple_loss=0.2963, pruned_loss=0.08702, over 4810.00 frames. ], tot_loss[loss=0.2141, simple_loss=0.2745, pruned_loss=0.07682, over 954962.85 frames. ], batch size: 40, lr: 3.90e-03, grad_scale: 16.0 +2023-03-26 06:32:56,370 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5149, 1.3939, 1.9320, 2.9283, 1.9799, 2.1045, 1.1266, 2.3217], + device='cuda:4'), covar=tensor([0.1900, 0.1568, 0.1231, 0.0656, 0.0878, 0.1447, 0.1704, 0.0668], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0118, 0.0135, 0.0165, 0.0102, 0.0141, 0.0128, 0.0103], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 06:33:39,236 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.174e+02 1.714e+02 2.125e+02 2.690e+02 4.211e+02, threshold=4.250e+02, percent-clipped=1.0 +2023-03-26 06:33:48,022 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33130.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:33:58,505 INFO [finetune.py:976] (4/7) Epoch 6, batch 4500, loss[loss=0.2378, simple_loss=0.2899, pruned_loss=0.09289, over 4917.00 frames. ], tot_loss[loss=0.2155, simple_loss=0.2763, pruned_loss=0.07738, over 956650.17 frames. ], batch size: 33, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:34:19,016 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33155.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:34:53,561 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 06:35:01,059 INFO [finetune.py:976] (4/7) Epoch 6, batch 4550, loss[loss=0.186, simple_loss=0.2507, pruned_loss=0.06058, over 4829.00 frames. ], tot_loss[loss=0.216, simple_loss=0.2772, pruned_loss=0.07741, over 954349.77 frames. ], batch size: 30, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:35:03,044 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0269, 2.0116, 2.6703, 1.6447, 2.2242, 2.2361, 1.8428, 2.5537], + device='cuda:4'), covar=tensor([0.1513, 0.1894, 0.1544, 0.2340, 0.0996, 0.1717, 0.2548, 0.0933], + device='cuda:4'), in_proj_covar=tensor([0.0203, 0.0204, 0.0198, 0.0195, 0.0182, 0.0219, 0.0216, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:35:11,971 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4701, 2.2762, 1.8054, 0.9516, 1.9721, 1.9057, 1.6881, 1.8589], + device='cuda:4'), covar=tensor([0.0776, 0.0850, 0.1714, 0.2102, 0.1538, 0.2164, 0.2244, 0.1214], + device='cuda:4'), in_proj_covar=tensor([0.0166, 0.0200, 0.0199, 0.0187, 0.0214, 0.0206, 0.0218, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:35:20,248 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9346, 1.7686, 1.4898, 1.8022, 1.9223, 1.6182, 2.3126, 1.9065], + device='cuda:4'), covar=tensor([0.1678, 0.3011, 0.3862, 0.3313, 0.3037, 0.1937, 0.3369, 0.2336], + device='cuda:4'), in_proj_covar=tensor([0.0171, 0.0192, 0.0237, 0.0256, 0.0233, 0.0192, 0.0213, 0.0192], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:35:33,158 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33216.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:35:44,646 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.755e+02 2.084e+02 2.335e+02 4.621e+02, threshold=4.168e+02, percent-clipped=2.0 +2023-03-26 06:35:52,799 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.2646, 4.5034, 4.8002, 5.1354, 4.9333, 4.6528, 5.3813, 1.7058], + device='cuda:4'), covar=tensor([0.0728, 0.0807, 0.0721, 0.0771, 0.1270, 0.1411, 0.0514, 0.5302], + device='cuda:4'), in_proj_covar=tensor([0.0349, 0.0238, 0.0271, 0.0290, 0.0330, 0.0280, 0.0300, 0.0294], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:36:05,046 INFO [finetune.py:976] (4/7) Epoch 6, batch 4600, loss[loss=0.1802, simple_loss=0.2458, pruned_loss=0.05734, over 4792.00 frames. ], tot_loss[loss=0.2145, simple_loss=0.2764, pruned_loss=0.07636, over 954596.11 frames. ], batch size: 45, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:36:26,342 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5201, 2.2861, 1.9361, 1.0184, 2.1099, 1.9410, 1.6947, 1.9711], + device='cuda:4'), covar=tensor([0.0803, 0.0838, 0.1578, 0.2126, 0.1628, 0.2359, 0.2301, 0.1197], + device='cuda:4'), in_proj_covar=tensor([0.0166, 0.0199, 0.0198, 0.0187, 0.0213, 0.0205, 0.0217, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:37:07,546 INFO [finetune.py:976] (4/7) Epoch 6, batch 4650, loss[loss=0.1884, simple_loss=0.244, pruned_loss=0.06637, over 4708.00 frames. ], tot_loss[loss=0.211, simple_loss=0.2723, pruned_loss=0.07487, over 954175.30 frames. ], batch size: 23, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:37:48,443 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33322.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:37:50,592 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.141e+02 1.572e+02 1.961e+02 2.434e+02 3.752e+02, threshold=3.921e+02, percent-clipped=0.0 +2023-03-26 06:37:57,761 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0910, 1.9116, 1.5848, 2.1022, 2.0670, 1.7968, 2.4077, 2.0941], + device='cuda:4'), covar=tensor([0.1575, 0.3148, 0.3865, 0.3250, 0.2778, 0.1801, 0.3937, 0.2252], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0191, 0.0235, 0.0254, 0.0232, 0.0191, 0.0212, 0.0191], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:38:10,951 INFO [finetune.py:976] (4/7) Epoch 6, batch 4700, loss[loss=0.1671, simple_loss=0.212, pruned_loss=0.06116, over 4304.00 frames. ], tot_loss[loss=0.2083, simple_loss=0.2688, pruned_loss=0.07386, over 954727.51 frames. ], batch size: 65, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:39:20,542 INFO [finetune.py:976] (4/7) Epoch 6, batch 4750, loss[loss=0.2049, simple_loss=0.263, pruned_loss=0.07339, over 4386.00 frames. ], tot_loss[loss=0.2062, simple_loss=0.2667, pruned_loss=0.07287, over 955051.17 frames. ], batch size: 19, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:40:04,162 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.633e+02 1.917e+02 2.350e+02 3.562e+02, threshold=3.835e+02, percent-clipped=0.0 +2023-03-26 06:40:04,244 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33425.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:40:24,230 INFO [finetune.py:976] (4/7) Epoch 6, batch 4800, loss[loss=0.2316, simple_loss=0.3022, pruned_loss=0.08048, over 4900.00 frames. ], tot_loss[loss=0.2089, simple_loss=0.2691, pruned_loss=0.0744, over 955317.79 frames. ], batch size: 36, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:41:07,489 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33472.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:41:28,611 INFO [finetune.py:976] (4/7) Epoch 6, batch 4850, loss[loss=0.2046, simple_loss=0.2625, pruned_loss=0.07332, over 4897.00 frames. ], tot_loss[loss=0.2114, simple_loss=0.2723, pruned_loss=0.07525, over 955742.24 frames. ], batch size: 32, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:41:29,359 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0428, 1.8923, 1.5230, 1.8352, 1.9737, 1.7121, 2.2452, 1.9767], + device='cuda:4'), covar=tensor([0.1570, 0.2725, 0.3614, 0.2962, 0.2921, 0.1810, 0.3430, 0.2160], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0192, 0.0236, 0.0255, 0.0233, 0.0192, 0.0213, 0.0192], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:41:42,310 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6559, 1.2377, 0.8173, 1.5490, 1.9662, 1.2249, 1.4918, 1.4739], + device='cuda:4'), covar=tensor([0.1476, 0.2042, 0.2114, 0.1180, 0.2019, 0.1950, 0.1384, 0.2020], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0098, 0.0114, 0.0093, 0.0123, 0.0096, 0.0100, 0.0093], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 06:41:59,902 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33511.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:42:14,022 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.198e+02 1.841e+02 2.130e+02 2.477e+02 4.983e+02, threshold=4.260e+02, percent-clipped=3.0 +2023-03-26 06:42:24,108 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33533.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:42:32,810 INFO [finetune.py:976] (4/7) Epoch 6, batch 4900, loss[loss=0.206, simple_loss=0.2734, pruned_loss=0.06928, over 4920.00 frames. ], tot_loss[loss=0.2132, simple_loss=0.2743, pruned_loss=0.07611, over 956232.12 frames. ], batch size: 41, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:43:24,806 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2210, 1.9924, 2.5640, 4.0743, 2.8038, 2.6945, 0.8551, 3.3582], + device='cuda:4'), covar=tensor([0.1731, 0.1406, 0.1398, 0.0483, 0.0789, 0.1539, 0.2249, 0.0499], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0119, 0.0136, 0.0167, 0.0103, 0.0141, 0.0129, 0.0103], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 06:43:36,451 INFO [finetune.py:976] (4/7) Epoch 6, batch 4950, loss[loss=0.2294, simple_loss=0.2934, pruned_loss=0.08271, over 4796.00 frames. ], tot_loss[loss=0.216, simple_loss=0.2774, pruned_loss=0.07731, over 955104.25 frames. ], batch size: 45, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:43:51,065 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-03-26 06:43:56,952 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8078, 3.8225, 3.6568, 1.7251, 3.9375, 2.8149, 0.7741, 2.8231], + device='cuda:4'), covar=tensor([0.2344, 0.1730, 0.1574, 0.3234, 0.1057, 0.1032, 0.4484, 0.1264], + device='cuda:4'), in_proj_covar=tensor([0.0155, 0.0172, 0.0163, 0.0129, 0.0156, 0.0123, 0.0146, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 06:44:20,464 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=33622.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:44:22,687 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.562e+01 1.670e+02 2.087e+02 2.382e+02 5.310e+02, threshold=4.173e+02, percent-clipped=3.0 +2023-03-26 06:44:41,929 INFO [finetune.py:976] (4/7) Epoch 6, batch 5000, loss[loss=0.1862, simple_loss=0.2493, pruned_loss=0.06159, over 4844.00 frames. ], tot_loss[loss=0.2134, simple_loss=0.2749, pruned_loss=0.07598, over 956284.98 frames. ], batch size: 31, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:45:14,312 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=33670.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:45:19,042 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33677.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:45:22,674 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 06:45:26,106 INFO [finetune.py:976] (4/7) Epoch 6, batch 5050, loss[loss=0.1953, simple_loss=0.258, pruned_loss=0.06633, over 4869.00 frames. ], tot_loss[loss=0.2116, simple_loss=0.2723, pruned_loss=0.07541, over 956575.31 frames. ], batch size: 34, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:45:50,301 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.621e+02 1.877e+02 2.380e+02 3.773e+02, threshold=3.754e+02, percent-clipped=0.0 +2023-03-26 06:45:50,420 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=33725.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:45:58,832 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33738.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 06:45:59,300 INFO [finetune.py:976] (4/7) Epoch 6, batch 5100, loss[loss=0.2038, simple_loss=0.272, pruned_loss=0.06783, over 4789.00 frames. ], tot_loss[loss=0.2069, simple_loss=0.2682, pruned_loss=0.07278, over 957164.96 frames. ], batch size: 28, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:46:21,522 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.97 vs. limit=2.0 +2023-03-26 06:46:22,565 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=33773.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:46:32,750 INFO [finetune.py:976] (4/7) Epoch 6, batch 5150, loss[loss=0.2118, simple_loss=0.282, pruned_loss=0.07079, over 4901.00 frames. ], tot_loss[loss=0.2066, simple_loss=0.2676, pruned_loss=0.07285, over 957503.64 frames. ], batch size: 43, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:46:40,798 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9666, 2.1868, 1.6733, 1.5588, 2.2513, 2.2548, 2.2130, 1.9334], + device='cuda:4'), covar=tensor([0.0361, 0.0322, 0.0521, 0.0362, 0.0266, 0.0765, 0.0241, 0.0345], + device='cuda:4'), in_proj_covar=tensor([0.0086, 0.0109, 0.0136, 0.0115, 0.0103, 0.0098, 0.0089, 0.0107], + device='cuda:4'), out_proj_covar=tensor([6.7220e-05, 8.5970e-05, 1.0877e-04, 9.0466e-05, 8.0896e-05, 7.3130e-05, + 6.7749e-05, 8.3543e-05], device='cuda:4') +2023-03-26 06:46:45,128 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 06:46:48,267 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=33811.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:46:48,914 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7906, 1.2305, 1.7328, 1.6513, 1.4481, 1.5065, 1.6063, 1.5891], + device='cuda:4'), covar=tensor([0.5317, 0.7042, 0.5331, 0.6072, 0.7344, 0.5383, 0.7735, 0.5302], + device='cuda:4'), in_proj_covar=tensor([0.0229, 0.0245, 0.0254, 0.0256, 0.0242, 0.0219, 0.0272, 0.0225], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:46:57,550 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.284e+02 1.768e+02 2.107e+02 2.614e+02 3.782e+02, threshold=4.214e+02, percent-clipped=1.0 +2023-03-26 06:46:59,465 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33828.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:47:06,626 INFO [finetune.py:976] (4/7) Epoch 6, batch 5200, loss[loss=0.2107, simple_loss=0.2771, pruned_loss=0.07216, over 4821.00 frames. ], tot_loss[loss=0.21, simple_loss=0.2719, pruned_loss=0.07407, over 958334.79 frames. ], batch size: 30, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:47:19,509 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=33859.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:47:29,211 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9960, 1.6847, 2.4922, 4.1394, 2.8616, 2.8006, 0.7922, 3.3396], + device='cuda:4'), covar=tensor([0.2000, 0.1764, 0.1609, 0.0509, 0.0850, 0.1470, 0.2295, 0.0558], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0118, 0.0136, 0.0166, 0.0102, 0.0141, 0.0128, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 06:47:44,971 INFO [finetune.py:976] (4/7) Epoch 6, batch 5250, loss[loss=0.2421, simple_loss=0.3021, pruned_loss=0.09106, over 4847.00 frames. ], tot_loss[loss=0.2129, simple_loss=0.275, pruned_loss=0.0754, over 957816.51 frames. ], batch size: 44, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:48:10,022 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.099e+02 1.728e+02 2.112e+02 2.559e+02 4.196e+02, threshold=4.224e+02, percent-clipped=0.0 +2023-03-26 06:48:18,715 INFO [finetune.py:976] (4/7) Epoch 6, batch 5300, loss[loss=0.2588, simple_loss=0.3137, pruned_loss=0.102, over 4828.00 frames. ], tot_loss[loss=0.2153, simple_loss=0.2771, pruned_loss=0.07678, over 956756.68 frames. ], batch size: 30, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:48:19,402 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33939.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:48:53,986 INFO [finetune.py:976] (4/7) Epoch 6, batch 5350, loss[loss=0.1975, simple_loss=0.2678, pruned_loss=0.06361, over 4894.00 frames. ], tot_loss[loss=0.2156, simple_loss=0.2774, pruned_loss=0.07687, over 957528.51 frames. ], batch size: 32, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:49:07,570 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34000.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:49:40,310 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.730e+02 2.013e+02 2.437e+02 5.230e+02, threshold=4.026e+02, percent-clipped=3.0 +2023-03-26 06:49:50,511 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34033.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 06:49:59,344 INFO [finetune.py:976] (4/7) Epoch 6, batch 5400, loss[loss=0.2077, simple_loss=0.2575, pruned_loss=0.07892, over 4769.00 frames. ], tot_loss[loss=0.2136, simple_loss=0.2752, pruned_loss=0.07603, over 958402.53 frames. ], batch size: 26, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:50:02,489 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34044.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:50:22,313 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1356, 3.5568, 3.7526, 3.9558, 3.9142, 3.6677, 4.2199, 1.4243], + device='cuda:4'), covar=tensor([0.0731, 0.0892, 0.0732, 0.0998, 0.1028, 0.1181, 0.0611, 0.4944], + device='cuda:4'), in_proj_covar=tensor([0.0354, 0.0242, 0.0274, 0.0294, 0.0334, 0.0285, 0.0303, 0.0297], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:50:51,576 INFO [finetune.py:976] (4/7) Epoch 6, batch 5450, loss[loss=0.1857, simple_loss=0.2538, pruned_loss=0.05881, over 4827.00 frames. ], tot_loss[loss=0.21, simple_loss=0.2709, pruned_loss=0.07458, over 957839.32 frames. ], batch size: 33, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:51:04,652 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34105.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:51:17,056 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.064e+02 1.588e+02 1.777e+02 2.163e+02 4.002e+02, threshold=3.553e+02, percent-clipped=0.0 +2023-03-26 06:51:19,939 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34128.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:51:27,297 INFO [finetune.py:976] (4/7) Epoch 6, batch 5500, loss[loss=0.1819, simple_loss=0.2406, pruned_loss=0.06158, over 4801.00 frames. ], tot_loss[loss=0.207, simple_loss=0.2673, pruned_loss=0.07335, over 957367.64 frames. ], batch size: 26, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:51:31,025 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34145.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:51:34,018 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9645, 1.8429, 1.5149, 1.7575, 1.7317, 1.7095, 1.6409, 2.5410], + device='cuda:4'), covar=tensor([0.5811, 0.6475, 0.4842, 0.6451, 0.5831, 0.3432, 0.6150, 0.2249], + device='cuda:4'), in_proj_covar=tensor([0.0281, 0.0258, 0.0221, 0.0283, 0.0239, 0.0203, 0.0245, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:51:50,703 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34176.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:52:00,938 INFO [finetune.py:976] (4/7) Epoch 6, batch 5550, loss[loss=0.2247, simple_loss=0.288, pruned_loss=0.08073, over 4904.00 frames. ], tot_loss[loss=0.2108, simple_loss=0.2707, pruned_loss=0.07546, over 955535.55 frames. ], batch size: 32, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:52:02,692 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-26 06:52:15,692 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34206.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:52:37,391 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.199e+02 1.703e+02 1.979e+02 2.376e+02 4.570e+02, threshold=3.959e+02, percent-clipped=2.0 +2023-03-26 06:52:54,046 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.56 vs. limit=2.0 +2023-03-26 06:52:55,412 INFO [finetune.py:976] (4/7) Epoch 6, batch 5600, loss[loss=0.2158, simple_loss=0.2919, pruned_loss=0.06987, over 4800.00 frames. ], tot_loss[loss=0.2137, simple_loss=0.2746, pruned_loss=0.07637, over 955217.12 frames. ], batch size: 29, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:53:07,668 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.0490, 3.6725, 3.8240, 3.7537, 3.5988, 3.4999, 4.2472, 1.5040], + device='cuda:4'), covar=tensor([0.1503, 0.1638, 0.1492, 0.2164, 0.2479, 0.2438, 0.1328, 0.7141], + device='cuda:4'), in_proj_covar=tensor([0.0351, 0.0241, 0.0273, 0.0293, 0.0333, 0.0284, 0.0302, 0.0295], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 06:53:26,849 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-26 06:53:54,498 INFO [finetune.py:976] (4/7) Epoch 6, batch 5650, loss[loss=0.2758, simple_loss=0.3323, pruned_loss=0.1096, over 4811.00 frames. ], tot_loss[loss=0.2159, simple_loss=0.2774, pruned_loss=0.07717, over 956006.77 frames. ], batch size: 45, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:53:58,433 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34295.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:54:35,325 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.805e+01 1.675e+02 2.029e+02 2.481e+02 4.265e+02, threshold=4.057e+02, percent-clipped=3.0 +2023-03-26 06:54:40,154 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34333.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:54:48,434 INFO [finetune.py:976] (4/7) Epoch 6, batch 5700, loss[loss=0.2295, simple_loss=0.283, pruned_loss=0.08804, over 4080.00 frames. ], tot_loss[loss=0.2125, simple_loss=0.2726, pruned_loss=0.07625, over 935740.41 frames. ], batch size: 18, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:55:38,389 INFO [finetune.py:976] (4/7) Epoch 7, batch 0, loss[loss=0.244, simple_loss=0.2994, pruned_loss=0.09434, over 4894.00 frames. ], tot_loss[loss=0.244, simple_loss=0.2994, pruned_loss=0.09434, over 4894.00 frames. ], batch size: 36, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:55:38,389 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 06:55:55,928 INFO [finetune.py:1010] (4/7) Epoch 7, validation: loss=0.165, simple_loss=0.2365, pruned_loss=0.04677, over 2265189.00 frames. +2023-03-26 06:55:55,928 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6301MB +2023-03-26 06:56:14,683 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34381.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:56:36,913 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34400.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:56:59,197 INFO [finetune.py:976] (4/7) Epoch 7, batch 50, loss[loss=0.1519, simple_loss=0.2239, pruned_loss=0.0399, over 4734.00 frames. ], tot_loss[loss=0.2193, simple_loss=0.2797, pruned_loss=0.07944, over 217238.81 frames. ], batch size: 27, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:56:59,315 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1213, 1.3318, 1.1819, 1.4049, 1.3742, 2.3934, 1.1870, 1.4360], + device='cuda:4'), covar=tensor([0.0995, 0.1590, 0.1178, 0.0926, 0.1512, 0.0395, 0.1405, 0.1550], + device='cuda:4'), in_proj_covar=tensor([0.0077, 0.0081, 0.0076, 0.0079, 0.0092, 0.0083, 0.0084, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 06:57:09,545 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.600e+02 2.022e+02 2.565e+02 5.766e+02, threshold=4.045e+02, percent-clipped=4.0 +2023-03-26 06:57:19,391 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 06:57:30,163 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-26 06:58:05,579 INFO [finetune.py:976] (4/7) Epoch 7, batch 100, loss[loss=0.1912, simple_loss=0.2585, pruned_loss=0.06198, over 4789.00 frames. ], tot_loss[loss=0.2099, simple_loss=0.2706, pruned_loss=0.07461, over 382057.20 frames. ], batch size: 51, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:58:45,260 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34501.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 06:59:06,468 INFO [finetune.py:976] (4/7) Epoch 7, batch 150, loss[loss=0.2461, simple_loss=0.2988, pruned_loss=0.09672, over 4718.00 frames. ], tot_loss[loss=0.2041, simple_loss=0.264, pruned_loss=0.07212, over 509083.35 frames. ], batch size: 59, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 06:59:16,116 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.65 vs. limit=2.0 +2023-03-26 06:59:17,661 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.007e+02 1.586e+02 1.902e+02 2.328e+02 6.438e+02, threshold=3.804e+02, percent-clipped=3.0 +2023-03-26 06:59:18,969 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.7009, 4.0210, 4.2567, 4.4782, 4.4130, 4.1635, 4.7905, 1.5610], + device='cuda:4'), covar=tensor([0.0677, 0.0937, 0.0676, 0.0900, 0.1140, 0.1263, 0.0570, 0.5552], + device='cuda:4'), in_proj_covar=tensor([0.0353, 0.0242, 0.0273, 0.0294, 0.0335, 0.0285, 0.0303, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 07:00:10,777 INFO [finetune.py:976] (4/7) Epoch 7, batch 200, loss[loss=0.2018, simple_loss=0.2601, pruned_loss=0.07168, over 4901.00 frames. ], tot_loss[loss=0.2062, simple_loss=0.2653, pruned_loss=0.07352, over 610661.29 frames. ], batch size: 43, lr: 3.89e-03, grad_scale: 32.0 +2023-03-26 07:00:19,831 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6460, 3.4218, 3.2632, 1.4633, 3.5714, 2.5770, 0.7829, 2.2876], + device='cuda:4'), covar=tensor([0.2326, 0.2090, 0.1712, 0.3621, 0.1161, 0.1138, 0.4495, 0.1584], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0171, 0.0162, 0.0127, 0.0155, 0.0122, 0.0145, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 07:00:42,229 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6843, 1.2308, 0.8485, 1.5624, 2.0142, 1.2983, 1.5534, 1.6496], + device='cuda:4'), covar=tensor([0.1594, 0.2166, 0.2222, 0.1297, 0.2114, 0.1974, 0.1407, 0.2009], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0098, 0.0115, 0.0093, 0.0124, 0.0095, 0.0100, 0.0093], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 07:00:42,259 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34593.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:00:43,463 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34595.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:01:13,423 INFO [finetune.py:976] (4/7) Epoch 7, batch 250, loss[loss=0.1884, simple_loss=0.2458, pruned_loss=0.06547, over 4769.00 frames. ], tot_loss[loss=0.2124, simple_loss=0.2716, pruned_loss=0.07662, over 686638.55 frames. ], batch size: 26, lr: 3.88e-03, grad_scale: 32.0 +2023-03-26 07:01:22,751 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34622.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:01:24,456 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.284e+02 1.737e+02 2.023e+02 2.550e+02 3.958e+02, threshold=4.047e+02, percent-clipped=1.0 +2023-03-26 07:01:25,175 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34626.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:01:25,845 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7899, 1.4947, 1.3679, 1.0189, 1.4623, 1.4944, 1.4294, 2.1315], + device='cuda:4'), covar=tensor([0.5736, 0.5083, 0.4264, 0.5342, 0.4931, 0.3146, 0.4921, 0.2185], + device='cuda:4'), in_proj_covar=tensor([0.0283, 0.0258, 0.0220, 0.0282, 0.0240, 0.0205, 0.0244, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 07:01:33,508 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1240, 1.8502, 1.9897, 0.8521, 2.1888, 2.4357, 1.9641, 1.8979], + device='cuda:4'), covar=tensor([0.0918, 0.0783, 0.0469, 0.0805, 0.0511, 0.0669, 0.0438, 0.0730], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0158, 0.0121, 0.0138, 0.0132, 0.0125, 0.0147, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.7059e-05, 1.1719e-04, 8.7643e-05, 1.0097e-04, 9.4957e-05, 9.2065e-05, + 1.0884e-04, 1.0781e-04], device='cuda:4') +2023-03-26 07:01:42,608 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.9990, 3.4190, 3.6308, 3.8486, 3.7503, 3.5021, 4.0606, 1.3422], + device='cuda:4'), covar=tensor([0.0744, 0.0882, 0.0737, 0.0858, 0.1100, 0.1293, 0.0697, 0.4881], + device='cuda:4'), in_proj_covar=tensor([0.0354, 0.0244, 0.0275, 0.0295, 0.0337, 0.0286, 0.0305, 0.0298], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 07:01:45,028 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34643.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:01:56,759 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3109, 1.4389, 1.6947, 1.7533, 1.5642, 3.2374, 1.2687, 1.5492], + device='cuda:4'), covar=tensor([0.1085, 0.1728, 0.1325, 0.1045, 0.1577, 0.0262, 0.1424, 0.1700], + device='cuda:4'), in_proj_covar=tensor([0.0078, 0.0081, 0.0077, 0.0080, 0.0092, 0.0083, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 07:01:57,408 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34654.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:02:14,153 INFO [finetune.py:976] (4/7) Epoch 7, batch 300, loss[loss=0.192, simple_loss=0.2648, pruned_loss=0.05961, over 4897.00 frames. ], tot_loss[loss=0.2139, simple_loss=0.2747, pruned_loss=0.07651, over 749559.28 frames. ], batch size: 32, lr: 3.88e-03, grad_scale: 32.0 +2023-03-26 07:02:34,477 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34683.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 07:02:36,960 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34687.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:02:55,170 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34700.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:03:04,178 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0336, 2.0234, 1.9445, 1.4468, 2.1872, 2.1071, 2.1655, 1.8125], + device='cuda:4'), covar=tensor([0.0702, 0.0653, 0.0856, 0.0953, 0.0489, 0.0856, 0.0661, 0.1043], + device='cuda:4'), in_proj_covar=tensor([0.0140, 0.0135, 0.0145, 0.0129, 0.0114, 0.0146, 0.0147, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 07:03:06,627 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34710.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:03:15,689 INFO [finetune.py:976] (4/7) Epoch 7, batch 350, loss[loss=0.2416, simple_loss=0.3119, pruned_loss=0.0856, over 4820.00 frames. ], tot_loss[loss=0.217, simple_loss=0.2781, pruned_loss=0.07792, over 795129.94 frames. ], batch size: 33, lr: 3.88e-03, grad_scale: 32.0 +2023-03-26 07:03:27,161 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.146e+02 1.664e+02 2.043e+02 2.496e+02 5.690e+02, threshold=4.087e+02, percent-clipped=3.0 +2023-03-26 07:03:54,746 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34748.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:04:07,396 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.5625, 1.6562, 1.6474, 0.9467, 1.7224, 1.8675, 1.8928, 1.4984], + device='cuda:4'), covar=tensor([0.1028, 0.0600, 0.0481, 0.0610, 0.0409, 0.0513, 0.0305, 0.0709], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0157, 0.0120, 0.0138, 0.0131, 0.0124, 0.0145, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.6603e-05, 1.1634e-04, 8.6745e-05, 1.0027e-04, 9.3842e-05, 9.1271e-05, + 1.0748e-04, 1.0688e-04], device='cuda:4') +2023-03-26 07:04:08,697 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 07:04:16,440 INFO [finetune.py:976] (4/7) Epoch 7, batch 400, loss[loss=0.2302, simple_loss=0.2893, pruned_loss=0.08557, over 4785.00 frames. ], tot_loss[loss=0.2151, simple_loss=0.2772, pruned_loss=0.07651, over 830567.29 frames. ], batch size: 26, lr: 3.88e-03, grad_scale: 32.0 +2023-03-26 07:04:24,030 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34771.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:04:55,075 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34801.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:05:14,337 INFO [finetune.py:976] (4/7) Epoch 7, batch 450, loss[loss=0.1823, simple_loss=0.2506, pruned_loss=0.05703, over 4896.00 frames. ], tot_loss[loss=0.2145, simple_loss=0.276, pruned_loss=0.0765, over 858049.74 frames. ], batch size: 37, lr: 3.88e-03, grad_scale: 32.0 +2023-03-26 07:05:14,465 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1899, 1.2791, 1.3048, 0.6651, 1.1580, 1.4773, 1.5368, 1.2632], + device='cuda:4'), covar=tensor([0.0818, 0.0550, 0.0435, 0.0508, 0.0461, 0.0463, 0.0296, 0.0646], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0157, 0.0120, 0.0137, 0.0131, 0.0124, 0.0145, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.6501e-05, 1.1611e-04, 8.6641e-05, 1.0014e-04, 9.3986e-05, 9.1348e-05, + 1.0741e-04, 1.0700e-04], device='cuda:4') +2023-03-26 07:05:25,983 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.133e+01 1.634e+02 1.827e+02 2.211e+02 3.915e+02, threshold=3.654e+02, percent-clipped=0.0 +2023-03-26 07:05:50,920 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=34849.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:06:11,592 INFO [finetune.py:976] (4/7) Epoch 7, batch 500, loss[loss=0.2348, simple_loss=0.286, pruned_loss=0.09177, over 4899.00 frames. ], tot_loss[loss=0.2104, simple_loss=0.2719, pruned_loss=0.07441, over 880993.87 frames. ], batch size: 35, lr: 3.88e-03, grad_scale: 32.0 +2023-03-26 07:07:15,208 INFO [finetune.py:976] (4/7) Epoch 7, batch 550, loss[loss=0.1538, simple_loss=0.2237, pruned_loss=0.0419, over 4742.00 frames. ], tot_loss[loss=0.2067, simple_loss=0.2678, pruned_loss=0.07274, over 898779.24 frames. ], batch size: 27, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:07:26,484 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.074e+02 1.634e+02 2.013e+02 2.383e+02 4.182e+02, threshold=4.026e+02, percent-clipped=3.0 +2023-03-26 07:07:53,594 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34949.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:08:14,352 INFO [finetune.py:976] (4/7) Epoch 7, batch 600, loss[loss=0.2469, simple_loss=0.305, pruned_loss=0.09443, over 4931.00 frames. ], tot_loss[loss=0.2074, simple_loss=0.2682, pruned_loss=0.07326, over 910535.45 frames. ], batch size: 33, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:08:31,420 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34978.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 07:08:34,223 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4967, 1.3543, 1.3822, 1.4571, 1.0421, 3.1818, 1.2617, 1.8019], + device='cuda:4'), covar=tensor([0.3204, 0.2422, 0.2081, 0.2308, 0.1985, 0.0202, 0.2841, 0.1318], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0114, 0.0118, 0.0122, 0.0117, 0.0098, 0.0101, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 07:08:35,300 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34982.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:09:16,192 INFO [finetune.py:976] (4/7) Epoch 7, batch 650, loss[loss=0.1856, simple_loss=0.2496, pruned_loss=0.06085, over 4866.00 frames. ], tot_loss[loss=0.2099, simple_loss=0.2716, pruned_loss=0.07413, over 921538.04 frames. ], batch size: 34, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:09:27,412 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.303e+02 1.723e+02 2.031e+02 2.472e+02 3.902e+02, threshold=4.061e+02, percent-clipped=0.0 +2023-03-26 07:10:17,352 INFO [finetune.py:976] (4/7) Epoch 7, batch 700, loss[loss=0.2081, simple_loss=0.2688, pruned_loss=0.07367, over 4907.00 frames. ], tot_loss[loss=0.2113, simple_loss=0.2733, pruned_loss=0.07458, over 925907.66 frames. ], batch size: 32, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:10:17,420 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35066.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 07:10:46,476 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35092.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:10:56,431 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0280, 1.9061, 1.5970, 1.8013, 1.7781, 1.7057, 1.7949, 2.5192], + device='cuda:4'), covar=tensor([0.5579, 0.6176, 0.4776, 0.5763, 0.5805, 0.3443, 0.6017, 0.2196], + device='cuda:4'), in_proj_covar=tensor([0.0283, 0.0258, 0.0221, 0.0282, 0.0240, 0.0205, 0.0245, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 07:11:11,742 INFO [finetune.py:976] (4/7) Epoch 7, batch 750, loss[loss=0.2566, simple_loss=0.3109, pruned_loss=0.1012, over 4229.00 frames. ], tot_loss[loss=0.2145, simple_loss=0.2766, pruned_loss=0.07622, over 932713.71 frames. ], batch size: 66, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:11:23,577 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.641e+02 2.027e+02 2.429e+02 3.682e+02, threshold=4.054e+02, percent-clipped=0.0 +2023-03-26 07:12:01,717 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6264, 1.4665, 2.2179, 3.5737, 2.4436, 2.4861, 0.8602, 2.7057], + device='cuda:4'), covar=tensor([0.1869, 0.1533, 0.1417, 0.0540, 0.0762, 0.1383, 0.2089, 0.0612], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0117, 0.0135, 0.0165, 0.0102, 0.0141, 0.0129, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 07:12:01,755 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35153.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:12:14,609 INFO [finetune.py:976] (4/7) Epoch 7, batch 800, loss[loss=0.1913, simple_loss=0.2651, pruned_loss=0.05875, over 4898.00 frames. ], tot_loss[loss=0.2129, simple_loss=0.2753, pruned_loss=0.07527, over 938290.32 frames. ], batch size: 35, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:13:05,891 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-03-26 07:13:17,536 INFO [finetune.py:976] (4/7) Epoch 7, batch 850, loss[loss=0.1822, simple_loss=0.2465, pruned_loss=0.05889, over 4922.00 frames. ], tot_loss[loss=0.2105, simple_loss=0.2723, pruned_loss=0.07436, over 943454.50 frames. ], batch size: 38, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:13:27,729 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.064e+02 1.645e+02 1.948e+02 2.227e+02 3.525e+02, threshold=3.897e+02, percent-clipped=0.0 +2023-03-26 07:13:32,073 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 07:13:55,211 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35249.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:14:14,656 INFO [finetune.py:976] (4/7) Epoch 7, batch 900, loss[loss=0.1669, simple_loss=0.2366, pruned_loss=0.04866, over 4894.00 frames. ], tot_loss[loss=0.2087, simple_loss=0.2701, pruned_loss=0.07366, over 947382.87 frames. ], batch size: 43, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:14:32,801 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35278.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:14:33,509 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8616, 0.9634, 1.7475, 1.6282, 1.5236, 1.5125, 1.5080, 1.5604], + device='cuda:4'), covar=tensor([0.4386, 0.5922, 0.4944, 0.5224, 0.6213, 0.4905, 0.6669, 0.4930], + device='cuda:4'), in_proj_covar=tensor([0.0229, 0.0243, 0.0254, 0.0255, 0.0243, 0.0219, 0.0272, 0.0225], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 07:14:35,296 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35282.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:14:55,291 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=35297.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:15:14,826 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6000, 3.3547, 3.1934, 1.4906, 3.4663, 2.5327, 0.9146, 2.2260], + device='cuda:4'), covar=tensor([0.2545, 0.2196, 0.1703, 0.3522, 0.1211, 0.1152, 0.4297, 0.1650], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0172, 0.0162, 0.0127, 0.0155, 0.0122, 0.0145, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 07:15:16,598 INFO [finetune.py:976] (4/7) Epoch 7, batch 950, loss[loss=0.1664, simple_loss=0.2382, pruned_loss=0.04733, over 4812.00 frames. ], tot_loss[loss=0.2077, simple_loss=0.2688, pruned_loss=0.07326, over 950945.29 frames. ], batch size: 25, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:15:31,371 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.193e+02 1.515e+02 1.811e+02 2.306e+02 3.628e+02, threshold=3.621e+02, percent-clipped=0.0 +2023-03-26 07:15:31,449 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=35326.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:15:33,895 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=35330.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:15:38,887 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7187, 2.3698, 1.8885, 1.0511, 2.0180, 2.1041, 1.8721, 2.0508], + device='cuda:4'), covar=tensor([0.0825, 0.0957, 0.1791, 0.2131, 0.1582, 0.2165, 0.2303, 0.1103], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0201, 0.0200, 0.0188, 0.0217, 0.0207, 0.0219, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 07:15:50,923 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35345.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:16:01,721 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.70 vs. limit=5.0 +2023-03-26 07:16:18,536 INFO [finetune.py:976] (4/7) Epoch 7, batch 1000, loss[loss=0.195, simple_loss=0.2571, pruned_loss=0.06643, over 4745.00 frames. ], tot_loss[loss=0.2092, simple_loss=0.2708, pruned_loss=0.07382, over 951552.25 frames. ], batch size: 26, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:16:18,647 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35366.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 07:16:27,465 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35374.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:17:08,095 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35406.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 07:17:17,935 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=35414.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 07:17:19,086 INFO [finetune.py:976] (4/7) Epoch 7, batch 1050, loss[loss=0.197, simple_loss=0.2611, pruned_loss=0.06645, over 4912.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.2752, pruned_loss=0.07515, over 953860.23 frames. ], batch size: 36, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:17:30,144 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.274e+02 1.696e+02 1.924e+02 2.371e+02 5.787e+02, threshold=3.848e+02, percent-clipped=4.0 +2023-03-26 07:17:41,407 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35435.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:17:59,815 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35448.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:18:20,381 INFO [finetune.py:976] (4/7) Epoch 7, batch 1100, loss[loss=0.2141, simple_loss=0.2881, pruned_loss=0.07001, over 4893.00 frames. ], tot_loss[loss=0.2147, simple_loss=0.2767, pruned_loss=0.07638, over 955336.94 frames. ], batch size: 43, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:18:33,604 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6404, 1.4909, 1.4960, 1.5934, 1.1569, 2.9953, 1.1245, 1.7241], + device='cuda:4'), covar=tensor([0.3539, 0.2529, 0.2145, 0.2422, 0.1897, 0.0269, 0.2750, 0.1335], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0114, 0.0118, 0.0122, 0.0116, 0.0098, 0.0101, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 07:18:50,231 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.21 vs. limit=5.0 +2023-03-26 07:19:01,451 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8930, 1.7357, 1.5523, 1.9413, 2.1497, 1.8924, 1.3471, 1.4732], + device='cuda:4'), covar=tensor([0.2312, 0.2209, 0.2037, 0.1659, 0.2043, 0.1203, 0.2793, 0.2024], + device='cuda:4'), in_proj_covar=tensor([0.0236, 0.0208, 0.0204, 0.0186, 0.0237, 0.0176, 0.0212, 0.0191], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 07:19:22,492 INFO [finetune.py:976] (4/7) Epoch 7, batch 1150, loss[loss=0.1951, simple_loss=0.2638, pruned_loss=0.06326, over 4888.00 frames. ], tot_loss[loss=0.2168, simple_loss=0.2787, pruned_loss=0.07744, over 957257.33 frames. ], batch size: 43, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:19:33,764 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.207e+02 1.764e+02 2.068e+02 2.432e+02 4.937e+02, threshold=4.137e+02, percent-clipped=2.0 +2023-03-26 07:19:41,298 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35530.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:20:24,974 INFO [finetune.py:976] (4/7) Epoch 7, batch 1200, loss[loss=0.1906, simple_loss=0.267, pruned_loss=0.05707, over 4740.00 frames. ], tot_loss[loss=0.2156, simple_loss=0.2769, pruned_loss=0.0772, over 955943.90 frames. ], batch size: 59, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:20:51,613 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35591.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:21:20,130 INFO [finetune.py:976] (4/7) Epoch 7, batch 1250, loss[loss=0.2158, simple_loss=0.2784, pruned_loss=0.07657, over 4816.00 frames. ], tot_loss[loss=0.2127, simple_loss=0.2734, pruned_loss=0.07601, over 957605.86 frames. ], batch size: 41, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:21:31,425 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.265e+02 1.688e+02 2.018e+02 2.672e+02 1.298e+03, threshold=4.035e+02, percent-clipped=4.0 +2023-03-26 07:21:38,445 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.14 vs. limit=5.0 +2023-03-26 07:22:18,864 INFO [finetune.py:976] (4/7) Epoch 7, batch 1300, loss[loss=0.2243, simple_loss=0.2923, pruned_loss=0.07813, over 4820.00 frames. ], tot_loss[loss=0.2092, simple_loss=0.2697, pruned_loss=0.07436, over 956488.48 frames. ], batch size: 38, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:23:06,116 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35701.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 07:23:06,771 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9333, 1.6930, 1.6609, 1.8992, 1.3898, 4.4905, 1.6664, 2.4427], + device='cuda:4'), covar=tensor([0.3310, 0.2483, 0.2156, 0.2341, 0.1864, 0.0104, 0.2664, 0.1230], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0115, 0.0119, 0.0123, 0.0117, 0.0099, 0.0102, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 07:23:25,711 INFO [finetune.py:976] (4/7) Epoch 7, batch 1350, loss[loss=0.1909, simple_loss=0.2481, pruned_loss=0.06682, over 4826.00 frames. ], tot_loss[loss=0.2086, simple_loss=0.2692, pruned_loss=0.07395, over 957785.76 frames. ], batch size: 30, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:23:38,291 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.109e+01 1.659e+02 1.871e+02 2.249e+02 4.421e+02, threshold=3.743e+02, percent-clipped=1.0 +2023-03-26 07:23:40,802 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35730.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:24:06,323 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35748.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:24:25,447 INFO [finetune.py:976] (4/7) Epoch 7, batch 1400, loss[loss=0.2142, simple_loss=0.2904, pruned_loss=0.06897, over 4840.00 frames. ], tot_loss[loss=0.2123, simple_loss=0.2738, pruned_loss=0.07536, over 956466.24 frames. ], batch size: 40, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:24:33,487 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35771.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:24:42,360 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-03-26 07:25:01,856 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=35796.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:25:20,904 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35811.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:25:23,858 INFO [finetune.py:976] (4/7) Epoch 7, batch 1450, loss[loss=0.1642, simple_loss=0.229, pruned_loss=0.04972, over 4776.00 frames. ], tot_loss[loss=0.213, simple_loss=0.275, pruned_loss=0.07554, over 957135.25 frames. ], batch size: 26, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:25:33,653 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.204e+02 1.737e+02 2.011e+02 2.560e+02 4.083e+02, threshold=4.021e+02, percent-clipped=3.0 +2023-03-26 07:25:43,688 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35832.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:26:24,785 INFO [finetune.py:976] (4/7) Epoch 7, batch 1500, loss[loss=0.2087, simple_loss=0.2709, pruned_loss=0.07322, over 4917.00 frames. ], tot_loss[loss=0.2142, simple_loss=0.2759, pruned_loss=0.07622, over 958570.92 frames. ], batch size: 33, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:26:32,658 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35872.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:26:47,875 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35886.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:27:11,196 INFO [finetune.py:976] (4/7) Epoch 7, batch 1550, loss[loss=0.1824, simple_loss=0.2519, pruned_loss=0.05646, over 4888.00 frames. ], tot_loss[loss=0.2139, simple_loss=0.2758, pruned_loss=0.07605, over 958167.27 frames. ], batch size: 35, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:27:17,688 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.940e+01 1.555e+02 1.902e+02 2.352e+02 4.828e+02, threshold=3.804e+02, percent-clipped=1.0 +2023-03-26 07:27:45,066 INFO [finetune.py:976] (4/7) Epoch 7, batch 1600, loss[loss=0.192, simple_loss=0.2559, pruned_loss=0.06399, over 4844.00 frames. ], tot_loss[loss=0.2116, simple_loss=0.2728, pruned_loss=0.07527, over 957272.61 frames. ], batch size: 47, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:28:16,184 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-26 07:28:25,891 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36001.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:28:35,457 INFO [finetune.py:976] (4/7) Epoch 7, batch 1650, loss[loss=0.1747, simple_loss=0.2452, pruned_loss=0.05214, over 4775.00 frames. ], tot_loss[loss=0.2083, simple_loss=0.2692, pruned_loss=0.07364, over 957949.24 frames. ], batch size: 28, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:28:41,918 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.242e+02 1.569e+02 1.867e+02 2.342e+02 3.778e+02, threshold=3.734e+02, percent-clipped=0.0 +2023-03-26 07:28:50,686 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36030.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:29:09,240 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36049.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 07:29:11,117 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.7718, 1.8595, 1.7989, 1.0813, 1.9240, 2.1143, 2.0819, 1.5700], + device='cuda:4'), covar=tensor([0.1000, 0.0613, 0.0488, 0.0645, 0.0375, 0.0534, 0.0321, 0.0738], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0157, 0.0120, 0.0137, 0.0132, 0.0124, 0.0145, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.6855e-05, 1.1560e-04, 8.7115e-05, 9.9725e-05, 9.4577e-05, 9.1317e-05, + 1.0736e-04, 1.0718e-04], device='cuda:4') +2023-03-26 07:29:25,210 INFO [finetune.py:976] (4/7) Epoch 7, batch 1700, loss[loss=0.1854, simple_loss=0.2353, pruned_loss=0.06776, over 4812.00 frames. ], tot_loss[loss=0.208, simple_loss=0.2683, pruned_loss=0.07389, over 959452.54 frames. ], batch size: 25, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:29:39,787 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36078.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:30:15,625 INFO [finetune.py:976] (4/7) Epoch 7, batch 1750, loss[loss=0.178, simple_loss=0.258, pruned_loss=0.04902, over 4761.00 frames. ], tot_loss[loss=0.2103, simple_loss=0.2709, pruned_loss=0.07486, over 959545.79 frames. ], batch size: 26, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:30:27,793 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.647e+02 1.960e+02 2.452e+02 4.962e+02, threshold=3.920e+02, percent-clipped=3.0 +2023-03-26 07:30:28,497 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36127.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:31:18,547 INFO [finetune.py:976] (4/7) Epoch 7, batch 1800, loss[loss=0.2248, simple_loss=0.2917, pruned_loss=0.07891, over 4858.00 frames. ], tot_loss[loss=0.2136, simple_loss=0.2748, pruned_loss=0.07618, over 956992.48 frames. ], batch size: 34, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:31:19,218 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36167.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:31:19,489 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-26 07:31:38,159 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36181.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:31:47,169 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36186.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:32:21,147 INFO [finetune.py:976] (4/7) Epoch 7, batch 1850, loss[loss=0.1844, simple_loss=0.2619, pruned_loss=0.05346, over 4761.00 frames. ], tot_loss[loss=0.215, simple_loss=0.276, pruned_loss=0.07695, over 957649.35 frames. ], batch size: 28, lr: 3.88e-03, grad_scale: 16.0 +2023-03-26 07:32:21,268 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5317, 2.3525, 2.8519, 1.8475, 2.7895, 2.9155, 2.1256, 2.9674], + device='cuda:4'), covar=tensor([0.1446, 0.1945, 0.1429, 0.2401, 0.0857, 0.1664, 0.2555, 0.0841], + device='cuda:4'), in_proj_covar=tensor([0.0201, 0.0203, 0.0196, 0.0195, 0.0180, 0.0219, 0.0216, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 07:32:23,093 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 07:32:33,456 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.247e+02 1.736e+02 2.131e+02 2.651e+02 6.216e+02, threshold=4.263e+02, percent-clipped=3.0 +2023-03-26 07:32:40,446 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36234.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:32:50,346 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36242.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:33:21,419 INFO [finetune.py:976] (4/7) Epoch 7, batch 1900, loss[loss=0.1818, simple_loss=0.2489, pruned_loss=0.05739, over 4781.00 frames. ], tot_loss[loss=0.215, simple_loss=0.2768, pruned_loss=0.0766, over 958039.39 frames. ], batch size: 25, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:34:25,475 INFO [finetune.py:976] (4/7) Epoch 7, batch 1950, loss[loss=0.1886, simple_loss=0.256, pruned_loss=0.06065, over 4928.00 frames. ], tot_loss[loss=0.2137, simple_loss=0.2754, pruned_loss=0.07606, over 958483.38 frames. ], batch size: 33, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:34:36,074 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 07:34:36,923 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.969e+01 1.685e+02 2.051e+02 2.475e+02 4.640e+02, threshold=4.103e+02, percent-clipped=3.0 +2023-03-26 07:34:46,845 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1482, 3.5691, 3.7107, 4.0011, 3.8756, 3.6305, 4.2303, 1.2530], + device='cuda:4'), covar=tensor([0.0773, 0.0828, 0.0876, 0.0865, 0.1239, 0.1567, 0.0710, 0.5188], + device='cuda:4'), in_proj_covar=tensor([0.0345, 0.0240, 0.0270, 0.0288, 0.0329, 0.0280, 0.0299, 0.0292], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 07:35:28,763 INFO [finetune.py:976] (4/7) Epoch 7, batch 2000, loss[loss=0.1811, simple_loss=0.2468, pruned_loss=0.05766, over 4394.00 frames. ], tot_loss[loss=0.211, simple_loss=0.2722, pruned_loss=0.07487, over 958467.75 frames. ], batch size: 19, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:36:30,440 INFO [finetune.py:976] (4/7) Epoch 7, batch 2050, loss[loss=0.1988, simple_loss=0.2531, pruned_loss=0.07232, over 4927.00 frames. ], tot_loss[loss=0.2077, simple_loss=0.2685, pruned_loss=0.07346, over 959865.38 frames. ], batch size: 46, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:36:43,671 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 7.910e+01 1.532e+02 1.893e+02 2.218e+02 7.941e+02, threshold=3.786e+02, percent-clipped=2.0 +2023-03-26 07:36:44,386 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36427.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:37:13,976 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-26 07:37:34,744 INFO [finetune.py:976] (4/7) Epoch 7, batch 2100, loss[loss=0.1514, simple_loss=0.2283, pruned_loss=0.03721, over 4895.00 frames. ], tot_loss[loss=0.2069, simple_loss=0.2677, pruned_loss=0.07307, over 959555.34 frames. ], batch size: 32, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:37:35,477 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36467.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:37:45,798 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36475.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:37:47,787 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.86 vs. limit=5.0 +2023-03-26 07:38:36,450 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36515.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:38:36,990 INFO [finetune.py:976] (4/7) Epoch 7, batch 2150, loss[loss=0.2475, simple_loss=0.3029, pruned_loss=0.09601, over 4918.00 frames. ], tot_loss[loss=0.2096, simple_loss=0.2707, pruned_loss=0.07428, over 959943.32 frames. ], batch size: 42, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:38:48,002 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.129e+02 1.787e+02 2.211e+02 2.590e+02 5.595e+02, threshold=4.423e+02, percent-clipped=4.0 +2023-03-26 07:39:05,369 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36537.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:39:35,928 INFO [finetune.py:976] (4/7) Epoch 7, batch 2200, loss[loss=0.2107, simple_loss=0.2646, pruned_loss=0.07841, over 4797.00 frames. ], tot_loss[loss=0.2116, simple_loss=0.2731, pruned_loss=0.07499, over 958656.58 frames. ], batch size: 25, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:40:25,412 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36605.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:40:27,769 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36608.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:40:35,327 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36612.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:40:38,178 INFO [finetune.py:976] (4/7) Epoch 7, batch 2250, loss[loss=0.2168, simple_loss=0.2676, pruned_loss=0.08303, over 4925.00 frames. ], tot_loss[loss=0.2134, simple_loss=0.275, pruned_loss=0.07595, over 958966.00 frames. ], batch size: 28, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:40:49,910 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.198e+02 1.735e+02 1.950e+02 2.446e+02 5.153e+02, threshold=3.899e+02, percent-clipped=1.0 +2023-03-26 07:40:50,018 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0576, 1.6662, 2.6430, 3.8286, 2.6265, 2.6469, 1.0390, 3.0612], + device='cuda:4'), covar=tensor([0.1713, 0.1653, 0.1220, 0.0586, 0.0794, 0.1717, 0.1878, 0.0528], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0117, 0.0134, 0.0165, 0.0101, 0.0140, 0.0128, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 07:41:41,216 INFO [finetune.py:976] (4/7) Epoch 7, batch 2300, loss[loss=0.1972, simple_loss=0.2601, pruned_loss=0.06713, over 4880.00 frames. ], tot_loss[loss=0.2141, simple_loss=0.2759, pruned_loss=0.07616, over 959549.07 frames. ], batch size: 32, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:41:41,338 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36666.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:41:43,641 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36669.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:41:51,484 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36673.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:42:18,126 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36700.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:42:33,924 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 07:42:36,646 INFO [finetune.py:976] (4/7) Epoch 7, batch 2350, loss[loss=0.1725, simple_loss=0.2324, pruned_loss=0.0563, over 4756.00 frames. ], tot_loss[loss=0.2122, simple_loss=0.2734, pruned_loss=0.07547, over 956564.04 frames. ], batch size: 27, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:42:36,749 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4226, 1.5085, 0.8577, 2.2743, 2.5699, 1.8675, 1.9613, 2.1172], + device='cuda:4'), covar=tensor([0.1395, 0.2142, 0.2265, 0.1145, 0.1914, 0.1899, 0.1463, 0.1954], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0098, 0.0114, 0.0093, 0.0124, 0.0096, 0.0100, 0.0093], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 07:42:47,524 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0109, 1.8018, 1.6326, 1.7285, 1.7288, 1.7336, 1.7242, 2.5561], + device='cuda:4'), covar=tensor([0.5434, 0.6727, 0.4657, 0.5999, 0.6027, 0.3409, 0.6057, 0.2104], + device='cuda:4'), in_proj_covar=tensor([0.0284, 0.0259, 0.0220, 0.0282, 0.0240, 0.0205, 0.0244, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 07:42:49,194 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.103e+02 1.509e+02 1.866e+02 2.321e+02 4.735e+02, threshold=3.732e+02, percent-clipped=2.0 +2023-03-26 07:43:25,426 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 07:43:29,351 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36761.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:43:37,712 INFO [finetune.py:976] (4/7) Epoch 7, batch 2400, loss[loss=0.192, simple_loss=0.2508, pruned_loss=0.06657, over 4868.00 frames. ], tot_loss[loss=0.2096, simple_loss=0.2704, pruned_loss=0.07444, over 956199.13 frames. ], batch size: 31, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:44:17,810 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-26 07:44:28,537 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36805.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:44:29,818 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-03-26 07:44:40,197 INFO [finetune.py:976] (4/7) Epoch 7, batch 2450, loss[loss=0.1828, simple_loss=0.2497, pruned_loss=0.05793, over 4759.00 frames. ], tot_loss[loss=0.2081, simple_loss=0.2681, pruned_loss=0.07405, over 953689.39 frames. ], batch size: 27, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:44:49,389 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3098, 2.1344, 2.7709, 1.7504, 2.5134, 2.6313, 2.0380, 2.6649], + device='cuda:4'), covar=tensor([0.1698, 0.2260, 0.1702, 0.2585, 0.1180, 0.1757, 0.2697, 0.1241], + device='cuda:4'), in_proj_covar=tensor([0.0200, 0.0202, 0.0194, 0.0194, 0.0180, 0.0218, 0.0214, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 07:44:51,699 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.798e+02 2.140e+02 2.594e+02 4.660e+02, threshold=4.281e+02, percent-clipped=3.0 +2023-03-26 07:45:00,766 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.55 vs. limit=5.0 +2023-03-26 07:45:10,244 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36837.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:45:31,450 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1210, 1.6896, 2.0281, 0.7388, 2.1774, 2.5857, 1.9211, 1.8516], + device='cuda:4'), covar=tensor([0.1294, 0.1459, 0.0652, 0.1071, 0.0804, 0.0579, 0.0841, 0.0974], + device='cuda:4'), in_proj_covar=tensor([0.0128, 0.0155, 0.0119, 0.0136, 0.0131, 0.0123, 0.0144, 0.0144], + device='cuda:4'), out_proj_covar=tensor([9.5829e-05, 1.1409e-04, 8.5795e-05, 9.9037e-05, 9.3850e-05, 9.0567e-05, + 1.0641e-04, 1.0592e-04], device='cuda:4') +2023-03-26 07:45:49,128 INFO [finetune.py:976] (4/7) Epoch 7, batch 2500, loss[loss=0.1999, simple_loss=0.2665, pruned_loss=0.06661, over 4822.00 frames. ], tot_loss[loss=0.2098, simple_loss=0.27, pruned_loss=0.07484, over 954871.06 frames. ], batch size: 39, lr: 3.87e-03, grad_scale: 16.0 +2023-03-26 07:45:49,267 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36866.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 07:45:49,865 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.2115, 1.1978, 1.3050, 0.4899, 1.2292, 1.5239, 1.5156, 1.2117], + device='cuda:4'), covar=tensor([0.1042, 0.0808, 0.0459, 0.0654, 0.0502, 0.0446, 0.0392, 0.0683], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0155, 0.0119, 0.0136, 0.0131, 0.0123, 0.0144, 0.0144], + device='cuda:4'), out_proj_covar=tensor([9.5943e-05, 1.1422e-04, 8.5866e-05, 9.9105e-05, 9.3946e-05, 9.0602e-05, + 1.0653e-04, 1.0606e-04], device='cuda:4') +2023-03-26 07:45:50,127 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 07:46:12,541 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=36885.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:46:33,404 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5271, 1.7118, 1.2935, 1.5051, 1.8012, 1.7793, 1.5694, 1.3680], + device='cuda:4'), covar=tensor([0.0411, 0.0249, 0.0538, 0.0300, 0.0251, 0.0410, 0.0298, 0.0395], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0111, 0.0139, 0.0116, 0.0104, 0.0100, 0.0091, 0.0109], + device='cuda:4'), out_proj_covar=tensor([6.9632e-05, 8.7247e-05, 1.1114e-04, 9.1753e-05, 8.2224e-05, 7.4498e-05, + 6.9154e-05, 8.5136e-05], device='cuda:4') +2023-03-26 07:46:44,109 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36911.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:46:52,300 INFO [finetune.py:976] (4/7) Epoch 7, batch 2550, loss[loss=0.1894, simple_loss=0.2622, pruned_loss=0.05828, over 4924.00 frames. ], tot_loss[loss=0.211, simple_loss=0.2725, pruned_loss=0.07478, over 955150.40 frames. ], batch size: 38, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:47:03,264 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.261e+02 1.620e+02 1.912e+02 2.307e+02 6.491e+02, threshold=3.825e+02, percent-clipped=1.0 +2023-03-26 07:47:45,969 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36958.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:47:47,752 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36961.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:47:55,200 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36964.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:47:56,360 INFO [finetune.py:976] (4/7) Epoch 7, batch 2600, loss[loss=0.2526, simple_loss=0.3043, pruned_loss=0.1005, over 4884.00 frames. ], tot_loss[loss=0.2142, simple_loss=0.2757, pruned_loss=0.07634, over 954853.92 frames. ], batch size: 32, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:47:57,656 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36968.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:48:04,500 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6196, 3.4583, 3.2239, 1.5544, 3.5390, 2.6319, 0.7389, 2.2997], + device='cuda:4'), covar=tensor([0.2440, 0.2017, 0.1881, 0.3727, 0.1119, 0.1083, 0.4726, 0.1687], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0172, 0.0161, 0.0128, 0.0152, 0.0122, 0.0145, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 07:48:05,179 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36972.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:48:41,498 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37004.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:48:53,699 INFO [finetune.py:976] (4/7) Epoch 7, batch 2650, loss[loss=0.2094, simple_loss=0.2675, pruned_loss=0.07563, over 4832.00 frames. ], tot_loss[loss=0.2137, simple_loss=0.2755, pruned_loss=0.07595, over 953939.87 frames. ], batch size: 30, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:49:01,191 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37019.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:49:05,451 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.627e+02 1.954e+02 2.393e+02 3.704e+02, threshold=3.907e+02, percent-clipped=0.0 +2023-03-26 07:49:41,698 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37056.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 07:49:47,723 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37065.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:49:48,223 INFO [finetune.py:976] (4/7) Epoch 7, batch 2700, loss[loss=0.1875, simple_loss=0.2567, pruned_loss=0.05914, over 4831.00 frames. ], tot_loss[loss=0.2122, simple_loss=0.2739, pruned_loss=0.07529, over 954288.06 frames. ], batch size: 38, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:49:51,940 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.8591, 1.5324, 1.5359, 0.9048, 1.6608, 1.8066, 1.7446, 1.5044], + device='cuda:4'), covar=tensor([0.0769, 0.0622, 0.0421, 0.0614, 0.0403, 0.0438, 0.0353, 0.0593], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0156, 0.0120, 0.0136, 0.0131, 0.0124, 0.0145, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.6165e-05, 1.1505e-04, 8.6666e-05, 9.9163e-05, 9.3989e-05, 9.1128e-05, + 1.0683e-04, 1.0680e-04], device='cuda:4') +2023-03-26 07:50:04,747 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-26 07:50:07,268 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7022, 1.2111, 0.8549, 1.7463, 2.0803, 1.4248, 1.4013, 1.6997], + device='cuda:4'), covar=tensor([0.1318, 0.1983, 0.2004, 0.1113, 0.1982, 0.1962, 0.1394, 0.1701], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0097, 0.0113, 0.0092, 0.0123, 0.0096, 0.0100, 0.0092], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 07:50:08,924 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.3127, 2.9225, 3.0400, 3.2357, 3.0867, 2.8900, 3.3355, 0.9951], + device='cuda:4'), covar=tensor([0.1035, 0.0971, 0.1006, 0.1015, 0.1542, 0.1560, 0.1139, 0.5010], + device='cuda:4'), in_proj_covar=tensor([0.0348, 0.0243, 0.0274, 0.0294, 0.0332, 0.0283, 0.0303, 0.0297], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 07:50:16,796 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1223, 2.0701, 2.1068, 1.3662, 2.2519, 2.2591, 2.1903, 1.7411], + device='cuda:4'), covar=tensor([0.0587, 0.0691, 0.0793, 0.1039, 0.0544, 0.0699, 0.0604, 0.1138], + device='cuda:4'), in_proj_covar=tensor([0.0138, 0.0134, 0.0144, 0.0127, 0.0114, 0.0145, 0.0147, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 07:50:22,113 INFO [finetune.py:976] (4/7) Epoch 7, batch 2750, loss[loss=0.2062, simple_loss=0.277, pruned_loss=0.06769, over 4829.00 frames. ], tot_loss[loss=0.2111, simple_loss=0.2718, pruned_loss=0.07525, over 955336.63 frames. ], batch size: 47, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:50:28,699 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.127e+02 1.629e+02 1.991e+02 2.307e+02 4.303e+02, threshold=3.983e+02, percent-clipped=1.0 +2023-03-26 07:50:58,191 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37161.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 07:51:03,501 INFO [finetune.py:976] (4/7) Epoch 7, batch 2800, loss[loss=0.1997, simple_loss=0.2641, pruned_loss=0.06767, over 4826.00 frames. ], tot_loss[loss=0.2072, simple_loss=0.2677, pruned_loss=0.07331, over 955837.76 frames. ], batch size: 30, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:52:00,005 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7577, 1.6815, 1.4847, 1.8566, 2.0136, 1.8249, 1.2751, 1.4357], + device='cuda:4'), covar=tensor([0.2206, 0.2106, 0.1935, 0.1712, 0.1990, 0.1189, 0.2851, 0.1917], + device='cuda:4'), in_proj_covar=tensor([0.0237, 0.0209, 0.0204, 0.0187, 0.0240, 0.0178, 0.0214, 0.0192], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 07:52:09,301 INFO [finetune.py:976] (4/7) Epoch 7, batch 2850, loss[loss=0.1825, simple_loss=0.2522, pruned_loss=0.05646, over 4935.00 frames. ], tot_loss[loss=0.2056, simple_loss=0.2665, pruned_loss=0.07237, over 956870.45 frames. ], batch size: 38, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:52:20,862 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.582e+02 1.929e+02 2.327e+02 4.539e+02, threshold=3.857e+02, percent-clipped=3.0 +2023-03-26 07:53:00,974 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37261.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:53:02,783 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37264.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:53:03,914 INFO [finetune.py:976] (4/7) Epoch 7, batch 2900, loss[loss=0.279, simple_loss=0.3393, pruned_loss=0.1094, over 4905.00 frames. ], tot_loss[loss=0.2085, simple_loss=0.2695, pruned_loss=0.07372, over 957203.87 frames. ], batch size: 36, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:53:09,377 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37267.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:53:09,981 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37268.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:53:09,998 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37268.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:53:56,553 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5291, 1.5236, 1.2802, 1.3740, 1.8299, 1.7601, 1.5459, 1.3283], + device='cuda:4'), covar=tensor([0.0315, 0.0312, 0.0526, 0.0315, 0.0193, 0.0434, 0.0281, 0.0408], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0111, 0.0139, 0.0116, 0.0104, 0.0100, 0.0091, 0.0110], + device='cuda:4'), out_proj_covar=tensor([6.9383e-05, 8.7204e-05, 1.1121e-04, 9.1629e-05, 8.2401e-05, 7.4558e-05, + 6.8916e-05, 8.5419e-05], device='cuda:4') +2023-03-26 07:54:04,930 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37309.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:54:06,758 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37312.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:54:07,998 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37314.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:54:13,848 INFO [finetune.py:976] (4/7) Epoch 7, batch 2950, loss[loss=0.188, simple_loss=0.2467, pruned_loss=0.06461, over 4377.00 frames. ], tot_loss[loss=0.2107, simple_loss=0.2722, pruned_loss=0.07462, over 956601.28 frames. ], batch size: 19, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:54:13,909 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37316.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:54:25,441 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.193e+02 1.702e+02 2.045e+02 2.514e+02 5.908e+02, threshold=4.090e+02, percent-clipped=3.0 +2023-03-26 07:54:27,404 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37329.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:55:00,473 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37356.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 07:55:03,443 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-26 07:55:05,686 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37360.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:55:08,104 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37364.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:55:09,242 INFO [finetune.py:976] (4/7) Epoch 7, batch 3000, loss[loss=0.2593, simple_loss=0.3191, pruned_loss=0.09977, over 4926.00 frames. ], tot_loss[loss=0.2133, simple_loss=0.2749, pruned_loss=0.07587, over 958140.13 frames. ], batch size: 33, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:55:09,242 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 07:55:25,743 INFO [finetune.py:1010] (4/7) Epoch 7, validation: loss=0.161, simple_loss=0.2327, pruned_loss=0.04464, over 2265189.00 frames. +2023-03-26 07:55:25,743 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6301MB +2023-03-26 07:55:51,453 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9038, 1.6950, 1.4588, 1.6339, 1.6210, 1.6261, 1.6135, 2.3483], + device='cuda:4'), covar=tensor([0.5453, 0.5889, 0.4334, 0.5602, 0.5294, 0.2982, 0.5553, 0.2248], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0259, 0.0221, 0.0282, 0.0241, 0.0205, 0.0245, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 07:55:54,505 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37404.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 07:56:02,292 INFO [finetune.py:976] (4/7) Epoch 7, batch 3050, loss[loss=0.2171, simple_loss=0.2806, pruned_loss=0.07679, over 4887.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.2749, pruned_loss=0.07537, over 957926.70 frames. ], batch size: 36, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:56:11,552 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37425.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:56:12,025 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.173e+02 1.581e+02 1.871e+02 2.387e+02 4.591e+02, threshold=3.742e+02, percent-clipped=1.0 +2023-03-26 07:56:48,983 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37461.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 07:56:57,055 INFO [finetune.py:976] (4/7) Epoch 7, batch 3100, loss[loss=0.2327, simple_loss=0.2863, pruned_loss=0.08952, over 4843.00 frames. ], tot_loss[loss=0.2105, simple_loss=0.2725, pruned_loss=0.07425, over 958013.26 frames. ], batch size: 49, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:57:50,878 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.5221, 3.0627, 2.8262, 1.5163, 3.0228, 2.6119, 2.4839, 2.6544], + device='cuda:4'), covar=tensor([0.0776, 0.0957, 0.1861, 0.2557, 0.1778, 0.2105, 0.1998, 0.1289], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0201, 0.0200, 0.0188, 0.0218, 0.0206, 0.0221, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 07:57:52,635 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37509.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 07:58:01,947 INFO [finetune.py:976] (4/7) Epoch 7, batch 3150, loss[loss=0.1638, simple_loss=0.2337, pruned_loss=0.04699, over 4907.00 frames. ], tot_loss[loss=0.2078, simple_loss=0.2693, pruned_loss=0.07317, over 958504.04 frames. ], batch size: 37, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:58:13,090 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.225e+02 1.704e+02 2.041e+02 2.515e+02 5.799e+02, threshold=4.081e+02, percent-clipped=3.0 +2023-03-26 07:58:20,737 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1951, 1.4386, 2.0169, 1.9077, 1.8251, 1.7775, 1.8368, 1.8609], + device='cuda:4'), covar=tensor([0.4050, 0.5456, 0.5054, 0.4965, 0.6477, 0.4570, 0.6371, 0.4435], + device='cuda:4'), in_proj_covar=tensor([0.0229, 0.0242, 0.0255, 0.0254, 0.0244, 0.0220, 0.0272, 0.0226], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 07:59:05,775 INFO [finetune.py:976] (4/7) Epoch 7, batch 3200, loss[loss=0.1645, simple_loss=0.229, pruned_loss=0.05001, over 4810.00 frames. ], tot_loss[loss=0.2045, simple_loss=0.2658, pruned_loss=0.07153, over 959393.44 frames. ], batch size: 45, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 07:59:06,472 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37567.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 07:59:23,850 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 08:00:08,891 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37614.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:00:14,410 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37615.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:00:14,959 INFO [finetune.py:976] (4/7) Epoch 7, batch 3250, loss[loss=0.2232, simple_loss=0.2929, pruned_loss=0.07674, over 4833.00 frames. ], tot_loss[loss=0.2048, simple_loss=0.2663, pruned_loss=0.07161, over 957560.88 frames. ], batch size: 39, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 08:00:24,954 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37624.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:00:26,116 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.216e+02 1.664e+02 1.918e+02 2.274e+02 4.430e+02, threshold=3.836e+02, percent-clipped=1.0 +2023-03-26 08:01:09,566 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37660.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:01:10,725 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37662.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:01:18,269 INFO [finetune.py:976] (4/7) Epoch 7, batch 3300, loss[loss=0.2329, simple_loss=0.2977, pruned_loss=0.08408, over 4898.00 frames. ], tot_loss[loss=0.2078, simple_loss=0.2698, pruned_loss=0.0729, over 956212.64 frames. ], batch size: 43, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 08:02:12,455 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37708.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:02:22,973 INFO [finetune.py:976] (4/7) Epoch 7, batch 3350, loss[loss=0.2566, simple_loss=0.3045, pruned_loss=0.1044, over 4833.00 frames. ], tot_loss[loss=0.2097, simple_loss=0.2718, pruned_loss=0.07378, over 956990.81 frames. ], batch size: 49, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 08:02:25,470 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37720.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:02:34,111 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.166e+02 1.767e+02 2.019e+02 2.457e+02 5.992e+02, threshold=4.038e+02, percent-clipped=4.0 +2023-03-26 08:03:28,098 INFO [finetune.py:976] (4/7) Epoch 7, batch 3400, loss[loss=0.2518, simple_loss=0.3197, pruned_loss=0.09197, over 4907.00 frames. ], tot_loss[loss=0.2105, simple_loss=0.2737, pruned_loss=0.07372, over 957335.26 frames. ], batch size: 37, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 08:03:48,724 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2559, 1.9097, 2.6644, 4.2874, 2.9894, 2.7748, 0.9214, 3.3922], + device='cuda:4'), covar=tensor([0.1666, 0.1459, 0.1341, 0.0510, 0.0726, 0.1345, 0.2064, 0.0463], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0117, 0.0135, 0.0165, 0.0101, 0.0140, 0.0128, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 08:04:32,126 INFO [finetune.py:976] (4/7) Epoch 7, batch 3450, loss[loss=0.2194, simple_loss=0.2834, pruned_loss=0.07764, over 4786.00 frames. ], tot_loss[loss=0.2093, simple_loss=0.2723, pruned_loss=0.07318, over 954616.84 frames. ], batch size: 29, lr: 3.87e-03, grad_scale: 32.0 +2023-03-26 08:04:43,341 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.715e+02 1.991e+02 2.496e+02 6.747e+02, threshold=3.982e+02, percent-clipped=3.0 +2023-03-26 08:05:23,836 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 08:05:36,356 INFO [finetune.py:976] (4/7) Epoch 7, batch 3500, loss[loss=0.2182, simple_loss=0.2654, pruned_loss=0.08551, over 4915.00 frames. ], tot_loss[loss=0.2092, simple_loss=0.2713, pruned_loss=0.07356, over 955837.72 frames. ], batch size: 46, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:06:41,167 INFO [finetune.py:976] (4/7) Epoch 7, batch 3550, loss[loss=0.1605, simple_loss=0.2325, pruned_loss=0.04424, over 4763.00 frames. ], tot_loss[loss=0.207, simple_loss=0.2687, pruned_loss=0.07262, over 957606.58 frames. ], batch size: 28, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:06:49,360 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 08:06:57,411 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37924.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:06:58,533 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.177e+02 1.559e+02 1.846e+02 2.185e+02 4.242e+02, threshold=3.693e+02, percent-clipped=1.0 +2023-03-26 08:07:02,904 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6655, 2.2212, 1.9864, 0.9954, 2.1805, 2.0175, 1.6793, 2.0125], + device='cuda:4'), covar=tensor([0.0700, 0.1056, 0.1679, 0.2216, 0.1573, 0.2249, 0.2234, 0.1162], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0200, 0.0199, 0.0188, 0.0217, 0.0205, 0.0220, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:07:08,785 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3653, 1.7682, 0.8224, 2.2773, 2.5692, 1.7811, 2.0733, 2.2001], + device='cuda:4'), covar=tensor([0.1387, 0.1921, 0.2100, 0.1126, 0.1762, 0.1936, 0.1336, 0.1888], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0098, 0.0114, 0.0093, 0.0124, 0.0097, 0.0101, 0.0093], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 08:07:43,975 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-26 08:07:52,118 INFO [finetune.py:976] (4/7) Epoch 7, batch 3600, loss[loss=0.1915, simple_loss=0.2583, pruned_loss=0.06233, over 4866.00 frames. ], tot_loss[loss=0.2047, simple_loss=0.2658, pruned_loss=0.07181, over 955532.22 frames. ], batch size: 44, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:08:01,700 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=37972.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:08:06,543 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37979.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:08:25,005 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4758, 1.0326, 0.7076, 1.3720, 1.8927, 0.6942, 1.2681, 1.4381], + device='cuda:4'), covar=tensor([0.1651, 0.2336, 0.1921, 0.1261, 0.2096, 0.2052, 0.1559, 0.2027], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0098, 0.0114, 0.0093, 0.0124, 0.0096, 0.0101, 0.0093], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 08:08:33,954 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3753, 1.2626, 1.6311, 2.3891, 1.6509, 2.0721, 0.8985, 1.9915], + device='cuda:4'), covar=tensor([0.1722, 0.1585, 0.1167, 0.0737, 0.0935, 0.1248, 0.1660, 0.0681], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0117, 0.0135, 0.0165, 0.0101, 0.0139, 0.0128, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 08:08:47,800 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9747, 1.6918, 1.5685, 1.6317, 1.9729, 2.0870, 1.8357, 1.5333], + device='cuda:4'), covar=tensor([0.0220, 0.0249, 0.0474, 0.0258, 0.0193, 0.0319, 0.0267, 0.0345], + device='cuda:4'), in_proj_covar=tensor([0.0088, 0.0110, 0.0137, 0.0115, 0.0103, 0.0099, 0.0090, 0.0108], + device='cuda:4'), out_proj_covar=tensor([6.8488e-05, 8.6588e-05, 1.0969e-04, 9.0146e-05, 8.1328e-05, 7.3766e-05, + 6.8117e-05, 8.4168e-05], device='cuda:4') +2023-03-26 08:08:59,069 INFO [finetune.py:976] (4/7) Epoch 7, batch 3650, loss[loss=0.2739, simple_loss=0.3322, pruned_loss=0.1078, over 4810.00 frames. ], tot_loss[loss=0.2081, simple_loss=0.2688, pruned_loss=0.07368, over 956038.09 frames. ], batch size: 41, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:09:07,237 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=38020.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:09:10,803 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.699e+02 2.068e+02 2.418e+02 4.148e+02, threshold=4.136e+02, percent-clipped=4.0 +2023-03-26 08:09:27,930 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=38040.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:09:28,523 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6602, 1.2129, 0.8217, 1.4921, 2.0381, 1.1437, 1.3596, 1.6038], + device='cuda:4'), covar=tensor([0.1644, 0.2304, 0.2085, 0.1357, 0.2043, 0.2147, 0.1621, 0.2049], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0098, 0.0114, 0.0093, 0.0124, 0.0097, 0.0101, 0.0093], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 08:09:46,800 INFO [finetune.py:976] (4/7) Epoch 7, batch 3700, loss[loss=0.1789, simple_loss=0.2386, pruned_loss=0.0596, over 4771.00 frames. ], tot_loss[loss=0.2099, simple_loss=0.2711, pruned_loss=0.07435, over 952146.50 frames. ], batch size: 26, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:09:48,562 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=38068.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:10:03,309 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-26 08:10:19,922 INFO [finetune.py:976] (4/7) Epoch 7, batch 3750, loss[loss=0.2468, simple_loss=0.2964, pruned_loss=0.09858, over 4864.00 frames. ], tot_loss[loss=0.2121, simple_loss=0.2736, pruned_loss=0.07528, over 952181.78 frames. ], batch size: 31, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:10:26,925 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.737e+01 1.627e+02 1.982e+02 2.503e+02 4.763e+02, threshold=3.965e+02, percent-clipped=1.0 +2023-03-26 08:10:30,109 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7982, 1.3802, 0.8291, 1.6647, 2.1217, 1.5185, 1.6065, 1.9245], + device='cuda:4'), covar=tensor([0.1373, 0.1878, 0.1979, 0.1117, 0.1831, 0.1902, 0.1308, 0.1583], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0098, 0.0114, 0.0093, 0.0124, 0.0096, 0.0101, 0.0093], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 08:10:57,173 INFO [finetune.py:976] (4/7) Epoch 7, batch 3800, loss[loss=0.2403, simple_loss=0.3057, pruned_loss=0.08743, over 4909.00 frames. ], tot_loss[loss=0.2145, simple_loss=0.2759, pruned_loss=0.07655, over 952473.90 frames. ], batch size: 37, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:11:04,426 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-26 08:11:28,163 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 08:11:30,370 INFO [finetune.py:976] (4/7) Epoch 7, batch 3850, loss[loss=0.1605, simple_loss=0.2312, pruned_loss=0.04489, over 4773.00 frames. ], tot_loss[loss=0.2117, simple_loss=0.2736, pruned_loss=0.07489, over 951902.07 frames. ], batch size: 29, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:11:43,043 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.610e+02 2.090e+02 2.406e+02 4.877e+02, threshold=4.181e+02, percent-clipped=2.0 +2023-03-26 08:12:02,983 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6467, 1.4707, 1.0591, 0.2783, 1.2771, 1.4850, 1.4044, 1.4495], + device='cuda:4'), covar=tensor([0.0887, 0.0827, 0.1312, 0.2040, 0.1489, 0.2171, 0.2293, 0.0856], + device='cuda:4'), in_proj_covar=tensor([0.0171, 0.0203, 0.0202, 0.0190, 0.0219, 0.0208, 0.0224, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:12:25,296 INFO [finetune.py:976] (4/7) Epoch 7, batch 3900, loss[loss=0.1582, simple_loss=0.2113, pruned_loss=0.05256, over 2784.00 frames. ], tot_loss[loss=0.2096, simple_loss=0.2711, pruned_loss=0.07406, over 952341.63 frames. ], batch size: 11, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:13:28,058 INFO [finetune.py:976] (4/7) Epoch 7, batch 3950, loss[loss=0.1783, simple_loss=0.2315, pruned_loss=0.0626, over 4827.00 frames. ], tot_loss[loss=0.2073, simple_loss=0.268, pruned_loss=0.07328, over 952032.10 frames. ], batch size: 25, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:13:45,514 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.693e+02 1.988e+02 2.374e+02 4.679e+02, threshold=3.976e+02, percent-clipped=1.0 +2023-03-26 08:13:56,367 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=38335.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:14:21,701 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=38360.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:14:22,866 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8648, 2.1149, 1.3195, 2.6693, 3.0373, 2.4368, 2.4954, 2.7657], + device='cuda:4'), covar=tensor([0.1165, 0.1855, 0.1883, 0.1032, 0.1552, 0.1468, 0.1204, 0.1729], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0098, 0.0115, 0.0093, 0.0124, 0.0096, 0.0101, 0.0093], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 08:14:25,158 INFO [finetune.py:976] (4/7) Epoch 7, batch 4000, loss[loss=0.2238, simple_loss=0.2843, pruned_loss=0.08163, over 4871.00 frames. ], tot_loss[loss=0.2057, simple_loss=0.2663, pruned_loss=0.07256, over 952686.01 frames. ], batch size: 34, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:15:29,787 INFO [finetune.py:976] (4/7) Epoch 7, batch 4050, loss[loss=0.2602, simple_loss=0.309, pruned_loss=0.1057, over 4817.00 frames. ], tot_loss[loss=0.2108, simple_loss=0.2712, pruned_loss=0.0752, over 952574.33 frames. ], batch size: 39, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:15:33,976 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=38421.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:15:42,069 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.248e+02 1.778e+02 2.129e+02 2.625e+02 5.238e+02, threshold=4.258e+02, percent-clipped=5.0 +2023-03-26 08:16:32,428 INFO [finetune.py:976] (4/7) Epoch 7, batch 4100, loss[loss=0.1813, simple_loss=0.2458, pruned_loss=0.0584, over 4862.00 frames. ], tot_loss[loss=0.2125, simple_loss=0.274, pruned_loss=0.07552, over 954098.93 frames. ], batch size: 31, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:16:40,201 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 08:17:31,591 INFO [finetune.py:976] (4/7) Epoch 7, batch 4150, loss[loss=0.2155, simple_loss=0.2818, pruned_loss=0.07463, over 4750.00 frames. ], tot_loss[loss=0.2133, simple_loss=0.275, pruned_loss=0.07582, over 956125.04 frames. ], batch size: 59, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:17:43,680 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.188e+02 1.723e+02 2.145e+02 2.598e+02 6.605e+02, threshold=4.291e+02, percent-clipped=2.0 +2023-03-26 08:18:25,239 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 08:18:34,212 INFO [finetune.py:976] (4/7) Epoch 7, batch 4200, loss[loss=0.238, simple_loss=0.2966, pruned_loss=0.08974, over 4869.00 frames. ], tot_loss[loss=0.2125, simple_loss=0.2744, pruned_loss=0.07527, over 953173.57 frames. ], batch size: 34, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:18:35,587 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9958, 1.9381, 1.6196, 2.0174, 1.9128, 1.8461, 1.8347, 2.6371], + device='cuda:4'), covar=tensor([0.5394, 0.7100, 0.4814, 0.6121, 0.6221, 0.3186, 0.6327, 0.2089], + device='cuda:4'), in_proj_covar=tensor([0.0284, 0.0260, 0.0221, 0.0281, 0.0241, 0.0206, 0.0245, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:19:34,113 INFO [finetune.py:976] (4/7) Epoch 7, batch 4250, loss[loss=0.2166, simple_loss=0.2778, pruned_loss=0.07767, over 4830.00 frames. ], tot_loss[loss=0.2097, simple_loss=0.2713, pruned_loss=0.07406, over 953869.69 frames. ], batch size: 33, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:19:44,840 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.606e+02 1.980e+02 2.259e+02 5.740e+02, threshold=3.960e+02, percent-clipped=2.0 +2023-03-26 08:20:02,312 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=38635.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:20:20,953 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6785, 1.4406, 1.3987, 1.4901, 1.7634, 1.7318, 1.5576, 1.2835], + device='cuda:4'), covar=tensor([0.0245, 0.0317, 0.0520, 0.0301, 0.0252, 0.0403, 0.0371, 0.0456], + device='cuda:4'), in_proj_covar=tensor([0.0088, 0.0111, 0.0139, 0.0115, 0.0104, 0.0100, 0.0091, 0.0109], + device='cuda:4'), out_proj_covar=tensor([6.9204e-05, 8.7284e-05, 1.1099e-04, 9.0365e-05, 8.1888e-05, 7.4340e-05, + 6.8689e-05, 8.4809e-05], device='cuda:4') +2023-03-26 08:20:38,695 INFO [finetune.py:976] (4/7) Epoch 7, batch 4300, loss[loss=0.2392, simple_loss=0.2921, pruned_loss=0.09318, over 4914.00 frames. ], tot_loss[loss=0.2088, simple_loss=0.27, pruned_loss=0.07384, over 955528.59 frames. ], batch size: 36, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:20:59,383 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=38683.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:21:40,757 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6540, 1.7174, 1.7220, 0.9449, 1.7898, 2.0420, 1.9396, 1.5396], + device='cuda:4'), covar=tensor([0.0934, 0.0662, 0.0417, 0.0619, 0.0380, 0.0453, 0.0355, 0.0697], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0155, 0.0120, 0.0136, 0.0131, 0.0123, 0.0144, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.5836e-05, 1.1438e-04, 8.6785e-05, 9.9168e-05, 9.3935e-05, 9.0872e-05, + 1.0621e-04, 1.0708e-04], device='cuda:4') +2023-03-26 08:21:41,239 INFO [finetune.py:976] (4/7) Epoch 7, batch 4350, loss[loss=0.2008, simple_loss=0.2577, pruned_loss=0.07198, over 4934.00 frames. ], tot_loss[loss=0.2063, simple_loss=0.2671, pruned_loss=0.07271, over 955793.57 frames. ], batch size: 33, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:21:41,323 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=38716.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:21:42,733 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-03-26 08:21:52,346 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.223e+02 1.679e+02 1.871e+02 2.197e+02 5.866e+02, threshold=3.741e+02, percent-clipped=4.0 +2023-03-26 08:22:43,659 INFO [finetune.py:976] (4/7) Epoch 7, batch 4400, loss[loss=0.2267, simple_loss=0.2898, pruned_loss=0.08182, over 4906.00 frames. ], tot_loss[loss=0.2074, simple_loss=0.2684, pruned_loss=0.0732, over 954578.72 frames. ], batch size: 36, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:23:42,200 INFO [finetune.py:976] (4/7) Epoch 7, batch 4450, loss[loss=0.2063, simple_loss=0.2853, pruned_loss=0.06366, over 4781.00 frames. ], tot_loss[loss=0.2094, simple_loss=0.2711, pruned_loss=0.07385, over 953918.93 frames. ], batch size: 29, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:23:51,917 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=38823.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:23:53,602 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.165e+02 1.712e+02 1.965e+02 2.330e+02 4.727e+02, threshold=3.929e+02, percent-clipped=4.0 +2023-03-26 08:24:02,099 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.65 vs. limit=2.0 +2023-03-26 08:24:44,784 INFO [finetune.py:976] (4/7) Epoch 7, batch 4500, loss[loss=0.2164, simple_loss=0.28, pruned_loss=0.0764, over 4835.00 frames. ], tot_loss[loss=0.2109, simple_loss=0.2723, pruned_loss=0.07478, over 953782.08 frames. ], batch size: 49, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:25:06,332 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=38884.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:25:49,102 INFO [finetune.py:976] (4/7) Epoch 7, batch 4550, loss[loss=0.2203, simple_loss=0.2876, pruned_loss=0.07647, over 4727.00 frames. ], tot_loss[loss=0.2126, simple_loss=0.2742, pruned_loss=0.07555, over 952362.70 frames. ], batch size: 59, lr: 3.86e-03, grad_scale: 64.0 +2023-03-26 08:25:59,508 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.041e+02 1.671e+02 2.000e+02 2.524e+02 3.434e+02, threshold=4.000e+02, percent-clipped=0.0 +2023-03-26 08:26:23,252 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 08:26:47,278 INFO [finetune.py:976] (4/7) Epoch 7, batch 4600, loss[loss=0.1545, simple_loss=0.2274, pruned_loss=0.04078, over 4812.00 frames. ], tot_loss[loss=0.2107, simple_loss=0.2727, pruned_loss=0.07435, over 953682.02 frames. ], batch size: 38, lr: 3.86e-03, grad_scale: 64.0 +2023-03-26 08:26:55,337 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6518, 1.6892, 2.1465, 2.0100, 1.8985, 4.3620, 1.6157, 1.9658], + device='cuda:4'), covar=tensor([0.0974, 0.1774, 0.1105, 0.0965, 0.1496, 0.0205, 0.1365, 0.1642], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0077, 0.0091, 0.0082, 0.0084, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 08:27:54,984 INFO [finetune.py:976] (4/7) Epoch 7, batch 4650, loss[loss=0.2204, simple_loss=0.2758, pruned_loss=0.08249, over 4791.00 frames. ], tot_loss[loss=0.2073, simple_loss=0.269, pruned_loss=0.07274, over 954225.75 frames. ], batch size: 29, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:27:55,086 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39016.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:28:06,170 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.761e+01 1.505e+02 1.924e+02 2.345e+02 4.238e+02, threshold=3.847e+02, percent-clipped=2.0 +2023-03-26 08:28:50,939 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=39064.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:28:57,994 INFO [finetune.py:976] (4/7) Epoch 7, batch 4700, loss[loss=0.1909, simple_loss=0.2481, pruned_loss=0.06679, over 4760.00 frames. ], tot_loss[loss=0.204, simple_loss=0.2658, pruned_loss=0.07107, over 954450.23 frames. ], batch size: 54, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:29:19,145 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8186, 1.6307, 2.1506, 3.5440, 2.5090, 2.4420, 1.0751, 2.7623], + device='cuda:4'), covar=tensor([0.1700, 0.1597, 0.1499, 0.0546, 0.0759, 0.1551, 0.1897, 0.0568], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0117, 0.0134, 0.0165, 0.0101, 0.0140, 0.0127, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 08:29:56,018 INFO [finetune.py:976] (4/7) Epoch 7, batch 4750, loss[loss=0.1692, simple_loss=0.2493, pruned_loss=0.04455, over 4849.00 frames. ], tot_loss[loss=0.2032, simple_loss=0.2645, pruned_loss=0.07098, over 953921.05 frames. ], batch size: 44, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:30:08,812 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.221e+02 1.609e+02 1.819e+02 2.206e+02 4.512e+02, threshold=3.638e+02, percent-clipped=2.0 +2023-03-26 08:30:17,390 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0029, 1.7928, 1.5534, 1.6584, 1.7441, 1.7361, 1.7212, 2.5477], + device='cuda:4'), covar=tensor([0.5193, 0.5747, 0.4057, 0.5332, 0.4924, 0.3055, 0.5200, 0.1904], + device='cuda:4'), in_proj_covar=tensor([0.0283, 0.0258, 0.0220, 0.0281, 0.0241, 0.0206, 0.0245, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:30:58,730 INFO [finetune.py:976] (4/7) Epoch 7, batch 4800, loss[loss=0.1975, simple_loss=0.2689, pruned_loss=0.063, over 4788.00 frames. ], tot_loss[loss=0.2064, simple_loss=0.2674, pruned_loss=0.07271, over 952927.91 frames. ], batch size: 29, lr: 3.86e-03, grad_scale: 32.0 +2023-03-26 08:31:12,897 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39179.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:31:22,034 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39185.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:31:32,222 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-26 08:31:57,092 INFO [finetune.py:976] (4/7) Epoch 7, batch 4850, loss[loss=0.2032, simple_loss=0.2672, pruned_loss=0.06962, over 4819.00 frames. ], tot_loss[loss=0.2106, simple_loss=0.2721, pruned_loss=0.07448, over 953978.73 frames. ], batch size: 30, lr: 3.86e-03, grad_scale: 16.0 +2023-03-26 08:32:06,049 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.691e+02 2.004e+02 2.499e+02 4.240e+02, threshold=4.008e+02, percent-clipped=2.0 +2023-03-26 08:32:18,174 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39246.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:32:30,570 INFO [finetune.py:976] (4/7) Epoch 7, batch 4900, loss[loss=0.2316, simple_loss=0.2823, pruned_loss=0.09047, over 4112.00 frames. ], tot_loss[loss=0.2108, simple_loss=0.2728, pruned_loss=0.07441, over 951410.94 frames. ], batch size: 66, lr: 3.86e-03, grad_scale: 16.0 +2023-03-26 08:32:31,113 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.66 vs. limit=5.0 +2023-03-26 08:32:43,187 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39284.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:32:48,599 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7109, 1.5266, 2.0623, 1.8588, 1.6771, 4.1416, 1.5978, 1.9017], + device='cuda:4'), covar=tensor([0.0950, 0.1865, 0.1150, 0.1048, 0.1736, 0.0204, 0.1445, 0.1738], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0076, 0.0078, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 08:33:03,516 INFO [finetune.py:976] (4/7) Epoch 7, batch 4950, loss[loss=0.2014, simple_loss=0.2661, pruned_loss=0.0683, over 4814.00 frames. ], tot_loss[loss=0.2111, simple_loss=0.2734, pruned_loss=0.07444, over 950804.66 frames. ], batch size: 33, lr: 3.86e-03, grad_scale: 16.0 +2023-03-26 08:33:09,351 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6007, 1.5356, 2.0906, 1.8952, 1.6918, 4.2696, 1.5833, 1.9390], + device='cuda:4'), covar=tensor([0.0975, 0.1897, 0.1123, 0.1033, 0.1710, 0.0186, 0.1392, 0.1711], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0082, 0.0076, 0.0078, 0.0093, 0.0084, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 08:33:12,724 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.619e+02 1.980e+02 2.423e+02 3.796e+02, threshold=3.961e+02, percent-clipped=0.0 +2023-03-26 08:33:24,223 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39345.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:33:37,225 INFO [finetune.py:976] (4/7) Epoch 7, batch 5000, loss[loss=0.1682, simple_loss=0.2332, pruned_loss=0.05157, over 4899.00 frames. ], tot_loss[loss=0.2076, simple_loss=0.2703, pruned_loss=0.07246, over 951877.68 frames. ], batch size: 46, lr: 3.86e-03, grad_scale: 16.0 +2023-03-26 08:34:10,938 INFO [finetune.py:976] (4/7) Epoch 7, batch 5050, loss[loss=0.1651, simple_loss=0.2384, pruned_loss=0.04591, over 4770.00 frames. ], tot_loss[loss=0.2056, simple_loss=0.2675, pruned_loss=0.07185, over 952226.61 frames. ], batch size: 27, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:34:19,594 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.096e+02 1.623e+02 1.955e+02 2.404e+02 3.498e+02, threshold=3.910e+02, percent-clipped=0.0 +2023-03-26 08:34:20,319 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6072, 1.5632, 2.1966, 1.9220, 1.7668, 4.1921, 1.4370, 1.9675], + device='cuda:4'), covar=tensor([0.0983, 0.1772, 0.1067, 0.1033, 0.1622, 0.0184, 0.1511, 0.1634], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 08:34:52,729 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39463.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:34:54,000 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-26 08:34:54,407 INFO [finetune.py:976] (4/7) Epoch 7, batch 5100, loss[loss=0.1773, simple_loss=0.2459, pruned_loss=0.05438, over 4908.00 frames. ], tot_loss[loss=0.2032, simple_loss=0.265, pruned_loss=0.07069, over 953110.98 frames. ], batch size: 43, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:35:05,350 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39479.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:35:07,042 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9547, 1.9130, 1.5602, 2.1255, 2.0288, 1.6274, 2.4277, 2.0114], + device='cuda:4'), covar=tensor([0.1779, 0.3191, 0.3804, 0.3351, 0.3003, 0.2133, 0.3956, 0.2431], + device='cuda:4'), in_proj_covar=tensor([0.0172, 0.0191, 0.0236, 0.0256, 0.0236, 0.0194, 0.0213, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:35:14,968 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1524, 1.8380, 1.5087, 0.6179, 1.7519, 1.7617, 1.5541, 1.8289], + device='cuda:4'), covar=tensor([0.0987, 0.0913, 0.1404, 0.1986, 0.1359, 0.2258, 0.2303, 0.0869], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0199, 0.0200, 0.0187, 0.0217, 0.0205, 0.0222, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:35:28,753 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 08:35:31,914 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3153, 1.4177, 1.4066, 0.7370, 1.3941, 1.6493, 1.6865, 1.3236], + device='cuda:4'), covar=tensor([0.0855, 0.0515, 0.0471, 0.0555, 0.0422, 0.0507, 0.0303, 0.0650], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0156, 0.0120, 0.0137, 0.0131, 0.0124, 0.0144, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.6286e-05, 1.1501e-04, 8.7040e-05, 9.9619e-05, 9.4362e-05, 9.1533e-05, + 1.0598e-04, 1.0725e-04], device='cuda:4') +2023-03-26 08:35:39,560 INFO [finetune.py:976] (4/7) Epoch 7, batch 5150, loss[loss=0.1755, simple_loss=0.2507, pruned_loss=0.05015, over 4736.00 frames. ], tot_loss[loss=0.2048, simple_loss=0.2662, pruned_loss=0.07173, over 953937.30 frames. ], batch size: 59, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:35:47,020 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39524.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:35:49,284 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=39527.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:35:49,816 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.405e+01 1.710e+02 2.016e+02 2.412e+02 5.054e+02, threshold=4.032e+02, percent-clipped=2.0 +2023-03-26 08:35:57,236 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.8026, 3.3114, 3.4823, 3.6475, 3.5560, 3.3201, 3.8677, 1.1469], + device='cuda:4'), covar=tensor([0.0976, 0.0950, 0.0848, 0.1123, 0.1546, 0.1673, 0.0934, 0.5706], + device='cuda:4'), in_proj_covar=tensor([0.0353, 0.0245, 0.0277, 0.0295, 0.0334, 0.0284, 0.0306, 0.0299], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:36:08,544 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39541.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:36:21,692 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3302, 2.2297, 1.8206, 2.4657, 2.3477, 1.8897, 2.8096, 2.2773], + device='cuda:4'), covar=tensor([0.1613, 0.3267, 0.3848, 0.3407, 0.3032, 0.2099, 0.3691, 0.2478], + device='cuda:4'), in_proj_covar=tensor([0.0172, 0.0191, 0.0236, 0.0256, 0.0236, 0.0194, 0.0213, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:36:24,637 INFO [finetune.py:976] (4/7) Epoch 7, batch 5200, loss[loss=0.2302, simple_loss=0.2968, pruned_loss=0.08185, over 4925.00 frames. ], tot_loss[loss=0.2087, simple_loss=0.2707, pruned_loss=0.07338, over 953642.64 frames. ], batch size: 38, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:36:29,071 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4843, 1.4146, 1.6302, 1.7976, 1.4948, 3.1589, 1.2358, 1.5271], + device='cuda:4'), covar=tensor([0.0986, 0.1794, 0.1387, 0.0984, 0.1633, 0.0249, 0.1526, 0.1707], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 08:36:33,205 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4063, 2.3069, 1.9371, 2.5897, 2.5427, 2.0301, 3.0422, 2.4843], + device='cuda:4'), covar=tensor([0.1521, 0.2973, 0.3624, 0.3343, 0.2662, 0.1813, 0.3177, 0.2113], + device='cuda:4'), in_proj_covar=tensor([0.0172, 0.0190, 0.0236, 0.0255, 0.0236, 0.0193, 0.0212, 0.0193], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:36:59,671 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9088, 1.6818, 1.4526, 1.5119, 1.6428, 1.6327, 1.6221, 2.3452], + device='cuda:4'), covar=tensor([0.5191, 0.5320, 0.4043, 0.5123, 0.4682, 0.3021, 0.5227, 0.1946], + device='cuda:4'), in_proj_covar=tensor([0.0282, 0.0257, 0.0219, 0.0279, 0.0240, 0.0205, 0.0244, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:37:07,820 INFO [finetune.py:976] (4/7) Epoch 7, batch 5250, loss[loss=0.23, simple_loss=0.2986, pruned_loss=0.08071, over 4733.00 frames. ], tot_loss[loss=0.2107, simple_loss=0.2729, pruned_loss=0.07423, over 952181.63 frames. ], batch size: 54, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:37:07,929 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8508, 1.3309, 0.8412, 1.6515, 2.1730, 1.2708, 1.3994, 1.5767], + device='cuda:4'), covar=tensor([0.1546, 0.2206, 0.2147, 0.1334, 0.1948, 0.1944, 0.1673, 0.2086], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0096, 0.0113, 0.0092, 0.0122, 0.0095, 0.0099, 0.0091], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 08:37:15,022 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.077e+02 1.709e+02 2.070e+02 2.577e+02 5.953e+02, threshold=4.140e+02, percent-clipped=1.0 +2023-03-26 08:37:25,989 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39639.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 08:37:26,030 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0938, 1.9296, 1.5823, 1.9792, 2.0613, 1.7281, 2.3687, 2.1026], + device='cuda:4'), covar=tensor([0.1500, 0.2659, 0.3851, 0.3031, 0.2795, 0.1948, 0.3501, 0.2181], + device='cuda:4'), in_proj_covar=tensor([0.0172, 0.0191, 0.0236, 0.0255, 0.0236, 0.0193, 0.0212, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:37:26,592 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39640.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:37:36,637 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3717, 1.4523, 2.0294, 1.7556, 1.7593, 3.9065, 1.3889, 1.7442], + device='cuda:4'), covar=tensor([0.1116, 0.1912, 0.1219, 0.1126, 0.1652, 0.0273, 0.1656, 0.1865], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 08:37:43,682 INFO [finetune.py:976] (4/7) Epoch 7, batch 5300, loss[loss=0.2162, simple_loss=0.2913, pruned_loss=0.07055, over 4894.00 frames. ], tot_loss[loss=0.21, simple_loss=0.2728, pruned_loss=0.07356, over 952759.98 frames. ], batch size: 36, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:38:17,427 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39700.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 08:38:24,284 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.87 vs. limit=5.0 +2023-03-26 08:38:32,615 INFO [finetune.py:976] (4/7) Epoch 7, batch 5350, loss[loss=0.1572, simple_loss=0.2367, pruned_loss=0.03884, over 4896.00 frames. ], tot_loss[loss=0.2098, simple_loss=0.2727, pruned_loss=0.07339, over 953525.67 frames. ], batch size: 36, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:38:40,829 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.565e+02 1.855e+02 2.323e+02 5.491e+02, threshold=3.710e+02, percent-clipped=1.0 +2023-03-26 08:39:15,941 INFO [finetune.py:976] (4/7) Epoch 7, batch 5400, loss[loss=0.1937, simple_loss=0.2536, pruned_loss=0.06684, over 4770.00 frames. ], tot_loss[loss=0.2061, simple_loss=0.2689, pruned_loss=0.07161, over 954820.50 frames. ], batch size: 28, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:39:16,659 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39767.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:39:50,814 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39815.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:39:51,311 INFO [finetune.py:976] (4/7) Epoch 7, batch 5450, loss[loss=0.2133, simple_loss=0.2644, pruned_loss=0.0811, over 4839.00 frames. ], tot_loss[loss=0.2042, simple_loss=0.2664, pruned_loss=0.071, over 955389.93 frames. ], batch size: 25, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:39:53,197 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39819.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:40:03,634 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.339e+01 1.599e+02 1.928e+02 2.299e+02 3.698e+02, threshold=3.856e+02, percent-clipped=0.0 +2023-03-26 08:40:03,759 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39828.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:40:16,806 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39841.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:40:24,777 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.48 vs. limit=5.0 +2023-03-26 08:40:54,098 INFO [finetune.py:976] (4/7) Epoch 7, batch 5500, loss[loss=0.2159, simple_loss=0.2726, pruned_loss=0.07961, over 4814.00 frames. ], tot_loss[loss=0.2012, simple_loss=0.2631, pruned_loss=0.06963, over 956229.61 frames. ], batch size: 38, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:41:05,352 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39876.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:41:18,431 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=39889.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:41:57,093 INFO [finetune.py:976] (4/7) Epoch 7, batch 5550, loss[loss=0.1597, simple_loss=0.2294, pruned_loss=0.04497, over 4761.00 frames. ], tot_loss[loss=0.2041, simple_loss=0.2657, pruned_loss=0.07125, over 954192.03 frames. ], batch size: 28, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:42:09,838 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.056e+02 1.647e+02 1.995e+02 2.278e+02 3.177e+02, threshold=3.991e+02, percent-clipped=0.0 +2023-03-26 08:42:19,594 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6520, 1.4169, 1.8886, 1.3222, 1.6419, 1.7734, 1.3607, 2.0074], + device='cuda:4'), covar=tensor([0.1372, 0.2286, 0.1396, 0.1943, 0.1045, 0.1500, 0.3199, 0.0889], + device='cuda:4'), in_proj_covar=tensor([0.0201, 0.0204, 0.0198, 0.0195, 0.0181, 0.0222, 0.0218, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:42:28,021 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39940.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:42:58,492 INFO [finetune.py:976] (4/7) Epoch 7, batch 5600, loss[loss=0.2107, simple_loss=0.2624, pruned_loss=0.07947, over 4759.00 frames. ], tot_loss[loss=0.2081, simple_loss=0.27, pruned_loss=0.07306, over 952327.48 frames. ], batch size: 27, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:43:20,785 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=39988.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:43:22,579 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9126, 4.5265, 4.3332, 2.4249, 4.6485, 3.6131, 0.7571, 3.1513], + device='cuda:4'), covar=tensor([0.2416, 0.1986, 0.1307, 0.2830, 0.0827, 0.0762, 0.4506, 0.1325], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0173, 0.0162, 0.0130, 0.0156, 0.0123, 0.0147, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 08:43:30,216 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39995.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 08:43:40,406 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.74 vs. limit=5.0 +2023-03-26 08:43:52,204 INFO [finetune.py:976] (4/7) Epoch 7, batch 5650, loss[loss=0.21, simple_loss=0.2728, pruned_loss=0.07366, over 4817.00 frames. ], tot_loss[loss=0.2108, simple_loss=0.274, pruned_loss=0.07378, over 954822.40 frames. ], batch size: 33, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:44:09,057 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.182e+02 1.647e+02 1.995e+02 2.469e+02 4.643e+02, threshold=3.989e+02, percent-clipped=3.0 +2023-03-26 08:44:50,695 INFO [finetune.py:976] (4/7) Epoch 7, batch 5700, loss[loss=0.1602, simple_loss=0.2229, pruned_loss=0.04875, over 4177.00 frames. ], tot_loss[loss=0.2075, simple_loss=0.2698, pruned_loss=0.07262, over 939796.94 frames. ], batch size: 18, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:45:42,072 INFO [finetune.py:976] (4/7) Epoch 8, batch 0, loss[loss=0.2131, simple_loss=0.2702, pruned_loss=0.07797, over 4691.00 frames. ], tot_loss[loss=0.2131, simple_loss=0.2702, pruned_loss=0.07797, over 4691.00 frames. ], batch size: 59, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:45:42,072 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 08:45:49,825 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4141, 1.2513, 1.2507, 1.3634, 1.6155, 1.4955, 1.3563, 1.2216], + device='cuda:4'), covar=tensor([0.0356, 0.0307, 0.0680, 0.0283, 0.0258, 0.0513, 0.0287, 0.0376], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0113, 0.0140, 0.0116, 0.0105, 0.0101, 0.0092, 0.0110], + device='cuda:4'), out_proj_covar=tensor([7.0022e-05, 8.8500e-05, 1.1228e-04, 9.1282e-05, 8.2386e-05, 7.5241e-05, + 6.9340e-05, 8.5588e-05], device='cuda:4') +2023-03-26 08:45:57,866 INFO [finetune.py:1010] (4/7) Epoch 8, validation: loss=0.1624, simple_loss=0.234, pruned_loss=0.04544, over 2265189.00 frames. +2023-03-26 08:45:57,866 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6336MB +2023-03-26 08:46:20,315 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40119.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:46:20,724 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 08:46:26,529 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40123.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:46:28,701 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.76 vs. limit=5.0 +2023-03-26 08:46:29,510 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.031e+02 1.580e+02 2.018e+02 2.508e+02 5.130e+02, threshold=4.036e+02, percent-clipped=1.0 +2023-03-26 08:46:41,221 INFO [finetune.py:976] (4/7) Epoch 8, batch 50, loss[loss=0.213, simple_loss=0.2712, pruned_loss=0.07744, over 4904.00 frames. ], tot_loss[loss=0.2122, simple_loss=0.2742, pruned_loss=0.07509, over 215184.47 frames. ], batch size: 36, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:47:08,173 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40167.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:47:09,484 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9725, 2.0568, 2.0230, 1.4367, 2.2394, 2.2467, 2.1382, 1.7581], + device='cuda:4'), covar=tensor([0.0598, 0.0615, 0.0761, 0.0929, 0.0502, 0.0612, 0.0636, 0.1102], + device='cuda:4'), in_proj_covar=tensor([0.0136, 0.0133, 0.0143, 0.0126, 0.0113, 0.0144, 0.0145, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:47:10,682 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40171.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:47:26,490 INFO [finetune.py:976] (4/7) Epoch 8, batch 100, loss[loss=0.2278, simple_loss=0.2948, pruned_loss=0.08041, over 4927.00 frames. ], tot_loss[loss=0.2091, simple_loss=0.27, pruned_loss=0.07406, over 380512.11 frames. ], batch size: 33, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:47:27,197 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40195.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:47:32,206 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0407, 1.9389, 1.6258, 2.0527, 2.0357, 1.7442, 2.4137, 2.1203], + device='cuda:4'), covar=tensor([0.1594, 0.3047, 0.3638, 0.3215, 0.2824, 0.1828, 0.3935, 0.2103], + device='cuda:4'), in_proj_covar=tensor([0.0173, 0.0191, 0.0236, 0.0256, 0.0236, 0.0194, 0.0212, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:47:42,316 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40218.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:47:48,309 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.570e+02 1.832e+02 2.394e+02 3.868e+02, threshold=3.663e+02, percent-clipped=0.0 +2023-03-26 08:47:59,305 INFO [finetune.py:976] (4/7) Epoch 8, batch 150, loss[loss=0.1831, simple_loss=0.2492, pruned_loss=0.05849, over 4767.00 frames. ], tot_loss[loss=0.2023, simple_loss=0.263, pruned_loss=0.07078, over 507254.17 frames. ], batch size: 26, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:48:07,720 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40256.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:48:18,415 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7057, 1.4795, 1.1264, 0.3113, 1.2446, 1.5092, 1.4429, 1.4577], + device='cuda:4'), covar=tensor([0.1040, 0.0802, 0.1268, 0.1934, 0.1398, 0.2406, 0.2229, 0.0904], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0201, 0.0203, 0.0188, 0.0219, 0.0207, 0.0224, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:48:18,418 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40272.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:48:22,657 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40279.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:48:33,051 INFO [finetune.py:976] (4/7) Epoch 8, batch 200, loss[loss=0.1708, simple_loss=0.2429, pruned_loss=0.04937, over 4897.00 frames. ], tot_loss[loss=0.2025, simple_loss=0.263, pruned_loss=0.07103, over 605790.90 frames. ], batch size: 43, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:48:33,764 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40295.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 08:48:55,743 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.652e+02 1.957e+02 2.371e+02 3.958e+02, threshold=3.914e+02, percent-clipped=3.0 +2023-03-26 08:48:57,113 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40330.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:48:58,976 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40333.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:49:05,955 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40343.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 08:49:06,468 INFO [finetune.py:976] (4/7) Epoch 8, batch 250, loss[loss=0.2214, simple_loss=0.2777, pruned_loss=0.08254, over 4816.00 frames. ], tot_loss[loss=0.2052, simple_loss=0.2664, pruned_loss=0.07203, over 680891.70 frames. ], batch size: 30, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:49:37,959 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40391.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:49:40,046 INFO [finetune.py:976] (4/7) Epoch 8, batch 300, loss[loss=0.2605, simple_loss=0.3029, pruned_loss=0.1091, over 4758.00 frames. ], tot_loss[loss=0.2083, simple_loss=0.2706, pruned_loss=0.07296, over 741365.74 frames. ], batch size: 27, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:49:59,864 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40423.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:50:07,993 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.190e+02 1.680e+02 2.022e+02 2.440e+02 4.521e+02, threshold=4.043e+02, percent-clipped=1.0 +2023-03-26 08:50:27,980 INFO [finetune.py:976] (4/7) Epoch 8, batch 350, loss[loss=0.1864, simple_loss=0.27, pruned_loss=0.05145, over 4862.00 frames. ], tot_loss[loss=0.2092, simple_loss=0.272, pruned_loss=0.07317, over 790844.92 frames. ], batch size: 34, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:51:00,917 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.90 vs. limit=5.0 +2023-03-26 08:51:01,400 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40471.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:51:01,450 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40471.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:51:26,233 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40492.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:51:27,354 INFO [finetune.py:976] (4/7) Epoch 8, batch 400, loss[loss=0.227, simple_loss=0.2858, pruned_loss=0.08408, over 4740.00 frames. ], tot_loss[loss=0.2111, simple_loss=0.274, pruned_loss=0.07412, over 826457.80 frames. ], batch size: 54, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:51:36,071 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8587, 1.7352, 1.5961, 1.8814, 2.2464, 1.8922, 1.4277, 1.4929], + device='cuda:4'), covar=tensor([0.1940, 0.1916, 0.1666, 0.1517, 0.1659, 0.1129, 0.2599, 0.1726], + device='cuda:4'), in_proj_covar=tensor([0.0237, 0.0210, 0.0204, 0.0188, 0.0240, 0.0179, 0.0215, 0.0193], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:51:52,932 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40519.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:51:58,842 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.662e+02 2.008e+02 2.590e+02 4.107e+02, threshold=4.016e+02, percent-clipped=2.0 +2023-03-26 08:52:11,109 INFO [finetune.py:976] (4/7) Epoch 8, batch 450, loss[loss=0.2264, simple_loss=0.2817, pruned_loss=0.08552, over 4822.00 frames. ], tot_loss[loss=0.2095, simple_loss=0.2723, pruned_loss=0.07338, over 854433.80 frames. ], batch size: 33, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:52:21,171 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40551.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:52:26,847 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40553.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:52:41,175 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40574.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:52:46,640 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7886, 1.7148, 1.5761, 1.8414, 2.1172, 1.8718, 1.3575, 1.4647], + device='cuda:4'), covar=tensor([0.2045, 0.2027, 0.1788, 0.1564, 0.1833, 0.1177, 0.2716, 0.1781], + device='cuda:4'), in_proj_covar=tensor([0.0238, 0.0210, 0.0205, 0.0188, 0.0240, 0.0179, 0.0215, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:52:53,146 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2230, 2.4317, 2.0639, 1.8547, 2.0331, 2.5545, 2.4026, 1.9723], + device='cuda:4'), covar=tensor([0.0519, 0.0487, 0.0811, 0.0797, 0.1353, 0.0507, 0.0515, 0.0923], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0131, 0.0141, 0.0124, 0.0112, 0.0142, 0.0143, 0.0158], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:52:54,279 INFO [finetune.py:976] (4/7) Epoch 8, batch 500, loss[loss=0.1828, simple_loss=0.2336, pruned_loss=0.066, over 4489.00 frames. ], tot_loss[loss=0.2069, simple_loss=0.269, pruned_loss=0.07237, over 876096.62 frames. ], batch size: 19, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:53:17,854 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.064e+02 1.648e+02 1.946e+02 2.379e+02 4.476e+02, threshold=3.892e+02, percent-clipped=1.0 +2023-03-26 08:53:17,931 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40628.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:53:28,115 INFO [finetune.py:976] (4/7) Epoch 8, batch 550, loss[loss=0.195, simple_loss=0.2444, pruned_loss=0.07281, over 4816.00 frames. ], tot_loss[loss=0.2033, simple_loss=0.265, pruned_loss=0.07079, over 895011.97 frames. ], batch size: 25, lr: 3.85e-03, grad_scale: 16.0 +2023-03-26 08:53:32,519 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40651.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:53:43,210 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3292, 2.2999, 2.2118, 2.3788, 2.9545, 2.3683, 2.2074, 1.7840], + device='cuda:4'), covar=tensor([0.2298, 0.2066, 0.1875, 0.1725, 0.1793, 0.1066, 0.2338, 0.1910], + device='cuda:4'), in_proj_covar=tensor([0.0237, 0.0209, 0.0204, 0.0188, 0.0239, 0.0178, 0.0214, 0.0193], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:53:56,758 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40686.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:54:01,543 INFO [finetune.py:976] (4/7) Epoch 8, batch 600, loss[loss=0.2333, simple_loss=0.2915, pruned_loss=0.08755, over 4760.00 frames. ], tot_loss[loss=0.2042, simple_loss=0.266, pruned_loss=0.07125, over 907293.18 frames. ], batch size: 54, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 08:54:14,589 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40712.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:54:15,216 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 08:54:24,591 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.141e+02 1.757e+02 2.080e+02 2.524e+02 4.426e+02, threshold=4.160e+02, percent-clipped=1.0 +2023-03-26 08:54:34,713 INFO [finetune.py:976] (4/7) Epoch 8, batch 650, loss[loss=0.1803, simple_loss=0.2797, pruned_loss=0.04047, over 4817.00 frames. ], tot_loss[loss=0.2081, simple_loss=0.2708, pruned_loss=0.07275, over 918237.09 frames. ], batch size: 39, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 08:55:08,422 INFO [finetune.py:976] (4/7) Epoch 8, batch 700, loss[loss=0.1973, simple_loss=0.2721, pruned_loss=0.06122, over 4237.00 frames. ], tot_loss[loss=0.2083, simple_loss=0.2718, pruned_loss=0.07244, over 925396.48 frames. ], batch size: 65, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 08:55:14,574 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.8398, 3.3684, 3.5308, 3.7129, 3.5624, 3.4452, 3.9339, 1.2152], + device='cuda:4'), covar=tensor([0.0914, 0.0858, 0.0825, 0.1048, 0.1488, 0.1497, 0.0881, 0.5081], + device='cuda:4'), in_proj_covar=tensor([0.0346, 0.0240, 0.0271, 0.0288, 0.0329, 0.0278, 0.0299, 0.0292], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:55:16,454 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3198, 1.8685, 2.1669, 2.1935, 1.9100, 1.9440, 2.1309, 1.9965], + device='cuda:4'), covar=tensor([0.4722, 0.5770, 0.4687, 0.5489, 0.6554, 0.4699, 0.6723, 0.4326], + device='cuda:4'), in_proj_covar=tensor([0.0230, 0.0242, 0.0254, 0.0254, 0.0245, 0.0222, 0.0272, 0.0226], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:55:31,876 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.288e+02 1.702e+02 1.948e+02 2.422e+02 4.930e+02, threshold=3.896e+02, percent-clipped=3.0 +2023-03-26 08:55:38,373 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.0628, 3.5109, 3.6708, 3.8954, 3.8276, 3.5888, 4.1432, 1.3932], + device='cuda:4'), covar=tensor([0.0723, 0.0723, 0.0828, 0.0888, 0.1180, 0.1472, 0.0700, 0.4863], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0241, 0.0273, 0.0290, 0.0330, 0.0279, 0.0301, 0.0294], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:55:51,726 INFO [finetune.py:976] (4/7) Epoch 8, batch 750, loss[loss=0.2544, simple_loss=0.3158, pruned_loss=0.09654, over 4759.00 frames. ], tot_loss[loss=0.2101, simple_loss=0.2735, pruned_loss=0.07333, over 931391.84 frames. ], batch size: 59, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 08:55:54,208 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40848.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:56:01,141 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40851.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:56:20,364 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40865.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:56:22,227 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6682, 1.2388, 0.8120, 1.5424, 2.0623, 1.2006, 1.4838, 1.6308], + device='cuda:4'), covar=tensor([0.1291, 0.1973, 0.1985, 0.1147, 0.1928, 0.2083, 0.1384, 0.1720], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0096, 0.0114, 0.0093, 0.0123, 0.0096, 0.0100, 0.0093], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 08:56:30,880 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2041, 3.5967, 3.8336, 4.0372, 3.9473, 3.7232, 4.2817, 1.4742], + device='cuda:4'), covar=tensor([0.0695, 0.0737, 0.0702, 0.0779, 0.1079, 0.1371, 0.0643, 0.5049], + device='cuda:4'), in_proj_covar=tensor([0.0348, 0.0241, 0.0272, 0.0291, 0.0330, 0.0279, 0.0301, 0.0294], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:56:32,110 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40874.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:56:55,830 INFO [finetune.py:976] (4/7) Epoch 8, batch 800, loss[loss=0.2268, simple_loss=0.2742, pruned_loss=0.08971, over 4296.00 frames. ], tot_loss[loss=0.2088, simple_loss=0.2723, pruned_loss=0.07262, over 936228.51 frames. ], batch size: 66, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 08:57:03,993 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40899.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:57:04,653 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40900.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:57:24,032 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40922.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:57:26,528 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40926.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:57:27,576 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.606e+02 1.985e+02 2.397e+02 9.945e+02, threshold=3.971e+02, percent-clipped=3.0 +2023-03-26 08:57:27,687 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40928.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:57:45,605 INFO [finetune.py:976] (4/7) Epoch 8, batch 850, loss[loss=0.1705, simple_loss=0.2252, pruned_loss=0.05788, over 4734.00 frames. ], tot_loss[loss=0.2068, simple_loss=0.2699, pruned_loss=0.07188, over 940147.89 frames. ], batch size: 23, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 08:58:00,368 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40961.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 08:58:10,969 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=40976.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:58:18,121 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40986.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:58:22,821 INFO [finetune.py:976] (4/7) Epoch 8, batch 900, loss[loss=0.2325, simple_loss=0.2827, pruned_loss=0.09113, over 4856.00 frames. ], tot_loss[loss=0.2061, simple_loss=0.2682, pruned_loss=0.07196, over 942603.53 frames. ], batch size: 44, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 08:58:25,205 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40997.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:58:31,500 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41007.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:58:46,167 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.525e+02 1.868e+02 2.283e+02 3.598e+02, threshold=3.736e+02, percent-clipped=0.0 +2023-03-26 08:58:50,351 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41034.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:58:56,855 INFO [finetune.py:976] (4/7) Epoch 8, batch 950, loss[loss=0.1946, simple_loss=0.253, pruned_loss=0.06813, over 4836.00 frames. ], tot_loss[loss=0.2051, simple_loss=0.2661, pruned_loss=0.07201, over 944129.81 frames. ], batch size: 30, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 08:58:57,584 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41045.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:59:06,040 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41058.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:59:26,238 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.58 vs. limit=5.0 +2023-03-26 08:59:30,583 INFO [finetune.py:976] (4/7) Epoch 8, batch 1000, loss[loss=0.2613, simple_loss=0.3252, pruned_loss=0.09867, over 4755.00 frames. ], tot_loss[loss=0.2076, simple_loss=0.2692, pruned_loss=0.07305, over 946993.72 frames. ], batch size: 54, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 08:59:36,019 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41102.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:59:38,374 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41106.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 08:59:40,810 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9122, 1.7486, 1.5531, 1.7497, 1.6497, 1.7220, 1.8114, 2.4600], + device='cuda:4'), covar=tensor([0.5203, 0.5912, 0.4175, 0.5665, 0.5403, 0.3126, 0.5104, 0.2264], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0257, 0.0220, 0.0280, 0.0241, 0.0205, 0.0244, 0.0206], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 08:59:52,969 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.084e+02 1.652e+02 2.000e+02 2.359e+02 4.809e+02, threshold=4.000e+02, percent-clipped=2.0 +2023-03-26 09:00:04,082 INFO [finetune.py:976] (4/7) Epoch 8, batch 1050, loss[loss=0.1718, simple_loss=0.2414, pruned_loss=0.05107, over 4689.00 frames. ], tot_loss[loss=0.2106, simple_loss=0.273, pruned_loss=0.07414, over 950331.51 frames. ], batch size: 23, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:00:06,628 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41148.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:00:13,574 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.74 vs. limit=5.0 +2023-03-26 09:00:16,271 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41163.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:00:27,514 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6092, 3.3970, 3.3369, 1.6174, 3.5426, 2.6375, 1.0127, 2.4317], + device='cuda:4'), covar=tensor([0.2567, 0.1793, 0.1480, 0.2946, 0.1015, 0.1022, 0.3641, 0.1309], + device='cuda:4'), in_proj_covar=tensor([0.0156, 0.0175, 0.0164, 0.0132, 0.0159, 0.0125, 0.0148, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 09:00:37,444 INFO [finetune.py:976] (4/7) Epoch 8, batch 1100, loss[loss=0.1983, simple_loss=0.263, pruned_loss=0.06678, over 4888.00 frames. ], tot_loss[loss=0.2108, simple_loss=0.2734, pruned_loss=0.07412, over 951853.73 frames. ], batch size: 32, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:00:37,758 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-26 09:00:38,741 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41196.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:00:50,655 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6865, 1.5624, 2.2860, 3.4228, 2.5027, 2.3875, 0.9809, 2.7586], + device='cuda:4'), covar=tensor([0.1767, 0.1530, 0.1287, 0.0565, 0.0721, 0.1359, 0.1970, 0.0578], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0117, 0.0135, 0.0165, 0.0101, 0.0139, 0.0127, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 09:00:54,970 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41221.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:00:59,684 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.124e+02 1.750e+02 2.155e+02 2.664e+02 4.791e+02, threshold=4.309e+02, percent-clipped=2.0 +2023-03-26 09:00:59,805 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3625, 1.4117, 1.5979, 1.7463, 1.5407, 3.1006, 1.3328, 1.5874], + device='cuda:4'), covar=tensor([0.0951, 0.1662, 0.1107, 0.0910, 0.1410, 0.0237, 0.1310, 0.1500], + device='cuda:4'), in_proj_covar=tensor([0.0077, 0.0082, 0.0076, 0.0079, 0.0093, 0.0083, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 09:01:17,462 INFO [finetune.py:976] (4/7) Epoch 8, batch 1150, loss[loss=0.1967, simple_loss=0.2686, pruned_loss=0.06235, over 4839.00 frames. ], tot_loss[loss=0.2119, simple_loss=0.2742, pruned_loss=0.07486, over 951502.40 frames. ], batch size: 30, lr: 3.84e-03, grad_scale: 32.0 +2023-03-26 09:01:32,471 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41256.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 09:01:43,715 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0565, 1.9359, 1.7443, 2.0450, 1.3053, 4.5723, 1.5901, 2.3858], + device='cuda:4'), covar=tensor([0.3118, 0.2343, 0.2066, 0.2156, 0.1796, 0.0114, 0.2398, 0.1262], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0116, 0.0120, 0.0123, 0.0117, 0.0099, 0.0101, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 09:02:12,805 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1910, 1.1986, 1.1027, 1.3192, 1.5228, 1.3938, 1.2269, 1.0845], + device='cuda:4'), covar=tensor([0.0330, 0.0270, 0.0535, 0.0244, 0.0212, 0.0350, 0.0288, 0.0386], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0112, 0.0141, 0.0117, 0.0105, 0.0101, 0.0092, 0.0110], + device='cuda:4'), out_proj_covar=tensor([7.0565e-05, 8.8272e-05, 1.1240e-04, 9.2144e-05, 8.2912e-05, 7.5243e-05, + 6.9284e-05, 8.5757e-05], device='cuda:4') +2023-03-26 09:02:15,112 INFO [finetune.py:976] (4/7) Epoch 8, batch 1200, loss[loss=0.1781, simple_loss=0.2462, pruned_loss=0.05502, over 4764.00 frames. ], tot_loss[loss=0.2092, simple_loss=0.2717, pruned_loss=0.07341, over 953361.98 frames. ], batch size: 26, lr: 3.84e-03, grad_scale: 32.0 +2023-03-26 09:02:18,746 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.5158, 4.7191, 5.0706, 5.3133, 5.2380, 4.9726, 5.5861, 1.7522], + device='cuda:4'), covar=tensor([0.0592, 0.0707, 0.0666, 0.0747, 0.0983, 0.1231, 0.0505, 0.5177], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0241, 0.0273, 0.0290, 0.0330, 0.0279, 0.0301, 0.0294], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:02:24,086 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41307.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:02:37,270 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.297e+01 1.635e+02 1.914e+02 2.289e+02 4.123e+02, threshold=3.829e+02, percent-clipped=0.0 +2023-03-26 09:02:51,370 INFO [finetune.py:976] (4/7) Epoch 8, batch 1250, loss[loss=0.173, simple_loss=0.2331, pruned_loss=0.05647, over 4800.00 frames. ], tot_loss[loss=0.2059, simple_loss=0.2678, pruned_loss=0.07197, over 952618.48 frames. ], batch size: 51, lr: 3.84e-03, grad_scale: 32.0 +2023-03-26 09:03:02,691 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41353.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:03:04,421 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41355.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:03:32,788 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.85 vs. limit=5.0 +2023-03-26 09:03:32,985 INFO [finetune.py:976] (4/7) Epoch 8, batch 1300, loss[loss=0.231, simple_loss=0.2932, pruned_loss=0.08442, over 4832.00 frames. ], tot_loss[loss=0.2039, simple_loss=0.2653, pruned_loss=0.07123, over 953717.33 frames. ], batch size: 47, lr: 3.84e-03, grad_scale: 32.0 +2023-03-26 09:03:37,903 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41401.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:03:56,247 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.077e+02 1.674e+02 1.900e+02 2.309e+02 4.379e+02, threshold=3.799e+02, percent-clipped=1.0 +2023-03-26 09:04:06,255 INFO [finetune.py:976] (4/7) Epoch 8, batch 1350, loss[loss=0.2377, simple_loss=0.2885, pruned_loss=0.09342, over 4899.00 frames. ], tot_loss[loss=0.205, simple_loss=0.2663, pruned_loss=0.07188, over 953793.58 frames. ], batch size: 32, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:04:16,297 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41458.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:04:30,682 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.75 vs. limit=5.0 +2023-03-26 09:04:39,878 INFO [finetune.py:976] (4/7) Epoch 8, batch 1400, loss[loss=0.2357, simple_loss=0.288, pruned_loss=0.09166, over 4825.00 frames. ], tot_loss[loss=0.2092, simple_loss=0.2712, pruned_loss=0.07358, over 952403.21 frames. ], batch size: 39, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:04:58,602 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41521.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:05:03,329 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.146e+02 1.704e+02 2.004e+02 2.444e+02 3.700e+02, threshold=4.008e+02, percent-clipped=0.0 +2023-03-26 09:05:12,597 INFO [finetune.py:976] (4/7) Epoch 8, batch 1450, loss[loss=0.2906, simple_loss=0.3382, pruned_loss=0.1215, over 4796.00 frames. ], tot_loss[loss=0.2109, simple_loss=0.2736, pruned_loss=0.07415, over 953920.05 frames. ], batch size: 51, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:05:21,961 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41556.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 09:05:30,810 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41569.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:05:30,875 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5717, 1.5653, 1.5155, 1.8489, 2.0417, 1.7358, 1.4067, 1.3757], + device='cuda:4'), covar=tensor([0.2323, 0.2188, 0.1862, 0.1551, 0.1959, 0.1239, 0.2654, 0.1910], + device='cuda:4'), in_proj_covar=tensor([0.0239, 0.0212, 0.0207, 0.0190, 0.0241, 0.0179, 0.0216, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:05:46,316 INFO [finetune.py:976] (4/7) Epoch 8, batch 1500, loss[loss=0.2121, simple_loss=0.2784, pruned_loss=0.07286, over 4847.00 frames. ], tot_loss[loss=0.2117, simple_loss=0.2747, pruned_loss=0.07442, over 955444.43 frames. ], batch size: 44, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:05:47,039 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2557, 1.2955, 1.3177, 1.5473, 1.5334, 2.9155, 1.2765, 1.5082], + device='cuda:4'), covar=tensor([0.1100, 0.1881, 0.1280, 0.1064, 0.1573, 0.0293, 0.1571, 0.1717], + device='cuda:4'), in_proj_covar=tensor([0.0077, 0.0082, 0.0076, 0.0079, 0.0093, 0.0083, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 09:05:53,544 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41604.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:05:58,238 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2095, 2.0459, 1.7851, 2.1180, 2.0386, 1.9390, 1.9058, 2.8773], + device='cuda:4'), covar=tensor([0.5306, 0.6760, 0.4576, 0.6197, 0.5519, 0.3259, 0.6242, 0.2075], + device='cuda:4'), in_proj_covar=tensor([0.0284, 0.0256, 0.0220, 0.0279, 0.0240, 0.0204, 0.0243, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:06:05,678 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.4333, 3.7889, 4.0184, 4.2978, 4.2093, 3.9045, 4.4715, 1.4050], + device='cuda:4'), covar=tensor([0.0638, 0.0690, 0.0850, 0.0754, 0.0974, 0.1435, 0.0589, 0.5046], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0243, 0.0275, 0.0293, 0.0332, 0.0282, 0.0302, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:06:10,825 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.231e+02 1.635e+02 1.924e+02 2.365e+02 3.634e+02, threshold=3.848e+02, percent-clipped=0.0 +2023-03-26 09:06:22,436 INFO [finetune.py:976] (4/7) Epoch 8, batch 1550, loss[loss=0.2379, simple_loss=0.2966, pruned_loss=0.08959, over 4823.00 frames. ], tot_loss[loss=0.2109, simple_loss=0.2743, pruned_loss=0.07379, over 955096.51 frames. ], batch size: 47, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:06:34,440 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41653.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:06:35,066 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41654.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:07:19,721 INFO [finetune.py:976] (4/7) Epoch 8, batch 1600, loss[loss=0.1708, simple_loss=0.2476, pruned_loss=0.047, over 4760.00 frames. ], tot_loss[loss=0.2083, simple_loss=0.271, pruned_loss=0.07275, over 955512.55 frames. ], batch size: 27, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:07:25,548 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41701.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:07:25,574 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41701.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:07:39,587 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41715.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 09:07:48,911 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.581e+02 1.949e+02 2.490e+02 4.755e+02, threshold=3.899e+02, percent-clipped=2.0 +2023-03-26 09:07:58,440 INFO [finetune.py:976] (4/7) Epoch 8, batch 1650, loss[loss=0.2291, simple_loss=0.2791, pruned_loss=0.0895, over 4914.00 frames. ], tot_loss[loss=0.2042, simple_loss=0.2667, pruned_loss=0.07082, over 952445.17 frames. ], batch size: 36, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:08:01,534 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41749.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:08:09,852 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41758.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:08:31,174 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 09:08:32,663 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4754, 2.3218, 2.0467, 2.5098, 2.4902, 2.1880, 2.8231, 2.4281], + device='cuda:4'), covar=tensor([0.1583, 0.2526, 0.3422, 0.2766, 0.2595, 0.1748, 0.2965, 0.2169], + device='cuda:4'), in_proj_covar=tensor([0.0173, 0.0190, 0.0235, 0.0255, 0.0236, 0.0194, 0.0212, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:08:42,690 INFO [finetune.py:976] (4/7) Epoch 8, batch 1700, loss[loss=0.2846, simple_loss=0.3307, pruned_loss=0.1192, over 4047.00 frames. ], tot_loss[loss=0.2019, simple_loss=0.264, pruned_loss=0.06988, over 951586.64 frames. ], batch size: 65, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:08:47,918 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5401, 1.4426, 1.3402, 1.4195, 0.9126, 3.1544, 1.1754, 1.8044], + device='cuda:4'), covar=tensor([0.3457, 0.2470, 0.2321, 0.2555, 0.2257, 0.0235, 0.2964, 0.1353], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0114, 0.0119, 0.0123, 0.0117, 0.0098, 0.0100, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 09:08:50,309 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=41806.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:08:50,377 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5437, 2.3455, 2.0066, 0.9502, 2.0585, 1.8715, 1.7536, 2.0353], + device='cuda:4'), covar=tensor([0.0842, 0.0906, 0.1405, 0.2175, 0.1508, 0.2535, 0.2245, 0.1083], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0201, 0.0204, 0.0189, 0.0220, 0.0208, 0.0224, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:08:57,315 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8055, 1.2832, 0.7582, 1.6753, 2.0690, 1.4795, 1.5595, 1.5533], + device='cuda:4'), covar=tensor([0.1352, 0.2083, 0.2172, 0.1227, 0.1989, 0.2021, 0.1365, 0.1938], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0097, 0.0114, 0.0092, 0.0123, 0.0096, 0.0101, 0.0092], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 09:09:06,693 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.178e+02 1.763e+02 2.034e+02 2.335e+02 4.675e+02, threshold=4.069e+02, percent-clipped=2.0 +2023-03-26 09:09:16,759 INFO [finetune.py:976] (4/7) Epoch 8, batch 1750, loss[loss=0.191, simple_loss=0.2574, pruned_loss=0.06229, over 4901.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.2658, pruned_loss=0.0705, over 954270.95 frames. ], batch size: 35, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:09:18,129 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4423, 2.3392, 2.0763, 2.3954, 2.1549, 4.9878, 2.2747, 3.0779], + device='cuda:4'), covar=tensor([0.2848, 0.2030, 0.1825, 0.2001, 0.1419, 0.0109, 0.2132, 0.0931], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0115, 0.0119, 0.0123, 0.0117, 0.0098, 0.0101, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 09:09:47,788 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2210, 2.0003, 1.5508, 0.6435, 1.7201, 1.8712, 1.6819, 1.7954], + device='cuda:4'), covar=tensor([0.1085, 0.0897, 0.1592, 0.2355, 0.1555, 0.2706, 0.2909, 0.0998], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0200, 0.0203, 0.0188, 0.0218, 0.0207, 0.0222, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:09:50,598 INFO [finetune.py:976] (4/7) Epoch 8, batch 1800, loss[loss=0.179, simple_loss=0.2685, pruned_loss=0.04473, over 4900.00 frames. ], tot_loss[loss=0.2055, simple_loss=0.2684, pruned_loss=0.0713, over 953363.94 frames. ], batch size: 37, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:09:58,005 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41906.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:09:58,587 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.3206, 3.7087, 3.9244, 4.1461, 4.0527, 3.7749, 4.3689, 1.3371], + device='cuda:4'), covar=tensor([0.0736, 0.0773, 0.0766, 0.0869, 0.1128, 0.1306, 0.0586, 0.5291], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0243, 0.0276, 0.0294, 0.0335, 0.0284, 0.0303, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:10:13,608 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.102e+02 1.792e+02 2.103e+02 2.633e+02 4.479e+02, threshold=4.207e+02, percent-clipped=2.0 +2023-03-26 09:10:23,654 INFO [finetune.py:976] (4/7) Epoch 8, batch 1850, loss[loss=0.1724, simple_loss=0.2491, pruned_loss=0.04783, over 4814.00 frames. ], tot_loss[loss=0.2092, simple_loss=0.2721, pruned_loss=0.07313, over 954457.38 frames. ], batch size: 39, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:10:26,680 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41948.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 09:10:38,695 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41967.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:10:57,330 INFO [finetune.py:976] (4/7) Epoch 8, batch 1900, loss[loss=0.1857, simple_loss=0.2606, pruned_loss=0.05541, over 4906.00 frames. ], tot_loss[loss=0.2098, simple_loss=0.2732, pruned_loss=0.07321, over 954790.75 frames. ], batch size: 36, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:11:08,256 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42009.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 09:11:08,451 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-26 09:11:08,800 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42010.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 09:11:21,639 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6018, 1.5383, 1.4284, 1.6036, 1.1122, 3.5126, 1.3697, 1.9439], + device='cuda:4'), covar=tensor([0.3750, 0.2684, 0.2329, 0.2394, 0.1952, 0.0195, 0.2723, 0.1328], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0115, 0.0119, 0.0123, 0.0117, 0.0099, 0.0100, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 09:11:22,119 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.207e+02 1.556e+02 1.920e+02 2.218e+02 3.872e+02, threshold=3.841e+02, percent-clipped=0.0 +2023-03-26 09:11:26,920 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4837, 2.2389, 2.8783, 1.7537, 2.7126, 2.7544, 2.1502, 2.8946], + device='cuda:4'), covar=tensor([0.1424, 0.1929, 0.1400, 0.2536, 0.0868, 0.1469, 0.2389, 0.0878], + device='cuda:4'), in_proj_covar=tensor([0.0199, 0.0202, 0.0196, 0.0195, 0.0179, 0.0218, 0.0217, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:11:32,121 INFO [finetune.py:976] (4/7) Epoch 8, batch 1950, loss[loss=0.2006, simple_loss=0.2556, pruned_loss=0.07279, over 4817.00 frames. ], tot_loss[loss=0.2068, simple_loss=0.2706, pruned_loss=0.07148, over 956115.52 frames. ], batch size: 38, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:12:30,959 INFO [finetune.py:976] (4/7) Epoch 8, batch 2000, loss[loss=0.2031, simple_loss=0.2677, pruned_loss=0.06923, over 4900.00 frames. ], tot_loss[loss=0.2041, simple_loss=0.2671, pruned_loss=0.07059, over 955566.37 frames. ], batch size: 36, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:12:56,438 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.048e+02 1.506e+02 1.840e+02 2.176e+02 3.856e+02, threshold=3.679e+02, percent-clipped=1.0 +2023-03-26 09:13:01,518 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-26 09:13:06,569 INFO [finetune.py:976] (4/7) Epoch 8, batch 2050, loss[loss=0.1805, simple_loss=0.2505, pruned_loss=0.05523, over 4851.00 frames. ], tot_loss[loss=0.2011, simple_loss=0.2637, pruned_loss=0.06922, over 956994.25 frames. ], batch size: 44, lr: 3.84e-03, grad_scale: 16.0 +2023-03-26 09:13:53,279 INFO [finetune.py:976] (4/7) Epoch 8, batch 2100, loss[loss=0.1615, simple_loss=0.24, pruned_loss=0.04156, over 4807.00 frames. ], tot_loss[loss=0.2014, simple_loss=0.2636, pruned_loss=0.06956, over 958919.80 frames. ], batch size: 25, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:14:03,809 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.43 vs. limit=5.0 +2023-03-26 09:14:16,280 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.710e+02 1.945e+02 2.376e+02 4.149e+02, threshold=3.889e+02, percent-clipped=2.0 +2023-03-26 09:14:26,997 INFO [finetune.py:976] (4/7) Epoch 8, batch 2150, loss[loss=0.2322, simple_loss=0.3029, pruned_loss=0.08074, over 4828.00 frames. ], tot_loss[loss=0.2036, simple_loss=0.2662, pruned_loss=0.07055, over 956686.83 frames. ], batch size: 47, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:14:38,894 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42262.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:14:49,899 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3590, 1.5455, 1.6833, 0.8673, 1.5985, 1.7614, 1.8303, 1.5037], + device='cuda:4'), covar=tensor([0.0884, 0.0554, 0.0381, 0.0556, 0.0432, 0.0733, 0.0316, 0.0563], + device='cuda:4'), in_proj_covar=tensor([0.0128, 0.0155, 0.0120, 0.0135, 0.0131, 0.0125, 0.0144, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.5425e-05, 1.1415e-04, 8.6627e-05, 9.8304e-05, 9.3754e-05, 9.1486e-05, + 1.0600e-04, 1.0758e-04], device='cuda:4') +2023-03-26 09:14:59,237 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.60 vs. limit=5.0 +2023-03-26 09:15:18,053 INFO [finetune.py:976] (4/7) Epoch 8, batch 2200, loss[loss=0.2283, simple_loss=0.294, pruned_loss=0.0813, over 4819.00 frames. ], tot_loss[loss=0.2074, simple_loss=0.2701, pruned_loss=0.07238, over 956484.16 frames. ], batch size: 38, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:15:25,342 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42304.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 09:15:29,043 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=42310.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 09:15:45,963 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.121e+02 1.543e+02 1.921e+02 2.479e+02 5.347e+02, threshold=3.843e+02, percent-clipped=1.0 +2023-03-26 09:16:07,395 INFO [finetune.py:976] (4/7) Epoch 8, batch 2250, loss[loss=0.1732, simple_loss=0.2454, pruned_loss=0.05053, over 4891.00 frames. ], tot_loss[loss=0.2103, simple_loss=0.2731, pruned_loss=0.07377, over 955106.97 frames. ], batch size: 35, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:16:27,423 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=42358.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:16:46,660 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8436, 1.7885, 1.7694, 1.0432, 1.8868, 1.9469, 1.8181, 1.5493], + device='cuda:4'), covar=tensor([0.0556, 0.0697, 0.0712, 0.0984, 0.0571, 0.0659, 0.0603, 0.1075], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0132, 0.0143, 0.0126, 0.0113, 0.0143, 0.0144, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:16:47,895 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42379.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:17:09,028 INFO [finetune.py:976] (4/7) Epoch 8, batch 2300, loss[loss=0.1836, simple_loss=0.2489, pruned_loss=0.05919, over 4753.00 frames. ], tot_loss[loss=0.2101, simple_loss=0.2733, pruned_loss=0.07344, over 956953.38 frames. ], batch size: 23, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:17:57,226 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.935e+01 1.473e+02 1.816e+02 2.175e+02 3.275e+02, threshold=3.633e+02, percent-clipped=0.0 +2023-03-26 09:18:06,506 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42440.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:18:09,263 INFO [finetune.py:976] (4/7) Epoch 8, batch 2350, loss[loss=0.2304, simple_loss=0.2936, pruned_loss=0.08359, over 4824.00 frames. ], tot_loss[loss=0.206, simple_loss=0.2694, pruned_loss=0.07135, over 957163.55 frames. ], batch size: 39, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:18:51,871 INFO [finetune.py:976] (4/7) Epoch 8, batch 2400, loss[loss=0.1801, simple_loss=0.2415, pruned_loss=0.05933, over 4819.00 frames. ], tot_loss[loss=0.2039, simple_loss=0.2666, pruned_loss=0.07054, over 955214.60 frames. ], batch size: 39, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:19:02,399 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42506.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:19:25,681 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.517e+02 1.798e+02 2.223e+02 5.682e+02, threshold=3.597e+02, percent-clipped=2.0 +2023-03-26 09:19:35,403 INFO [finetune.py:976] (4/7) Epoch 8, batch 2450, loss[loss=0.1566, simple_loss=0.2206, pruned_loss=0.04635, over 4758.00 frames. ], tot_loss[loss=0.2013, simple_loss=0.2635, pruned_loss=0.06956, over 955722.70 frames. ], batch size: 27, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:19:44,411 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.38 vs. limit=5.0 +2023-03-26 09:19:47,984 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=42562.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:19:51,987 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42567.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:20:08,913 INFO [finetune.py:976] (4/7) Epoch 8, batch 2500, loss[loss=0.1952, simple_loss=0.2395, pruned_loss=0.07545, over 4676.00 frames. ], tot_loss[loss=0.2007, simple_loss=0.2636, pruned_loss=0.06894, over 956494.85 frames. ], batch size: 23, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:20:16,185 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=42604.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 09:20:20,730 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=42610.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:20:33,719 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.056e+02 1.709e+02 1.979e+02 2.316e+02 5.134e+02, threshold=3.959e+02, percent-clipped=4.0 +2023-03-26 09:20:42,883 INFO [finetune.py:976] (4/7) Epoch 8, batch 2550, loss[loss=0.2104, simple_loss=0.2761, pruned_loss=0.07235, over 4906.00 frames. ], tot_loss[loss=0.2053, simple_loss=0.2687, pruned_loss=0.07097, over 956197.56 frames. ], batch size: 35, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:20:46,540 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42649.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:20:53,417 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=42652.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 09:21:22,990 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4377, 1.3558, 1.4103, 0.7910, 1.5551, 1.5076, 1.4439, 1.2460], + device='cuda:4'), covar=tensor([0.0605, 0.0721, 0.0720, 0.0952, 0.0695, 0.0712, 0.0631, 0.1207], + device='cuda:4'), in_proj_covar=tensor([0.0135, 0.0133, 0.0144, 0.0126, 0.0114, 0.0144, 0.0144, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:21:25,356 INFO [finetune.py:976] (4/7) Epoch 8, batch 2600, loss[loss=0.1748, simple_loss=0.2491, pruned_loss=0.05024, over 4740.00 frames. ], tot_loss[loss=0.2076, simple_loss=0.2709, pruned_loss=0.07215, over 954076.49 frames. ], batch size: 27, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:21:36,667 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42710.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:21:39,891 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-26 09:21:49,602 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.153e+02 1.771e+02 2.168e+02 2.787e+02 4.495e+02, threshold=4.337e+02, percent-clipped=4.0 +2023-03-26 09:21:53,824 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42735.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:21:54,486 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4347, 1.5036, 1.3558, 1.3935, 1.7489, 1.7089, 1.5159, 1.3449], + device='cuda:4'), covar=tensor([0.0343, 0.0275, 0.0486, 0.0271, 0.0205, 0.0421, 0.0260, 0.0354], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0112, 0.0140, 0.0116, 0.0105, 0.0102, 0.0092, 0.0110], + device='cuda:4'), out_proj_covar=tensor([7.0625e-05, 8.7823e-05, 1.1233e-04, 9.1601e-05, 8.2565e-05, 7.5564e-05, + 6.9261e-05, 8.5333e-05], device='cuda:4') +2023-03-26 09:21:56,271 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7810, 1.7314, 1.7572, 1.1186, 1.8479, 1.8712, 1.7987, 1.5090], + device='cuda:4'), covar=tensor([0.0601, 0.0651, 0.0735, 0.0960, 0.0594, 0.0706, 0.0640, 0.1058], + device='cuda:4'), in_proj_covar=tensor([0.0135, 0.0133, 0.0144, 0.0126, 0.0114, 0.0144, 0.0144, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:21:57,896 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-03-26 09:21:59,147 INFO [finetune.py:976] (4/7) Epoch 8, batch 2650, loss[loss=0.1636, simple_loss=0.2195, pruned_loss=0.05385, over 4766.00 frames. ], tot_loss[loss=0.2075, simple_loss=0.2711, pruned_loss=0.07199, over 950855.18 frames. ], batch size: 27, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:22:17,206 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 09:22:20,035 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.64 vs. limit=2.0 +2023-03-26 09:22:40,399 INFO [finetune.py:976] (4/7) Epoch 8, batch 2700, loss[loss=0.2043, simple_loss=0.2524, pruned_loss=0.07815, over 4829.00 frames. ], tot_loss[loss=0.2075, simple_loss=0.2705, pruned_loss=0.07222, over 950933.50 frames. ], batch size: 30, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:22:56,979 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 09:23:07,282 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 09:23:27,702 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.159e+02 1.604e+02 1.897e+02 2.218e+02 3.599e+02, threshold=3.793e+02, percent-clipped=0.0 +2023-03-26 09:23:46,886 INFO [finetune.py:976] (4/7) Epoch 8, batch 2750, loss[loss=0.1825, simple_loss=0.2252, pruned_loss=0.06988, over 4174.00 frames. ], tot_loss[loss=0.2043, simple_loss=0.2666, pruned_loss=0.07096, over 950767.52 frames. ], batch size: 18, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:23:51,809 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5607, 1.4677, 1.9732, 3.1639, 2.1704, 2.3987, 0.8756, 2.5664], + device='cuda:4'), covar=tensor([0.1745, 0.1489, 0.1265, 0.0585, 0.0818, 0.1304, 0.1947, 0.0554], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0116, 0.0133, 0.0165, 0.0102, 0.0138, 0.0126, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 09:23:55,957 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42859.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:24:02,855 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42862.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:24:14,741 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42879.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:24:23,351 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7676, 1.8079, 1.7872, 1.0790, 1.9811, 1.9365, 1.8522, 1.6895], + device='cuda:4'), covar=tensor([0.0653, 0.0724, 0.0771, 0.1063, 0.0563, 0.0749, 0.0676, 0.1012], + device='cuda:4'), in_proj_covar=tensor([0.0135, 0.0133, 0.0144, 0.0125, 0.0114, 0.0144, 0.0144, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:24:30,584 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.4539, 3.8406, 3.9906, 4.2876, 4.2017, 3.9127, 4.5048, 1.3125], + device='cuda:4'), covar=tensor([0.0675, 0.0773, 0.0800, 0.0859, 0.1015, 0.1254, 0.0573, 0.5396], + device='cuda:4'), in_proj_covar=tensor([0.0348, 0.0240, 0.0275, 0.0292, 0.0330, 0.0282, 0.0300, 0.0293], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:24:31,686 INFO [finetune.py:976] (4/7) Epoch 8, batch 2800, loss[loss=0.2117, simple_loss=0.2705, pruned_loss=0.07643, over 4813.00 frames. ], tot_loss[loss=0.2019, simple_loss=0.2639, pruned_loss=0.0699, over 953422.28 frames. ], batch size: 45, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:24:38,763 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 09:24:41,500 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1536, 1.9015, 2.3246, 1.7322, 2.1814, 2.3775, 1.9547, 2.5770], + device='cuda:4'), covar=tensor([0.1183, 0.1888, 0.1422, 0.1690, 0.0941, 0.1173, 0.2473, 0.0669], + device='cuda:4'), in_proj_covar=tensor([0.0204, 0.0206, 0.0199, 0.0198, 0.0184, 0.0223, 0.0221, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:24:48,560 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42920.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:24:54,864 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.137e+02 1.632e+02 1.941e+02 2.379e+02 3.960e+02, threshold=3.882e+02, percent-clipped=2.0 +2023-03-26 09:25:02,624 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42940.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:25:04,921 INFO [finetune.py:976] (4/7) Epoch 8, batch 2850, loss[loss=0.1848, simple_loss=0.242, pruned_loss=0.06377, over 4796.00 frames. ], tot_loss[loss=0.2004, simple_loss=0.2621, pruned_loss=0.06942, over 951639.92 frames. ], batch size: 25, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:25:38,296 INFO [finetune.py:976] (4/7) Epoch 8, batch 2900, loss[loss=0.2015, simple_loss=0.2747, pruned_loss=0.06413, over 4810.00 frames. ], tot_loss[loss=0.2036, simple_loss=0.2652, pruned_loss=0.07098, over 951395.45 frames. ], batch size: 40, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:25:45,629 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43005.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:26:03,401 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.926e+01 1.731e+02 1.970e+02 2.373e+02 5.777e+02, threshold=3.941e+02, percent-clipped=2.0 +2023-03-26 09:26:13,032 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43035.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:26:17,167 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6572, 1.4761, 1.9379, 1.3001, 1.6357, 1.8077, 1.4261, 1.9837], + device='cuda:4'), covar=tensor([0.1212, 0.2017, 0.1325, 0.1862, 0.0928, 0.1395, 0.2711, 0.0776], + device='cuda:4'), in_proj_covar=tensor([0.0202, 0.0205, 0.0197, 0.0197, 0.0183, 0.0221, 0.0219, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:26:18,884 INFO [finetune.py:976] (4/7) Epoch 8, batch 2950, loss[loss=0.1983, simple_loss=0.2663, pruned_loss=0.06513, over 4796.00 frames. ], tot_loss[loss=0.2053, simple_loss=0.2678, pruned_loss=0.07139, over 950094.43 frames. ], batch size: 41, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:26:44,526 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43083.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:26:52,568 INFO [finetune.py:976] (4/7) Epoch 8, batch 3000, loss[loss=0.2167, simple_loss=0.2752, pruned_loss=0.07909, over 4920.00 frames. ], tot_loss[loss=0.2076, simple_loss=0.27, pruned_loss=0.07258, over 950110.08 frames. ], batch size: 33, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:26:52,568 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 09:26:56,422 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0615, 1.8046, 1.7778, 2.1278, 2.2764, 2.0361, 1.3216, 1.8606], + device='cuda:4'), covar=tensor([0.1768, 0.1920, 0.1662, 0.1403, 0.1512, 0.1008, 0.2421, 0.1625], + device='cuda:4'), in_proj_covar=tensor([0.0238, 0.0210, 0.0206, 0.0188, 0.0241, 0.0180, 0.0215, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:26:57,205 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8513, 1.0135, 1.8129, 1.6850, 1.5406, 1.5046, 1.5545, 1.6059], + device='cuda:4'), covar=tensor([0.4562, 0.5408, 0.4588, 0.4951, 0.6120, 0.4482, 0.6251, 0.4270], + device='cuda:4'), in_proj_covar=tensor([0.0231, 0.0241, 0.0253, 0.0254, 0.0245, 0.0222, 0.0273, 0.0227], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:27:10,878 INFO [finetune.py:1010] (4/7) Epoch 8, validation: loss=0.16, simple_loss=0.2311, pruned_loss=0.04446, over 2265189.00 frames. +2023-03-26 09:27:10,878 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6336MB +2023-03-26 09:27:49,850 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.022e+02 1.686e+02 2.049e+02 2.426e+02 3.920e+02, threshold=4.099e+02, percent-clipped=0.0 +2023-03-26 09:28:00,451 INFO [finetune.py:976] (4/7) Epoch 8, batch 3050, loss[loss=0.2249, simple_loss=0.2554, pruned_loss=0.0972, over 4044.00 frames. ], tot_loss[loss=0.2082, simple_loss=0.2707, pruned_loss=0.07281, over 949781.70 frames. ], batch size: 17, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:28:01,744 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4211, 1.4098, 1.2800, 1.4390, 1.6794, 1.6458, 1.4801, 1.2277], + device='cuda:4'), covar=tensor([0.0300, 0.0298, 0.0549, 0.0266, 0.0222, 0.0477, 0.0266, 0.0402], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0111, 0.0140, 0.0116, 0.0104, 0.0101, 0.0091, 0.0110], + device='cuda:4'), out_proj_covar=tensor([7.0728e-05, 8.7395e-05, 1.1220e-04, 9.1027e-05, 8.1735e-05, 7.5294e-05, + 6.8800e-05, 8.5027e-05], device='cuda:4') +2023-03-26 09:28:13,442 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43162.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:28:14,068 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0295, 1.7955, 1.7428, 1.9841, 1.6081, 4.6493, 1.7076, 2.4054], + device='cuda:4'), covar=tensor([0.3174, 0.2276, 0.2013, 0.2208, 0.1626, 0.0104, 0.2423, 0.1218], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0115, 0.0118, 0.0122, 0.0116, 0.0098, 0.0100, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 09:28:23,322 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.56 vs. limit=5.0 +2023-03-26 09:28:36,035 INFO [finetune.py:976] (4/7) Epoch 8, batch 3100, loss[loss=0.1587, simple_loss=0.2319, pruned_loss=0.04273, over 4902.00 frames. ], tot_loss[loss=0.2059, simple_loss=0.2682, pruned_loss=0.07184, over 948362.90 frames. ], batch size: 46, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:28:52,833 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43210.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:28:59,343 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.0430, 3.3185, 2.9209, 2.2569, 3.1129, 3.4229, 3.2459, 2.8728], + device='cuda:4'), covar=tensor([0.0488, 0.0436, 0.0645, 0.0817, 0.0512, 0.0531, 0.0545, 0.0824], + device='cuda:4'), in_proj_covar=tensor([0.0135, 0.0132, 0.0143, 0.0125, 0.0114, 0.0143, 0.0144, 0.0159], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:29:01,133 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43215.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:29:03,637 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3009, 2.0998, 1.7152, 0.8206, 1.8838, 1.8448, 1.6697, 2.0184], + device='cuda:4'), covar=tensor([0.0929, 0.0864, 0.1434, 0.2082, 0.1464, 0.2364, 0.2334, 0.0903], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0201, 0.0202, 0.0187, 0.0218, 0.0206, 0.0222, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:29:14,666 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.165e+02 1.652e+02 1.930e+02 2.337e+02 4.149e+02, threshold=3.860e+02, percent-clipped=1.0 +2023-03-26 09:29:21,984 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8681, 4.7431, 4.5570, 2.4068, 4.7483, 3.4649, 0.8491, 3.4219], + device='cuda:4'), covar=tensor([0.2338, 0.1636, 0.1217, 0.3070, 0.0805, 0.0926, 0.4737, 0.1424], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0171, 0.0159, 0.0128, 0.0155, 0.0122, 0.0145, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 09:29:23,818 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43235.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:29:32,727 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 09:29:34,859 INFO [finetune.py:976] (4/7) Epoch 8, batch 3150, loss[loss=0.193, simple_loss=0.2439, pruned_loss=0.07101, over 4796.00 frames. ], tot_loss[loss=0.2022, simple_loss=0.2642, pruned_loss=0.0701, over 950411.43 frames. ], batch size: 29, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:30:08,688 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 09:30:24,889 INFO [finetune.py:976] (4/7) Epoch 8, batch 3200, loss[loss=0.1787, simple_loss=0.2474, pruned_loss=0.05505, over 4768.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.2608, pruned_loss=0.06866, over 950349.10 frames. ], batch size: 28, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:30:33,160 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43305.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:30:33,770 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6029, 1.5340, 1.4329, 1.5303, 1.1510, 3.3551, 1.3006, 1.7762], + device='cuda:4'), covar=tensor([0.4186, 0.3065, 0.2477, 0.2937, 0.1947, 0.0292, 0.2612, 0.1385], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0115, 0.0119, 0.0123, 0.0117, 0.0098, 0.0100, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 09:30:40,791 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7904, 1.6875, 1.5290, 1.8823, 2.3099, 1.8636, 1.5113, 1.4554], + device='cuda:4'), covar=tensor([0.2209, 0.2108, 0.1996, 0.1644, 0.1823, 0.1197, 0.2506, 0.1970], + device='cuda:4'), in_proj_covar=tensor([0.0235, 0.0207, 0.0203, 0.0186, 0.0238, 0.0177, 0.0212, 0.0192], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:30:49,479 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.668e+02 2.087e+02 2.541e+02 1.424e+03, threshold=4.174e+02, percent-clipped=3.0 +2023-03-26 09:31:03,654 INFO [finetune.py:976] (4/7) Epoch 8, batch 3250, loss[loss=0.184, simple_loss=0.2396, pruned_loss=0.06418, over 4754.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.2613, pruned_loss=0.06894, over 950673.48 frames. ], batch size: 27, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:31:15,391 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43353.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:31:16,742 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-03-26 09:31:59,348 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.42 vs. limit=5.0 +2023-03-26 09:31:59,776 INFO [finetune.py:976] (4/7) Epoch 8, batch 3300, loss[loss=0.1889, simple_loss=0.2563, pruned_loss=0.06071, over 4781.00 frames. ], tot_loss[loss=0.2027, simple_loss=0.265, pruned_loss=0.07022, over 950521.90 frames. ], batch size: 29, lr: 3.83e-03, grad_scale: 16.0 +2023-03-26 09:32:24,784 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8372, 1.1243, 1.7120, 1.7400, 1.5507, 1.5198, 1.6457, 1.6076], + device='cuda:4'), covar=tensor([0.3873, 0.5122, 0.4022, 0.4598, 0.5372, 0.4337, 0.5576, 0.3976], + device='cuda:4'), in_proj_covar=tensor([0.0232, 0.0241, 0.0254, 0.0254, 0.0245, 0.0222, 0.0272, 0.0227], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:32:45,627 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.223e+02 1.691e+02 2.029e+02 2.462e+02 4.055e+02, threshold=4.059e+02, percent-clipped=0.0 +2023-03-26 09:32:53,444 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7076, 1.5648, 1.3679, 1.1610, 1.5557, 1.5240, 1.5201, 2.0888], + device='cuda:4'), covar=tensor([0.5050, 0.5226, 0.3812, 0.4677, 0.4498, 0.2839, 0.4767, 0.2202], + device='cuda:4'), in_proj_covar=tensor([0.0284, 0.0256, 0.0219, 0.0278, 0.0239, 0.0204, 0.0243, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:33:04,670 INFO [finetune.py:976] (4/7) Epoch 8, batch 3350, loss[loss=0.1771, simple_loss=0.2524, pruned_loss=0.05094, over 4859.00 frames. ], tot_loss[loss=0.2049, simple_loss=0.2675, pruned_loss=0.07115, over 951025.08 frames. ], batch size: 44, lr: 3.83e-03, grad_scale: 32.0 +2023-03-26 09:33:35,198 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4626, 1.5298, 1.7694, 1.7865, 1.4228, 3.4275, 1.2758, 1.6001], + device='cuda:4'), covar=tensor([0.0956, 0.1727, 0.1210, 0.0938, 0.1650, 0.0226, 0.1485, 0.1640], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0076, 0.0079, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 09:33:37,078 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9215, 1.8323, 1.7017, 1.9998, 2.4251, 2.0946, 1.3463, 1.6030], + device='cuda:4'), covar=tensor([0.2319, 0.2136, 0.2035, 0.1856, 0.1924, 0.1120, 0.2785, 0.1990], + device='cuda:4'), in_proj_covar=tensor([0.0237, 0.0209, 0.0205, 0.0188, 0.0240, 0.0179, 0.0214, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:33:48,801 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8027, 1.5179, 2.1620, 3.3412, 2.4173, 2.3979, 0.9881, 2.5730], + device='cuda:4'), covar=tensor([0.1558, 0.1388, 0.1195, 0.0514, 0.0715, 0.1638, 0.1804, 0.0540], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0117, 0.0135, 0.0165, 0.0102, 0.0139, 0.0127, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 09:33:51,120 INFO [finetune.py:976] (4/7) Epoch 8, batch 3400, loss[loss=0.2311, simple_loss=0.2819, pruned_loss=0.09016, over 4681.00 frames. ], tot_loss[loss=0.2071, simple_loss=0.2698, pruned_loss=0.07219, over 949657.47 frames. ], batch size: 23, lr: 3.83e-03, grad_scale: 32.0 +2023-03-26 09:34:01,035 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7314, 2.5540, 2.3684, 2.8330, 2.5923, 2.5388, 2.4857, 3.4869], + device='cuda:4'), covar=tensor([0.4410, 0.5370, 0.3734, 0.4699, 0.4629, 0.2690, 0.5116, 0.1612], + device='cuda:4'), in_proj_covar=tensor([0.0284, 0.0257, 0.0220, 0.0278, 0.0239, 0.0205, 0.0243, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:34:04,487 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=43514.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:34:05,089 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43515.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:34:13,569 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.78 vs. limit=2.0 +2023-03-26 09:34:15,392 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.089e+02 1.653e+02 1.866e+02 2.233e+02 4.638e+02, threshold=3.733e+02, percent-clipped=1.0 +2023-03-26 09:34:19,633 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43535.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:34:25,009 INFO [finetune.py:976] (4/7) Epoch 8, batch 3450, loss[loss=0.2015, simple_loss=0.2601, pruned_loss=0.07142, over 4852.00 frames. ], tot_loss[loss=0.2055, simple_loss=0.2687, pruned_loss=0.07116, over 951371.66 frames. ], batch size: 44, lr: 3.83e-03, grad_scale: 32.0 +2023-03-26 09:34:43,305 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43563.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:34:55,908 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=43575.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:35:02,200 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=43583.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:35:08,907 INFO [finetune.py:976] (4/7) Epoch 8, batch 3500, loss[loss=0.2304, simple_loss=0.3006, pruned_loss=0.08015, over 4906.00 frames. ], tot_loss[loss=0.2029, simple_loss=0.2657, pruned_loss=0.07002, over 950513.94 frames. ], batch size: 46, lr: 3.83e-03, grad_scale: 32.0 +2023-03-26 09:35:32,858 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=2.05 vs. limit=2.0 +2023-03-26 09:35:34,548 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.199e+02 1.655e+02 2.028e+02 2.395e+02 4.370e+02, threshold=4.057e+02, percent-clipped=4.0 +2023-03-26 09:35:42,077 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-26 09:35:44,696 INFO [finetune.py:976] (4/7) Epoch 8, batch 3550, loss[loss=0.1874, simple_loss=0.2452, pruned_loss=0.06483, over 4844.00 frames. ], tot_loss[loss=0.2019, simple_loss=0.2639, pruned_loss=0.06999, over 951854.61 frames. ], batch size: 47, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:35:56,218 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.57 vs. limit=5.0 +2023-03-26 09:36:04,171 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8603, 1.1397, 1.8081, 1.7708, 1.5722, 1.5442, 1.6199, 1.6610], + device='cuda:4'), covar=tensor([0.4381, 0.5353, 0.4586, 0.4484, 0.6090, 0.4375, 0.5872, 0.4338], + device='cuda:4'), in_proj_covar=tensor([0.0232, 0.0241, 0.0254, 0.0254, 0.0245, 0.0222, 0.0272, 0.0227], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:36:17,761 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.95 vs. limit=5.0 +2023-03-26 09:36:18,005 INFO [finetune.py:976] (4/7) Epoch 8, batch 3600, loss[loss=0.2461, simple_loss=0.3074, pruned_loss=0.09235, over 4920.00 frames. ], tot_loss[loss=0.2007, simple_loss=0.2625, pruned_loss=0.06944, over 953891.50 frames. ], batch size: 42, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:36:40,282 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.156e+02 1.641e+02 1.836e+02 2.142e+02 3.900e+02, threshold=3.673e+02, percent-clipped=0.0 +2023-03-26 09:37:03,252 INFO [finetune.py:976] (4/7) Epoch 8, batch 3650, loss[loss=0.2264, simple_loss=0.2999, pruned_loss=0.07648, over 4181.00 frames. ], tot_loss[loss=0.2048, simple_loss=0.2663, pruned_loss=0.07162, over 953549.80 frames. ], batch size: 65, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:37:41,714 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=43787.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:37:53,054 INFO [finetune.py:976] (4/7) Epoch 8, batch 3700, loss[loss=0.2115, simple_loss=0.2612, pruned_loss=0.08089, over 4710.00 frames. ], tot_loss[loss=0.207, simple_loss=0.2692, pruned_loss=0.07241, over 950911.24 frames. ], batch size: 23, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:38:25,662 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-03-26 09:38:37,141 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.066e+01 1.637e+02 2.055e+02 2.501e+02 4.825e+02, threshold=4.110e+02, percent-clipped=4.0 +2023-03-26 09:38:56,857 INFO [finetune.py:976] (4/7) Epoch 8, batch 3750, loss[loss=0.2098, simple_loss=0.274, pruned_loss=0.07282, over 4899.00 frames. ], tot_loss[loss=0.2083, simple_loss=0.2709, pruned_loss=0.07283, over 951852.50 frames. ], batch size: 37, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:39:00,430 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=43848.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:39:13,801 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43870.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:39:30,695 INFO [finetune.py:976] (4/7) Epoch 8, batch 3800, loss[loss=0.2741, simple_loss=0.3271, pruned_loss=0.1105, over 4810.00 frames. ], tot_loss[loss=0.2096, simple_loss=0.2728, pruned_loss=0.07322, over 951982.64 frames. ], batch size: 38, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:39:40,903 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=43909.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:39:51,927 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5435, 1.0396, 0.8210, 1.4175, 1.9368, 0.7481, 1.2507, 1.3883], + device='cuda:4'), covar=tensor([0.1387, 0.2146, 0.1800, 0.1179, 0.1908, 0.1975, 0.1408, 0.1926], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0097, 0.0114, 0.0091, 0.0123, 0.0095, 0.0100, 0.0092], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 09:39:52,560 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9860, 1.9627, 2.0169, 1.2891, 2.1352, 2.0881, 1.9600, 1.6992], + device='cuda:4'), covar=tensor([0.0579, 0.0757, 0.0714, 0.0939, 0.0520, 0.0758, 0.0652, 0.1213], + device='cuda:4'), in_proj_covar=tensor([0.0137, 0.0133, 0.0145, 0.0126, 0.0115, 0.0144, 0.0145, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:40:01,520 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.167e+02 1.597e+02 1.980e+02 2.453e+02 5.062e+02, threshold=3.959e+02, percent-clipped=3.0 +2023-03-26 09:40:16,146 INFO [finetune.py:976] (4/7) Epoch 8, batch 3850, loss[loss=0.2256, simple_loss=0.2725, pruned_loss=0.08937, over 4814.00 frames. ], tot_loss[loss=0.2068, simple_loss=0.2703, pruned_loss=0.07169, over 953119.50 frames. ], batch size: 33, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:40:26,757 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=43959.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:40:33,727 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=43970.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:40:54,207 INFO [finetune.py:976] (4/7) Epoch 8, batch 3900, loss[loss=0.2013, simple_loss=0.2614, pruned_loss=0.07064, over 4880.00 frames. ], tot_loss[loss=0.203, simple_loss=0.2668, pruned_loss=0.06966, over 955443.51 frames. ], batch size: 32, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:41:17,077 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=44020.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:41:20,258 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 09:41:22,369 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.530e+02 1.807e+02 2.225e+02 3.627e+02, threshold=3.614e+02, percent-clipped=0.0 +2023-03-26 09:41:32,527 INFO [finetune.py:976] (4/7) Epoch 8, batch 3950, loss[loss=0.1736, simple_loss=0.2366, pruned_loss=0.05525, over 4755.00 frames. ], tot_loss[loss=0.2005, simple_loss=0.2634, pruned_loss=0.06877, over 956056.11 frames. ], batch size: 23, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:41:53,403 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6691, 2.4313, 2.8953, 1.8060, 2.7603, 2.9636, 2.3325, 3.0597], + device='cuda:4'), covar=tensor([0.1497, 0.1804, 0.1644, 0.2396, 0.1039, 0.1429, 0.2395, 0.1002], + device='cuda:4'), in_proj_covar=tensor([0.0203, 0.0205, 0.0199, 0.0197, 0.0181, 0.0221, 0.0221, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:42:06,489 INFO [finetune.py:976] (4/7) Epoch 8, batch 4000, loss[loss=0.1804, simple_loss=0.2408, pruned_loss=0.06001, over 4738.00 frames. ], tot_loss[loss=0.2017, simple_loss=0.2638, pruned_loss=0.06975, over 956571.20 frames. ], batch size: 23, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:42:37,515 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.793e+02 2.145e+02 2.588e+02 4.712e+02, threshold=4.291e+02, percent-clipped=10.0 +2023-03-26 09:42:56,498 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44143.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:42:57,050 INFO [finetune.py:976] (4/7) Epoch 8, batch 4050, loss[loss=0.1857, simple_loss=0.2336, pruned_loss=0.06894, over 4329.00 frames. ], tot_loss[loss=0.2044, simple_loss=0.267, pruned_loss=0.07094, over 956352.44 frames. ], batch size: 19, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:43:00,708 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4307, 2.0536, 2.7543, 1.6600, 2.5166, 2.5374, 2.0452, 2.7207], + device='cuda:4'), covar=tensor([0.1522, 0.2206, 0.1995, 0.2772, 0.0977, 0.1777, 0.2731, 0.1031], + device='cuda:4'), in_proj_covar=tensor([0.0203, 0.0206, 0.0200, 0.0197, 0.0182, 0.0222, 0.0222, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:43:21,558 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2157, 1.9683, 2.1913, 0.8046, 2.4424, 2.7273, 2.1139, 1.8932], + device='cuda:4'), covar=tensor([0.1011, 0.0868, 0.0545, 0.0817, 0.0576, 0.0513, 0.0543, 0.0736], + device='cuda:4'), in_proj_covar=tensor([0.0127, 0.0154, 0.0119, 0.0135, 0.0131, 0.0124, 0.0144, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.4607e-05, 1.1348e-04, 8.6079e-05, 9.7966e-05, 9.3567e-05, 9.0754e-05, + 1.0570e-04, 1.0718e-04], device='cuda:4') +2023-03-26 09:43:31,077 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44170.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:43:49,918 INFO [finetune.py:976] (4/7) Epoch 8, batch 4100, loss[loss=0.1851, simple_loss=0.2592, pruned_loss=0.05552, over 4902.00 frames. ], tot_loss[loss=0.2057, simple_loss=0.2687, pruned_loss=0.0714, over 953835.25 frames. ], batch size: 35, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:44:06,904 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 09:44:15,791 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=44218.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:44:22,431 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.086e+02 1.696e+02 1.895e+02 2.360e+02 4.949e+02, threshold=3.791e+02, percent-clipped=1.0 +2023-03-26 09:44:22,532 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0451, 1.6489, 2.4865, 3.9367, 2.6244, 2.6812, 1.0916, 3.1361], + device='cuda:4'), covar=tensor([0.1693, 0.1550, 0.1336, 0.0488, 0.0775, 0.1498, 0.1838, 0.0474], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0118, 0.0136, 0.0167, 0.0103, 0.0140, 0.0128, 0.0103], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 09:44:28,005 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6542, 1.4067, 2.2179, 3.4779, 2.3223, 2.4117, 0.8763, 2.6294], + device='cuda:4'), covar=tensor([0.1854, 0.1632, 0.1440, 0.0535, 0.0822, 0.1489, 0.2169, 0.0634], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0118, 0.0137, 0.0167, 0.0103, 0.0141, 0.0128, 0.0103], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 09:44:31,519 INFO [finetune.py:976] (4/7) Epoch 8, batch 4150, loss[loss=0.2461, simple_loss=0.3094, pruned_loss=0.09137, over 4919.00 frames. ], tot_loss[loss=0.2071, simple_loss=0.27, pruned_loss=0.07208, over 953399.21 frames. ], batch size: 33, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:44:46,755 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44265.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:45:07,274 INFO [finetune.py:976] (4/7) Epoch 8, batch 4200, loss[loss=0.149, simple_loss=0.2212, pruned_loss=0.03838, over 4756.00 frames. ], tot_loss[loss=0.2058, simple_loss=0.2691, pruned_loss=0.07122, over 953012.33 frames. ], batch size: 28, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:45:28,260 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0151, 1.8783, 1.5339, 1.7409, 1.7723, 1.7191, 1.7638, 2.5522], + device='cuda:4'), covar=tensor([0.4836, 0.5473, 0.4211, 0.4965, 0.4605, 0.2967, 0.4789, 0.1844], + device='cuda:4'), in_proj_covar=tensor([0.0284, 0.0259, 0.0220, 0.0279, 0.0240, 0.0205, 0.0243, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:45:30,593 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44315.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:45:40,454 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.242e+02 1.746e+02 1.975e+02 2.337e+02 5.106e+02, threshold=3.951e+02, percent-clipped=1.0 +2023-03-26 09:45:54,478 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 09:45:54,906 INFO [finetune.py:976] (4/7) Epoch 8, batch 4250, loss[loss=0.1942, simple_loss=0.2653, pruned_loss=0.06156, over 4913.00 frames. ], tot_loss[loss=0.2051, simple_loss=0.2681, pruned_loss=0.07106, over 954023.16 frames. ], batch size: 36, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:46:27,765 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0498, 2.2052, 2.0767, 1.5116, 2.2175, 2.2136, 2.0532, 1.8469], + device='cuda:4'), covar=tensor([0.0626, 0.0588, 0.0785, 0.0892, 0.0514, 0.0809, 0.0735, 0.1034], + device='cuda:4'), in_proj_covar=tensor([0.0136, 0.0133, 0.0145, 0.0125, 0.0116, 0.0145, 0.0145, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:46:32,567 INFO [finetune.py:976] (4/7) Epoch 8, batch 4300, loss[loss=0.1958, simple_loss=0.2658, pruned_loss=0.06295, over 4914.00 frames. ], tot_loss[loss=0.2014, simple_loss=0.2644, pruned_loss=0.06918, over 954234.19 frames. ], batch size: 37, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:46:35,136 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0364, 2.0270, 1.8943, 2.2629, 2.6995, 2.0825, 1.8142, 1.5655], + device='cuda:4'), covar=tensor([0.2284, 0.2128, 0.1950, 0.1720, 0.1940, 0.1218, 0.2503, 0.1995], + device='cuda:4'), in_proj_covar=tensor([0.0235, 0.0208, 0.0205, 0.0187, 0.0239, 0.0178, 0.0213, 0.0193], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:46:56,829 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.550e+02 1.851e+02 2.365e+02 4.860e+02, threshold=3.701e+02, percent-clipped=2.0 +2023-03-26 09:47:05,843 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44443.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:47:06,352 INFO [finetune.py:976] (4/7) Epoch 8, batch 4350, loss[loss=0.1697, simple_loss=0.2386, pruned_loss=0.05046, over 4756.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.2618, pruned_loss=0.06848, over 954409.63 frames. ], batch size: 27, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:47:37,687 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=44491.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:47:39,435 INFO [finetune.py:976] (4/7) Epoch 8, batch 4400, loss[loss=0.2042, simple_loss=0.2677, pruned_loss=0.07036, over 4792.00 frames. ], tot_loss[loss=0.2006, simple_loss=0.2629, pruned_loss=0.06913, over 954781.24 frames. ], batch size: 51, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:48:17,052 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=44527.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:48:18,616 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.261e+02 1.746e+02 1.995e+02 2.515e+02 6.158e+02, threshold=3.991e+02, percent-clipped=2.0 +2023-03-26 09:48:28,720 INFO [finetune.py:976] (4/7) Epoch 8, batch 4450, loss[loss=0.2447, simple_loss=0.3014, pruned_loss=0.09402, over 4753.00 frames. ], tot_loss[loss=0.2037, simple_loss=0.2665, pruned_loss=0.07043, over 954767.58 frames. ], batch size: 54, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:48:37,006 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=44550.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:48:37,700 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.62 vs. limit=2.0 +2023-03-26 09:48:51,545 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44565.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:49:19,339 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=44588.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:49:28,024 INFO [finetune.py:976] (4/7) Epoch 8, batch 4500, loss[loss=0.2284, simple_loss=0.2782, pruned_loss=0.08925, over 4816.00 frames. ], tot_loss[loss=0.2057, simple_loss=0.2685, pruned_loss=0.07147, over 954554.78 frames. ], batch size: 25, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:49:47,675 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=44611.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:49:48,883 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=44613.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:49:50,681 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44615.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:50:00,000 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.305e+02 1.720e+02 2.026e+02 2.465e+02 5.780e+02, threshold=4.053e+02, percent-clipped=2.0 +2023-03-26 09:50:10,542 INFO [finetune.py:976] (4/7) Epoch 8, batch 4550, loss[loss=0.1435, simple_loss=0.2186, pruned_loss=0.03414, over 4759.00 frames. ], tot_loss[loss=0.2062, simple_loss=0.2696, pruned_loss=0.07139, over 955216.14 frames. ], batch size: 27, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:50:27,737 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=44663.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:50:52,807 INFO [finetune.py:976] (4/7) Epoch 8, batch 4600, loss[loss=0.2016, simple_loss=0.2621, pruned_loss=0.07052, over 4848.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.2673, pruned_loss=0.06978, over 956703.07 frames. ], batch size: 44, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:51:15,457 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.778e+01 1.566e+02 1.839e+02 2.114e+02 3.234e+02, threshold=3.678e+02, percent-clipped=0.0 +2023-03-26 09:51:25,985 INFO [finetune.py:976] (4/7) Epoch 8, batch 4650, loss[loss=0.2023, simple_loss=0.2662, pruned_loss=0.06919, over 4820.00 frames. ], tot_loss[loss=0.2024, simple_loss=0.2656, pruned_loss=0.06954, over 952531.75 frames. ], batch size: 40, lr: 3.82e-03, grad_scale: 32.0 +2023-03-26 09:51:59,446 INFO [finetune.py:976] (4/7) Epoch 8, batch 4700, loss[loss=0.1708, simple_loss=0.2389, pruned_loss=0.05138, over 4760.00 frames. ], tot_loss[loss=0.2005, simple_loss=0.2635, pruned_loss=0.06874, over 951388.70 frames. ], batch size: 27, lr: 3.82e-03, grad_scale: 16.0 +2023-03-26 09:52:06,783 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6514, 1.4892, 1.5489, 1.5973, 1.0638, 3.0094, 1.1719, 1.6597], + device='cuda:4'), covar=tensor([0.3448, 0.2504, 0.2070, 0.2238, 0.1910, 0.0249, 0.2681, 0.1288], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0115, 0.0120, 0.0123, 0.0116, 0.0098, 0.0101, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 09:52:09,635 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8183, 1.8094, 1.9009, 1.2254, 1.8725, 1.8675, 1.8356, 1.5946], + device='cuda:4'), covar=tensor([0.0519, 0.0613, 0.0603, 0.0807, 0.0589, 0.0610, 0.0596, 0.1062], + device='cuda:4'), in_proj_covar=tensor([0.0135, 0.0132, 0.0144, 0.0125, 0.0114, 0.0144, 0.0144, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:52:22,748 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.499e+02 1.877e+02 2.319e+02 4.193e+02, threshold=3.754e+02, percent-clipped=1.0 +2023-03-26 09:52:32,232 INFO [finetune.py:976] (4/7) Epoch 8, batch 4750, loss[loss=0.1947, simple_loss=0.2634, pruned_loss=0.06298, over 4939.00 frames. ], tot_loss[loss=0.1985, simple_loss=0.2609, pruned_loss=0.06805, over 952205.64 frames. ], batch size: 33, lr: 3.82e-03, grad_scale: 16.0 +2023-03-26 09:52:58,203 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44883.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:52:58,264 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0726, 1.8966, 1.9463, 0.8744, 2.1241, 2.4421, 2.0300, 1.8554], + device='cuda:4'), covar=tensor([0.1210, 0.0749, 0.0534, 0.0750, 0.0632, 0.0579, 0.0516, 0.0772], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0158, 0.0122, 0.0138, 0.0134, 0.0126, 0.0147, 0.0150], + device='cuda:4'), out_proj_covar=tensor([9.7489e-05, 1.1624e-04, 8.8554e-05, 1.0003e-04, 9.5908e-05, 9.2513e-05, + 1.0823e-04, 1.1029e-04], device='cuda:4') +2023-03-26 09:53:05,275 INFO [finetune.py:976] (4/7) Epoch 8, batch 4800, loss[loss=0.1936, simple_loss=0.2504, pruned_loss=0.0684, over 4708.00 frames. ], tot_loss[loss=0.201, simple_loss=0.2631, pruned_loss=0.0694, over 951368.07 frames. ], batch size: 23, lr: 3.82e-03, grad_scale: 16.0 +2023-03-26 09:53:15,869 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44906.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:53:41,286 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.589e+02 1.932e+02 2.383e+02 4.430e+02, threshold=3.864e+02, percent-clipped=2.0 +2023-03-26 09:53:52,419 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0403, 2.0911, 1.7841, 1.6742, 2.4023, 2.4023, 2.1646, 1.9921], + device='cuda:4'), covar=tensor([0.0335, 0.0355, 0.0579, 0.0368, 0.0242, 0.0511, 0.0323, 0.0391], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0110, 0.0140, 0.0115, 0.0103, 0.0101, 0.0091, 0.0108], + device='cuda:4'), out_proj_covar=tensor([6.9433e-05, 8.5950e-05, 1.1171e-04, 9.0358e-05, 8.0674e-05, 7.4877e-05, + 6.8868e-05, 8.3833e-05], device='cuda:4') +2023-03-26 09:53:55,401 INFO [finetune.py:976] (4/7) Epoch 8, batch 4850, loss[loss=0.2564, simple_loss=0.3209, pruned_loss=0.09596, over 4846.00 frames. ], tot_loss[loss=0.2046, simple_loss=0.2673, pruned_loss=0.07093, over 952220.47 frames. ], batch size: 44, lr: 3.82e-03, grad_scale: 16.0 +2023-03-26 09:54:30,839 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6929, 1.5479, 1.4119, 1.3177, 1.7659, 1.4319, 1.8131, 1.7216], + device='cuda:4'), covar=tensor([0.1530, 0.2256, 0.3149, 0.2726, 0.2550, 0.1805, 0.3204, 0.1929], + device='cuda:4'), in_proj_covar=tensor([0.0172, 0.0189, 0.0233, 0.0254, 0.0235, 0.0194, 0.0211, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:54:40,772 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5670, 1.4308, 1.4673, 1.5796, 1.2661, 3.4020, 1.3027, 1.7564], + device='cuda:4'), covar=tensor([0.3542, 0.2513, 0.2190, 0.2291, 0.1751, 0.0202, 0.2775, 0.1376], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0115, 0.0120, 0.0123, 0.0117, 0.0098, 0.0101, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 09:54:57,804 INFO [finetune.py:976] (4/7) Epoch 8, batch 4900, loss[loss=0.221, simple_loss=0.2937, pruned_loss=0.07411, over 4908.00 frames. ], tot_loss[loss=0.2072, simple_loss=0.27, pruned_loss=0.07219, over 952799.48 frames. ], batch size: 43, lr: 3.82e-03, grad_scale: 16.0 +2023-03-26 09:55:23,358 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 09:55:25,592 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.182e+02 1.694e+02 2.008e+02 2.325e+02 4.035e+02, threshold=4.016e+02, percent-clipped=1.0 +2023-03-26 09:55:43,911 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8722, 1.2878, 1.6563, 1.6939, 1.5422, 1.5619, 1.5802, 1.6169], + device='cuda:4'), covar=tensor([0.6140, 0.7019, 0.6102, 0.6434, 0.7654, 0.6152, 0.8183, 0.5984], + device='cuda:4'), in_proj_covar=tensor([0.0232, 0.0241, 0.0252, 0.0254, 0.0246, 0.0223, 0.0273, 0.0227], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:55:44,371 INFO [finetune.py:976] (4/7) Epoch 8, batch 4950, loss[loss=0.2108, simple_loss=0.277, pruned_loss=0.0723, over 4883.00 frames. ], tot_loss[loss=0.2067, simple_loss=0.2703, pruned_loss=0.07156, over 950273.44 frames. ], batch size: 35, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 09:55:57,051 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45056.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:56:02,506 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 09:56:03,635 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.0299, 2.6449, 2.5554, 1.4309, 2.6394, 2.1301, 1.9795, 2.3117], + device='cuda:4'), covar=tensor([0.1035, 0.0935, 0.1751, 0.2274, 0.1835, 0.2401, 0.2225, 0.1298], + device='cuda:4'), in_proj_covar=tensor([0.0167, 0.0200, 0.0201, 0.0188, 0.0216, 0.0205, 0.0221, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:56:13,242 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45081.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:56:18,037 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3792, 2.9022, 2.7491, 1.1850, 3.0016, 2.2798, 0.8028, 1.9511], + device='cuda:4'), covar=tensor([0.2500, 0.2163, 0.1863, 0.3357, 0.1314, 0.1162, 0.3743, 0.1678], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0172, 0.0161, 0.0129, 0.0156, 0.0122, 0.0147, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 09:56:21,542 INFO [finetune.py:976] (4/7) Epoch 8, batch 5000, loss[loss=0.1603, simple_loss=0.2218, pruned_loss=0.04943, over 4794.00 frames. ], tot_loss[loss=0.2028, simple_loss=0.2669, pruned_loss=0.06933, over 950508.42 frames. ], batch size: 26, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 09:56:37,070 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45117.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 09:56:45,216 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.621e+02 1.919e+02 2.483e+02 3.797e+02, threshold=3.837e+02, percent-clipped=0.0 +2023-03-26 09:56:52,662 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45142.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:56:53,773 INFO [finetune.py:976] (4/7) Epoch 8, batch 5050, loss[loss=0.2055, simple_loss=0.265, pruned_loss=0.07304, over 4837.00 frames. ], tot_loss[loss=0.2014, simple_loss=0.2644, pruned_loss=0.06916, over 951573.98 frames. ], batch size: 33, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 09:56:55,594 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7059, 1.6261, 1.6680, 1.7826, 1.1424, 3.7362, 1.4415, 1.9851], + device='cuda:4'), covar=tensor([0.3273, 0.2409, 0.2037, 0.2229, 0.1908, 0.0168, 0.2627, 0.1315], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0116, 0.0120, 0.0123, 0.0117, 0.0098, 0.0101, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 09:57:02,260 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9770, 1.6798, 2.3115, 2.3409, 2.0983, 4.4978, 1.7376, 2.3245], + device='cuda:4'), covar=tensor([0.0864, 0.1657, 0.0994, 0.0906, 0.1392, 0.0228, 0.1356, 0.1455], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0092, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 09:57:10,991 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8071, 1.2203, 0.8458, 1.6364, 2.1587, 1.3896, 1.4396, 1.7473], + device='cuda:4'), covar=tensor([0.1362, 0.2017, 0.2148, 0.1182, 0.1893, 0.2093, 0.1437, 0.1905], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0097, 0.0115, 0.0093, 0.0123, 0.0096, 0.0101, 0.0093], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 09:57:19,961 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45183.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:57:26,576 INFO [finetune.py:976] (4/7) Epoch 8, batch 5100, loss[loss=0.179, simple_loss=0.2307, pruned_loss=0.06364, over 4805.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2613, pruned_loss=0.06777, over 954601.94 frames. ], batch size: 45, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 09:57:34,994 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45206.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:57:36,796 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6616, 3.3666, 3.2651, 1.4703, 3.4476, 2.4932, 0.7944, 2.3555], + device='cuda:4'), covar=tensor([0.2485, 0.2025, 0.1667, 0.3612, 0.1213, 0.1123, 0.4517, 0.1570], + device='cuda:4'), in_proj_covar=tensor([0.0154, 0.0174, 0.0161, 0.0129, 0.0156, 0.0123, 0.0147, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 09:57:37,427 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1557, 1.9580, 1.9578, 0.8028, 2.1921, 2.4253, 2.1202, 1.8279], + device='cuda:4'), covar=tensor([0.0941, 0.0697, 0.0506, 0.0724, 0.0449, 0.0499, 0.0444, 0.0620], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0158, 0.0122, 0.0138, 0.0134, 0.0126, 0.0147, 0.0150], + device='cuda:4'), out_proj_covar=tensor([9.7325e-05, 1.1635e-04, 8.8465e-05, 1.0023e-04, 9.5704e-05, 9.2397e-05, + 1.0815e-04, 1.1013e-04], device='cuda:4') +2023-03-26 09:57:55,113 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.141e+02 1.630e+02 1.903e+02 2.262e+02 3.588e+02, threshold=3.806e+02, percent-clipped=0.0 +2023-03-26 09:57:55,793 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=45231.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:58:02,966 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.9261, 3.4897, 3.6225, 3.8271, 3.6718, 3.4800, 3.9791, 1.3968], + device='cuda:4'), covar=tensor([0.0962, 0.0779, 0.0823, 0.1068, 0.1435, 0.1472, 0.0776, 0.5045], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0242, 0.0276, 0.0294, 0.0333, 0.0283, 0.0302, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:58:04,092 INFO [finetune.py:976] (4/7) Epoch 8, batch 5150, loss[loss=0.2122, simple_loss=0.2813, pruned_loss=0.07159, over 4819.00 frames. ], tot_loss[loss=0.1997, simple_loss=0.2619, pruned_loss=0.06878, over 954463.71 frames. ], batch size: 39, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 09:58:11,268 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=45254.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 09:58:36,779 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0909, 1.9540, 1.8287, 2.1766, 1.5612, 4.6945, 1.8906, 2.4788], + device='cuda:4'), covar=tensor([0.3194, 0.2405, 0.1956, 0.2092, 0.1638, 0.0089, 0.2315, 0.1220], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0115, 0.0119, 0.0122, 0.0116, 0.0098, 0.0100, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 09:58:40,360 INFO [finetune.py:976] (4/7) Epoch 8, batch 5200, loss[loss=0.1873, simple_loss=0.2529, pruned_loss=0.06083, over 4891.00 frames. ], tot_loss[loss=0.2019, simple_loss=0.2649, pruned_loss=0.06944, over 954207.30 frames. ], batch size: 35, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 09:59:09,668 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.225e+02 1.659e+02 1.911e+02 2.326e+02 4.760e+02, threshold=3.822e+02, percent-clipped=1.0 +2023-03-26 09:59:11,061 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3746, 2.1935, 1.8338, 0.8367, 1.9874, 1.7977, 1.6109, 1.9549], + device='cuda:4'), covar=tensor([0.0897, 0.0828, 0.1590, 0.2215, 0.1515, 0.2334, 0.2235, 0.1080], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0201, 0.0202, 0.0189, 0.0217, 0.0206, 0.0223, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 09:59:18,772 INFO [finetune.py:976] (4/7) Epoch 8, batch 5250, loss[loss=0.222, simple_loss=0.2823, pruned_loss=0.08085, over 4857.00 frames. ], tot_loss[loss=0.2023, simple_loss=0.2663, pruned_loss=0.06919, over 955065.44 frames. ], batch size: 44, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 09:59:27,669 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45348.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 09:59:38,331 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-03-26 10:00:03,243 INFO [finetune.py:976] (4/7) Epoch 8, batch 5300, loss[loss=0.1891, simple_loss=0.2533, pruned_loss=0.06242, over 4753.00 frames. ], tot_loss[loss=0.2052, simple_loss=0.2691, pruned_loss=0.07063, over 954808.99 frames. ], batch size: 27, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:00:06,191 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7765, 3.8740, 3.7887, 2.0403, 3.9854, 2.8694, 0.7536, 2.7740], + device='cuda:4'), covar=tensor([0.2376, 0.1714, 0.1454, 0.3064, 0.0955, 0.0992, 0.4685, 0.1383], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0173, 0.0160, 0.0129, 0.0156, 0.0122, 0.0146, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 10:00:16,221 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45409.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 10:00:18,478 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=45412.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 10:00:30,691 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.820e+01 1.516e+02 1.858e+02 2.399e+02 4.469e+02, threshold=3.716e+02, percent-clipped=2.0 +2023-03-26 10:00:39,974 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=45437.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:00:49,409 INFO [finetune.py:976] (4/7) Epoch 8, batch 5350, loss[loss=0.2004, simple_loss=0.2635, pruned_loss=0.06866, over 4811.00 frames. ], tot_loss[loss=0.2057, simple_loss=0.27, pruned_loss=0.07074, over 956001.22 frames. ], batch size: 33, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:01:02,980 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5963, 1.6237, 1.8375, 1.9412, 1.8959, 3.5681, 1.6033, 1.8849], + device='cuda:4'), covar=tensor([0.0962, 0.1681, 0.1005, 0.0927, 0.1351, 0.0232, 0.1301, 0.1554], + device='cuda:4'), in_proj_covar=tensor([0.0077, 0.0082, 0.0076, 0.0078, 0.0093, 0.0083, 0.0086, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 10:01:47,392 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 10:01:54,085 INFO [finetune.py:976] (4/7) Epoch 8, batch 5400, loss[loss=0.2274, simple_loss=0.2839, pruned_loss=0.08542, over 4861.00 frames. ], tot_loss[loss=0.2029, simple_loss=0.2667, pruned_loss=0.06959, over 955894.33 frames. ], batch size: 47, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:02:34,313 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.106e+02 1.640e+02 2.019e+02 2.439e+02 4.086e+02, threshold=4.037e+02, percent-clipped=1.0 +2023-03-26 10:02:42,401 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1259, 1.9937, 1.4716, 1.9730, 2.0273, 1.6927, 2.6980, 2.0368], + device='cuda:4'), covar=tensor([0.1374, 0.2344, 0.3769, 0.3339, 0.2826, 0.1774, 0.2735, 0.2031], + device='cuda:4'), in_proj_covar=tensor([0.0173, 0.0190, 0.0235, 0.0256, 0.0237, 0.0195, 0.0212, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:02:54,209 INFO [finetune.py:976] (4/7) Epoch 8, batch 5450, loss[loss=0.1952, simple_loss=0.2552, pruned_loss=0.06756, over 4923.00 frames. ], tot_loss[loss=0.2005, simple_loss=0.2637, pruned_loss=0.06865, over 955338.19 frames. ], batch size: 46, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:03:54,239 INFO [finetune.py:976] (4/7) Epoch 8, batch 5500, loss[loss=0.198, simple_loss=0.2531, pruned_loss=0.07147, over 4867.00 frames. ], tot_loss[loss=0.1978, simple_loss=0.2603, pruned_loss=0.06763, over 955564.05 frames. ], batch size: 31, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:04:01,643 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4021, 2.8373, 2.7581, 1.1359, 2.9646, 2.1358, 0.8472, 1.8489], + device='cuda:4'), covar=tensor([0.2690, 0.2223, 0.1817, 0.3813, 0.1502, 0.1150, 0.4047, 0.1726], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0174, 0.0161, 0.0130, 0.0157, 0.0123, 0.0147, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 10:04:09,866 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 10:04:15,383 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4921, 1.3710, 1.4077, 1.4417, 0.8356, 2.2450, 0.7460, 1.3045], + device='cuda:4'), covar=tensor([0.3538, 0.2659, 0.2234, 0.2479, 0.2039, 0.0393, 0.2836, 0.1412], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0115, 0.0119, 0.0122, 0.0116, 0.0098, 0.0100, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 10:04:18,374 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.004e+02 1.550e+02 1.829e+02 2.210e+02 3.729e+02, threshold=3.658e+02, percent-clipped=0.0 +2023-03-26 10:04:28,412 INFO [finetune.py:976] (4/7) Epoch 8, batch 5550, loss[loss=0.1517, simple_loss=0.2155, pruned_loss=0.04399, over 4689.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.2622, pruned_loss=0.06847, over 956086.69 frames. ], batch size: 23, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:05:19,423 INFO [finetune.py:976] (4/7) Epoch 8, batch 5600, loss[loss=0.1852, simple_loss=0.2554, pruned_loss=0.05747, over 4801.00 frames. ], tot_loss[loss=0.202, simple_loss=0.2655, pruned_loss=0.0693, over 956086.28 frames. ], batch size: 25, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:05:24,106 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45702.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:05:25,258 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=45704.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 10:05:29,917 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45712.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 10:05:40,352 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.219e+02 1.739e+02 1.993e+02 2.505e+02 5.014e+02, threshold=3.987e+02, percent-clipped=3.0 +2023-03-26 10:05:44,489 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45737.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:05:48,900 INFO [finetune.py:976] (4/7) Epoch 8, batch 5650, loss[loss=0.2599, simple_loss=0.3299, pruned_loss=0.09499, over 4901.00 frames. ], tot_loss[loss=0.2043, simple_loss=0.2683, pruned_loss=0.07012, over 955698.27 frames. ], batch size: 43, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:05:56,844 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0218, 1.6295, 2.4138, 3.8558, 2.6931, 2.5384, 0.5946, 3.1233], + device='cuda:4'), covar=tensor([0.1808, 0.1616, 0.1370, 0.0547, 0.0782, 0.1723, 0.2251, 0.0518], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0118, 0.0135, 0.0167, 0.0103, 0.0140, 0.0127, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 10:05:58,592 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=45760.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:06:00,387 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45763.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:06:06,229 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 10:06:06,752 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45774.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:06:08,531 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4854, 2.2554, 2.0078, 2.2324, 2.3257, 2.0765, 2.6860, 2.3869], + device='cuda:4'), covar=tensor([0.1385, 0.2524, 0.3297, 0.3005, 0.2787, 0.1742, 0.3271, 0.2050], + device='cuda:4'), in_proj_covar=tensor([0.0173, 0.0190, 0.0234, 0.0255, 0.0238, 0.0195, 0.0213, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:06:13,160 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=45785.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:06:18,815 INFO [finetune.py:976] (4/7) Epoch 8, batch 5700, loss[loss=0.2028, simple_loss=0.2463, pruned_loss=0.07967, over 3976.00 frames. ], tot_loss[loss=0.2029, simple_loss=0.2654, pruned_loss=0.07025, over 940661.07 frames. ], batch size: 17, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:06:54,870 INFO [finetune.py:976] (4/7) Epoch 9, batch 0, loss[loss=0.2312, simple_loss=0.3097, pruned_loss=0.07639, over 4901.00 frames. ], tot_loss[loss=0.2312, simple_loss=0.3097, pruned_loss=0.07639, over 4901.00 frames. ], batch size: 46, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:06:54,870 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 10:07:00,914 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6552, 1.4391, 1.9822, 2.8366, 1.9483, 2.2665, 0.9075, 2.2763], + device='cuda:4'), covar=tensor([0.1791, 0.1593, 0.1232, 0.0753, 0.1002, 0.1211, 0.1904, 0.0725], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0118, 0.0135, 0.0166, 0.0103, 0.0140, 0.0127, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 10:07:11,027 INFO [finetune.py:1010] (4/7) Epoch 9, validation: loss=0.1616, simple_loss=0.233, pruned_loss=0.04515, over 2265189.00 frames. +2023-03-26 10:07:11,028 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6336MB +2023-03-26 10:07:17,601 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.132e+01 1.600e+02 1.914e+02 2.307e+02 4.538e+02, threshold=3.829e+02, percent-clipped=2.0 +2023-03-26 10:07:22,768 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45835.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:07:33,763 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5776, 1.1591, 0.8594, 1.5452, 2.0005, 1.2355, 1.3488, 1.5461], + device='cuda:4'), covar=tensor([0.1521, 0.2114, 0.2016, 0.1184, 0.1948, 0.2178, 0.1478, 0.1984], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0097, 0.0114, 0.0092, 0.0123, 0.0096, 0.0100, 0.0093], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 10:07:46,513 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45859.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:07:55,654 INFO [finetune.py:976] (4/7) Epoch 9, batch 50, loss[loss=0.2099, simple_loss=0.2687, pruned_loss=0.07552, over 4889.00 frames. ], tot_loss[loss=0.2053, simple_loss=0.2715, pruned_loss=0.06953, over 216775.47 frames. ], batch size: 36, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:08:13,654 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6414, 1.6270, 1.6874, 0.9095, 1.7949, 1.8814, 1.9125, 1.4485], + device='cuda:4'), covar=tensor([0.0969, 0.0670, 0.0415, 0.0644, 0.0327, 0.0648, 0.0319, 0.0739], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0157, 0.0121, 0.0136, 0.0132, 0.0125, 0.0147, 0.0148], + device='cuda:4'), out_proj_covar=tensor([9.6434e-05, 1.1582e-04, 8.7317e-05, 9.9096e-05, 9.4817e-05, 9.1873e-05, + 1.0774e-04, 1.0883e-04], device='cuda:4') +2023-03-26 10:08:35,699 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45920.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:08:36,813 INFO [finetune.py:976] (4/7) Epoch 9, batch 100, loss[loss=0.2143, simple_loss=0.2704, pruned_loss=0.07916, over 4741.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2626, pruned_loss=0.06707, over 381653.72 frames. ], batch size: 54, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:08:42,568 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.760e+02 2.008e+02 2.423e+02 3.807e+02, threshold=4.016e+02, percent-clipped=0.0 +2023-03-26 10:09:09,545 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-26 10:09:10,374 INFO [finetune.py:976] (4/7) Epoch 9, batch 150, loss[loss=0.1726, simple_loss=0.233, pruned_loss=0.05616, over 4908.00 frames. ], tot_loss[loss=0.1956, simple_loss=0.259, pruned_loss=0.06608, over 508282.97 frames. ], batch size: 36, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:09:32,061 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46004.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 10:09:49,161 INFO [finetune.py:976] (4/7) Epoch 9, batch 200, loss[loss=0.204, simple_loss=0.2696, pruned_loss=0.06919, over 4927.00 frames. ], tot_loss[loss=0.1975, simple_loss=0.2591, pruned_loss=0.0679, over 608618.90 frames. ], batch size: 38, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:09:58,662 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.086e+02 1.678e+02 2.056e+02 2.461e+02 4.455e+02, threshold=4.113e+02, percent-clipped=4.0 +2023-03-26 10:10:22,924 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46052.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 10:10:26,560 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46058.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:10:36,486 INFO [finetune.py:976] (4/7) Epoch 9, batch 250, loss[loss=0.1801, simple_loss=0.2453, pruned_loss=0.0574, over 4823.00 frames. ], tot_loss[loss=0.201, simple_loss=0.2629, pruned_loss=0.06956, over 686779.12 frames. ], batch size: 30, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:10:52,304 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.62 vs. limit=5.0 +2023-03-26 10:10:55,084 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-03-26 10:10:56,452 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 10:11:03,451 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-26 10:11:08,999 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 10:11:09,157 INFO [finetune.py:976] (4/7) Epoch 9, batch 300, loss[loss=0.2184, simple_loss=0.2915, pruned_loss=0.07266, over 4812.00 frames. ], tot_loss[loss=0.2039, simple_loss=0.2666, pruned_loss=0.07058, over 746862.49 frames. ], batch size: 45, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:11:14,967 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.304e+02 1.737e+02 2.021e+02 2.354e+02 3.684e+02, threshold=4.042e+02, percent-clipped=0.0 +2023-03-26 10:11:15,051 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46130.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:11:30,628 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8880, 2.1168, 1.4393, 2.7097, 3.0713, 2.5168, 2.3544, 2.6574], + device='cuda:4'), covar=tensor([0.1172, 0.1744, 0.1800, 0.0939, 0.1522, 0.1445, 0.1357, 0.1721], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0096, 0.0114, 0.0092, 0.0123, 0.0095, 0.0100, 0.0092], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 10:11:33,606 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5285, 1.3663, 1.2548, 1.5092, 1.6421, 1.5467, 1.0198, 1.3413], + device='cuda:4'), covar=tensor([0.1927, 0.1922, 0.1780, 0.1453, 0.1524, 0.1191, 0.2429, 0.1700], + device='cuda:4'), in_proj_covar=tensor([0.0237, 0.0207, 0.0204, 0.0186, 0.0239, 0.0177, 0.0212, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:11:41,958 INFO [finetune.py:976] (4/7) Epoch 9, batch 350, loss[loss=0.2631, simple_loss=0.325, pruned_loss=0.1006, over 4847.00 frames. ], tot_loss[loss=0.2062, simple_loss=0.2693, pruned_loss=0.07154, over 792415.11 frames. ], batch size: 49, lr: 3.81e-03, grad_scale: 16.0 +2023-03-26 10:12:11,375 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46215.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:12:17,562 INFO [finetune.py:976] (4/7) Epoch 9, batch 400, loss[loss=0.189, simple_loss=0.2611, pruned_loss=0.05847, over 4819.00 frames. ], tot_loss[loss=0.2061, simple_loss=0.2696, pruned_loss=0.07129, over 829209.96 frames. ], batch size: 38, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:12:23,439 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.951e+01 1.676e+02 2.053e+02 2.418e+02 4.627e+02, threshold=4.106e+02, percent-clipped=2.0 +2023-03-26 10:13:00,671 INFO [finetune.py:976] (4/7) Epoch 9, batch 450, loss[loss=0.1715, simple_loss=0.2397, pruned_loss=0.05162, over 4708.00 frames. ], tot_loss[loss=0.2047, simple_loss=0.2682, pruned_loss=0.07061, over 856799.91 frames. ], batch size: 23, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:13:03,229 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46276.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:13:26,992 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.33 vs. limit=5.0 +2023-03-26 10:13:35,969 INFO [finetune.py:976] (4/7) Epoch 9, batch 500, loss[loss=0.2076, simple_loss=0.2613, pruned_loss=0.07696, over 4902.00 frames. ], tot_loss[loss=0.2025, simple_loss=0.2659, pruned_loss=0.0696, over 877327.03 frames. ], batch size: 35, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:13:45,335 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.146e+02 1.586e+02 1.900e+02 2.408e+02 3.619e+02, threshold=3.799e+02, percent-clipped=0.0 +2023-03-26 10:13:54,721 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46337.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:13:59,925 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-26 10:14:08,381 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46358.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:14:09,078 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-03-26 10:14:17,264 INFO [finetune.py:976] (4/7) Epoch 9, batch 550, loss[loss=0.2809, simple_loss=0.3171, pruned_loss=0.1223, over 4829.00 frames. ], tot_loss[loss=0.2003, simple_loss=0.263, pruned_loss=0.06875, over 894720.57 frames. ], batch size: 38, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:14:39,977 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46406.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:14:43,745 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3548, 1.5143, 1.4956, 0.8011, 1.4332, 1.7465, 1.7350, 1.3540], + device='cuda:4'), covar=tensor([0.0941, 0.0523, 0.0411, 0.0568, 0.0422, 0.0483, 0.0300, 0.0558], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0156, 0.0121, 0.0136, 0.0132, 0.0125, 0.0146, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.6270e-05, 1.1497e-04, 8.7723e-05, 9.8639e-05, 9.4303e-05, 9.1793e-05, + 1.0692e-04, 1.0814e-04], device='cuda:4') +2023-03-26 10:14:50,106 INFO [finetune.py:976] (4/7) Epoch 9, batch 600, loss[loss=0.1994, simple_loss=0.2747, pruned_loss=0.06205, over 4811.00 frames. ], tot_loss[loss=0.2011, simple_loss=0.2634, pruned_loss=0.06937, over 908756.63 frames. ], batch size: 41, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:14:54,847 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.149e+02 1.702e+02 1.969e+02 2.390e+02 4.680e+02, threshold=3.938e+02, percent-clipped=3.0 +2023-03-26 10:14:54,940 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46430.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:15:08,730 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9330, 1.6679, 1.5208, 1.5640, 1.6894, 1.5767, 1.6465, 2.3462], + device='cuda:4'), covar=tensor([0.4152, 0.4775, 0.3620, 0.4435, 0.4260, 0.2694, 0.4640, 0.1804], + device='cuda:4'), in_proj_covar=tensor([0.0287, 0.0261, 0.0222, 0.0280, 0.0243, 0.0208, 0.0245, 0.0209], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:15:32,712 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.47 vs. limit=5.0 +2023-03-26 10:15:36,215 INFO [finetune.py:976] (4/7) Epoch 9, batch 650, loss[loss=0.1853, simple_loss=0.2577, pruned_loss=0.05643, over 4826.00 frames. ], tot_loss[loss=0.2053, simple_loss=0.2682, pruned_loss=0.07119, over 917609.58 frames. ], batch size: 25, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:15:40,463 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46478.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:15:51,992 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-26 10:16:05,402 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46515.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:16:09,544 INFO [finetune.py:976] (4/7) Epoch 9, batch 700, loss[loss=0.1676, simple_loss=0.2352, pruned_loss=0.04999, over 4806.00 frames. ], tot_loss[loss=0.2068, simple_loss=0.2702, pruned_loss=0.07165, over 927516.88 frames. ], batch size: 51, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:16:14,888 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.666e+02 2.010e+02 2.529e+02 4.289e+02, threshold=4.019e+02, percent-clipped=2.0 +2023-03-26 10:16:26,204 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0451, 0.7947, 0.9814, 1.0838, 1.2155, 1.1190, 1.0302, 0.9204], + device='cuda:4'), covar=tensor([0.0269, 0.0305, 0.0528, 0.0270, 0.0247, 0.0392, 0.0256, 0.0361], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0110, 0.0139, 0.0115, 0.0102, 0.0101, 0.0091, 0.0108], + device='cuda:4'), out_proj_covar=tensor([6.9912e-05, 8.6186e-05, 1.1125e-04, 9.0722e-05, 8.0368e-05, 7.5266e-05, + 6.8636e-05, 8.3810e-05], device='cuda:4') +2023-03-26 10:16:37,474 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46563.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:16:42,850 INFO [finetune.py:976] (4/7) Epoch 9, batch 750, loss[loss=0.224, simple_loss=0.2832, pruned_loss=0.08242, over 4238.00 frames. ], tot_loss[loss=0.2082, simple_loss=0.2715, pruned_loss=0.07245, over 932667.97 frames. ], batch size: 65, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:16:58,835 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46595.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:17:03,705 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46602.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:17:15,989 INFO [finetune.py:976] (4/7) Epoch 9, batch 800, loss[loss=0.205, simple_loss=0.2494, pruned_loss=0.08033, over 4925.00 frames. ], tot_loss[loss=0.2067, simple_loss=0.2703, pruned_loss=0.0716, over 936274.51 frames. ], batch size: 33, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:17:20,814 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.216e+01 1.561e+02 1.863e+02 2.153e+02 3.377e+02, threshold=3.726e+02, percent-clipped=0.0 +2023-03-26 10:17:22,505 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46632.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:17:39,095 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46656.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 10:17:40,324 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 10:17:43,775 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46663.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:17:49,097 INFO [finetune.py:976] (4/7) Epoch 9, batch 850, loss[loss=0.2028, simple_loss=0.2471, pruned_loss=0.07926, over 3967.00 frames. ], tot_loss[loss=0.2057, simple_loss=0.2686, pruned_loss=0.07138, over 938408.87 frames. ], batch size: 17, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:18:34,822 INFO [finetune.py:976] (4/7) Epoch 9, batch 900, loss[loss=0.1735, simple_loss=0.2321, pruned_loss=0.05742, over 4711.00 frames. ], tot_loss[loss=0.2023, simple_loss=0.2649, pruned_loss=0.06981, over 941920.67 frames. ], batch size: 59, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:18:39,659 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.365e+01 1.501e+02 1.768e+02 2.130e+02 3.855e+02, threshold=3.537e+02, percent-clipped=1.0 +2023-03-26 10:19:10,157 INFO [finetune.py:976] (4/7) Epoch 9, batch 950, loss[loss=0.1385, simple_loss=0.1999, pruned_loss=0.03854, over 4844.00 frames. ], tot_loss[loss=0.2004, simple_loss=0.2624, pruned_loss=0.06918, over 942391.53 frames. ], batch size: 25, lr: 3.80e-03, grad_scale: 32.0 +2023-03-26 10:19:34,828 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8187, 1.6481, 2.2209, 1.3936, 1.9206, 2.0869, 1.6041, 2.3309], + device='cuda:4'), covar=tensor([0.1328, 0.1948, 0.1451, 0.2191, 0.0941, 0.1430, 0.2749, 0.0734], + device='cuda:4'), in_proj_covar=tensor([0.0201, 0.0206, 0.0197, 0.0195, 0.0183, 0.0220, 0.0220, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:19:36,024 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46810.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:19:44,264 INFO [finetune.py:976] (4/7) Epoch 9, batch 1000, loss[loss=0.1921, simple_loss=0.2694, pruned_loss=0.05744, over 4744.00 frames. ], tot_loss[loss=0.2032, simple_loss=0.2661, pruned_loss=0.07011, over 946098.28 frames. ], batch size: 54, lr: 3.80e-03, grad_scale: 32.0 +2023-03-26 10:19:49,056 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.128e+02 1.690e+02 1.992e+02 2.479e+02 5.334e+02, threshold=3.983e+02, percent-clipped=2.0 +2023-03-26 10:20:19,558 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0597, 1.7211, 2.4516, 1.5327, 2.0633, 2.3011, 1.6995, 2.4718], + device='cuda:4'), covar=tensor([0.1391, 0.2236, 0.1530, 0.2333, 0.1043, 0.1561, 0.2724, 0.0794], + device='cuda:4'), in_proj_covar=tensor([0.0201, 0.0206, 0.0197, 0.0195, 0.0183, 0.0220, 0.0220, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:20:22,581 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46871.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:20:23,062 INFO [finetune.py:976] (4/7) Epoch 9, batch 1050, loss[loss=0.2028, simple_loss=0.2615, pruned_loss=0.07201, over 4897.00 frames. ], tot_loss[loss=0.2063, simple_loss=0.2698, pruned_loss=0.07142, over 947968.33 frames. ], batch size: 35, lr: 3.80e-03, grad_scale: 32.0 +2023-03-26 10:21:02,278 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 10:21:05,009 INFO [finetune.py:976] (4/7) Epoch 9, batch 1100, loss[loss=0.1859, simple_loss=0.253, pruned_loss=0.05936, over 4854.00 frames. ], tot_loss[loss=0.2056, simple_loss=0.2695, pruned_loss=0.07083, over 948052.22 frames. ], batch size: 31, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:21:10,488 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.274e+02 1.654e+02 2.044e+02 2.534e+02 3.814e+02, threshold=4.089e+02, percent-clipped=0.0 +2023-03-26 10:21:11,165 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46932.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:21:23,043 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46951.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 10:21:28,249 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46958.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:21:37,702 INFO [finetune.py:976] (4/7) Epoch 9, batch 1150, loss[loss=0.1666, simple_loss=0.2441, pruned_loss=0.04461, over 4827.00 frames. ], tot_loss[loss=0.2062, simple_loss=0.2702, pruned_loss=0.07112, over 949831.88 frames. ], batch size: 47, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:21:42,584 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=46980.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:22:10,618 INFO [finetune.py:976] (4/7) Epoch 9, batch 1200, loss[loss=0.1676, simple_loss=0.2323, pruned_loss=0.05145, over 4804.00 frames. ], tot_loss[loss=0.204, simple_loss=0.2677, pruned_loss=0.07018, over 950113.73 frames. ], batch size: 29, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:22:16,140 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.814e+01 1.625e+02 2.018e+02 2.406e+02 3.989e+02, threshold=4.036e+02, percent-clipped=0.0 +2023-03-26 10:22:18,758 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0184, 1.7849, 1.6183, 1.7282, 1.7129, 1.6906, 1.7027, 2.4561], + device='cuda:4'), covar=tensor([0.4400, 0.4990, 0.3743, 0.4764, 0.4663, 0.2784, 0.4694, 0.1770], + device='cuda:4'), in_proj_covar=tensor([0.0286, 0.0260, 0.0222, 0.0280, 0.0242, 0.0209, 0.0245, 0.0208], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:22:43,588 INFO [finetune.py:976] (4/7) Epoch 9, batch 1250, loss[loss=0.1955, simple_loss=0.2637, pruned_loss=0.06361, over 4892.00 frames. ], tot_loss[loss=0.1998, simple_loss=0.2638, pruned_loss=0.06793, over 952013.23 frames. ], batch size: 35, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:22:48,957 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6199, 1.7770, 1.8402, 1.0071, 1.8157, 2.0404, 1.9549, 1.5361], + device='cuda:4'), covar=tensor([0.0904, 0.0618, 0.0414, 0.0559, 0.0396, 0.0417, 0.0311, 0.0744], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0156, 0.0121, 0.0135, 0.0132, 0.0125, 0.0146, 0.0148], + device='cuda:4'), out_proj_covar=tensor([9.6402e-05, 1.1499e-04, 8.7440e-05, 9.8108e-05, 9.4240e-05, 9.1911e-05, + 1.0730e-04, 1.0852e-04], device='cuda:4') +2023-03-26 10:22:49,792 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.97 vs. limit=5.0 +2023-03-26 10:22:52,605 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-26 10:22:57,778 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.77 vs. limit=5.0 +2023-03-26 10:23:05,539 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-03-26 10:23:21,437 INFO [finetune.py:976] (4/7) Epoch 9, batch 1300, loss[loss=0.2572, simple_loss=0.3057, pruned_loss=0.1044, over 4186.00 frames. ], tot_loss[loss=0.1978, simple_loss=0.261, pruned_loss=0.06733, over 950647.62 frames. ], batch size: 65, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:23:31,785 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.169e+01 1.658e+02 1.956e+02 2.404e+02 5.414e+02, threshold=3.912e+02, percent-clipped=2.0 +2023-03-26 10:23:33,438 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.70 vs. limit=5.0 +2023-03-26 10:23:35,008 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7266, 1.1485, 0.9958, 1.6741, 2.0843, 1.4456, 1.5475, 1.5556], + device='cuda:4'), covar=tensor([0.1467, 0.2304, 0.2057, 0.1234, 0.2029, 0.2190, 0.1495, 0.2090], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0096, 0.0113, 0.0092, 0.0122, 0.0096, 0.0100, 0.0092], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 10:23:58,428 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=47166.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:23:59,659 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 10:24:03,031 INFO [finetune.py:976] (4/7) Epoch 9, batch 1350, loss[loss=0.2304, simple_loss=0.2726, pruned_loss=0.09405, over 4761.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.2625, pruned_loss=0.0686, over 952228.64 frames. ], batch size: 27, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:24:03,380 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 10:24:36,401 INFO [finetune.py:976] (4/7) Epoch 9, batch 1400, loss[loss=0.1447, simple_loss=0.229, pruned_loss=0.03025, over 4819.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.2671, pruned_loss=0.06978, over 952851.06 frames. ], batch size: 39, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:24:42,785 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.739e+02 2.002e+02 2.347e+02 4.228e+02, threshold=4.005e+02, percent-clipped=2.0 +2023-03-26 10:24:49,008 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47241.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:24:49,608 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7029, 1.1076, 0.8720, 1.6722, 2.0485, 1.5691, 1.5166, 1.5417], + device='cuda:4'), covar=tensor([0.1589, 0.2347, 0.2190, 0.1296, 0.2031, 0.2154, 0.1518, 0.2169], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0096, 0.0113, 0.0092, 0.0121, 0.0095, 0.0100, 0.0092], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 10:24:55,111 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=47251.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:24:59,406 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=47258.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:25:09,252 INFO [finetune.py:976] (4/7) Epoch 9, batch 1450, loss[loss=0.2132, simple_loss=0.2801, pruned_loss=0.07314, over 4896.00 frames. ], tot_loss[loss=0.205, simple_loss=0.2689, pruned_loss=0.07054, over 952725.43 frames. ], batch size: 32, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:25:27,569 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=47299.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:25:29,412 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47302.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:25:36,208 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=47306.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:25:53,869 INFO [finetune.py:976] (4/7) Epoch 9, batch 1500, loss[loss=0.1947, simple_loss=0.2669, pruned_loss=0.06127, over 4783.00 frames. ], tot_loss[loss=0.2074, simple_loss=0.2714, pruned_loss=0.07164, over 953919.21 frames. ], batch size: 28, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:26:00,804 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.103e+02 1.648e+02 1.983e+02 2.305e+02 4.092e+02, threshold=3.967e+02, percent-clipped=1.0 +2023-03-26 10:26:06,402 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 10:26:18,934 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5841, 3.4208, 3.2207, 1.4820, 3.5484, 2.7312, 0.8116, 2.3470], + device='cuda:4'), covar=tensor([0.2271, 0.2413, 0.1752, 0.3474, 0.1140, 0.1050, 0.4389, 0.1589], + device='cuda:4'), in_proj_covar=tensor([0.0154, 0.0173, 0.0160, 0.0129, 0.0155, 0.0122, 0.0146, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 10:26:26,875 INFO [finetune.py:976] (4/7) Epoch 9, batch 1550, loss[loss=0.2061, simple_loss=0.2604, pruned_loss=0.07592, over 4833.00 frames. ], tot_loss[loss=0.2058, simple_loss=0.2701, pruned_loss=0.07081, over 954928.45 frames. ], batch size: 49, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:26:33,213 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47379.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:26:44,905 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9644, 1.7711, 1.7697, 1.9042, 1.9240, 4.5700, 1.9209, 2.4429], + device='cuda:4'), covar=tensor([0.3133, 0.2428, 0.2020, 0.2246, 0.1472, 0.0090, 0.2217, 0.1143], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0115, 0.0119, 0.0123, 0.0116, 0.0098, 0.0100, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 10:27:00,841 INFO [finetune.py:976] (4/7) Epoch 9, batch 1600, loss[loss=0.28, simple_loss=0.3178, pruned_loss=0.1211, over 4820.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.2672, pruned_loss=0.06981, over 955913.50 frames. ], batch size: 39, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:27:07,814 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.155e+02 1.632e+02 1.991e+02 2.321e+02 5.028e+02, threshold=3.982e+02, percent-clipped=1.0 +2023-03-26 10:27:14,451 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47440.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:27:30,689 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=47466.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:27:34,231 INFO [finetune.py:976] (4/7) Epoch 9, batch 1650, loss[loss=0.2081, simple_loss=0.2708, pruned_loss=0.07275, over 4898.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.2631, pruned_loss=0.06783, over 956802.00 frames. ], batch size: 32, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:28:02,974 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=47514.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:28:07,771 INFO [finetune.py:976] (4/7) Epoch 9, batch 1700, loss[loss=0.1315, simple_loss=0.2067, pruned_loss=0.02817, over 4764.00 frames. ], tot_loss[loss=0.1977, simple_loss=0.2609, pruned_loss=0.06724, over 957953.15 frames. ], batch size: 27, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:28:13,223 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.617e+02 1.864e+02 2.209e+02 5.015e+02, threshold=3.728e+02, percent-clipped=2.0 +2023-03-26 10:28:47,737 INFO [finetune.py:976] (4/7) Epoch 9, batch 1750, loss[loss=0.2105, simple_loss=0.2779, pruned_loss=0.07158, over 4916.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.2642, pruned_loss=0.06881, over 957470.13 frames. ], batch size: 36, lr: 3.80e-03, grad_scale: 16.0 +2023-03-26 10:29:13,718 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=47597.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:29:24,084 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0862, 1.6492, 2.3116, 1.4256, 2.1257, 2.2159, 1.6933, 2.4868], + device='cuda:4'), covar=tensor([0.1254, 0.2182, 0.1433, 0.2247, 0.0875, 0.1381, 0.2739, 0.0793], + device='cuda:4'), in_proj_covar=tensor([0.0201, 0.0206, 0.0197, 0.0195, 0.0181, 0.0220, 0.0219, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:29:29,377 INFO [finetune.py:976] (4/7) Epoch 9, batch 1800, loss[loss=0.2251, simple_loss=0.2869, pruned_loss=0.08169, over 4817.00 frames. ], tot_loss[loss=0.2044, simple_loss=0.2682, pruned_loss=0.0703, over 955386.14 frames. ], batch size: 38, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:29:34,861 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.172e+02 1.736e+02 1.994e+02 2.573e+02 6.193e+02, threshold=3.989e+02, percent-clipped=4.0 +2023-03-26 10:30:02,647 INFO [finetune.py:976] (4/7) Epoch 9, batch 1850, loss[loss=0.1997, simple_loss=0.2604, pruned_loss=0.06954, over 4904.00 frames. ], tot_loss[loss=0.2045, simple_loss=0.268, pruned_loss=0.07048, over 952366.46 frames. ], batch size: 37, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:30:35,896 INFO [finetune.py:976] (4/7) Epoch 9, batch 1900, loss[loss=0.1857, simple_loss=0.2508, pruned_loss=0.06035, over 4736.00 frames. ], tot_loss[loss=0.2052, simple_loss=0.2693, pruned_loss=0.07059, over 953259.54 frames. ], batch size: 54, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:30:46,260 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.162e+02 1.608e+02 1.835e+02 2.236e+02 3.803e+02, threshold=3.670e+02, percent-clipped=0.0 +2023-03-26 10:30:53,008 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=47735.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:30:54,875 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47738.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:31:21,752 INFO [finetune.py:976] (4/7) Epoch 9, batch 1950, loss[loss=0.1712, simple_loss=0.2529, pruned_loss=0.04476, over 4852.00 frames. ], tot_loss[loss=0.2048, simple_loss=0.269, pruned_loss=0.07028, over 955181.81 frames. ], batch size: 34, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:31:31,524 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47788.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:31:39,613 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47799.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:31:46,710 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47809.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 10:31:55,009 INFO [finetune.py:976] (4/7) Epoch 9, batch 2000, loss[loss=0.1863, simple_loss=0.2563, pruned_loss=0.05816, over 4829.00 frames. ], tot_loss[loss=0.2032, simple_loss=0.2669, pruned_loss=0.06972, over 957137.23 frames. ], batch size: 33, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:32:00,456 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.198e+02 1.527e+02 1.824e+02 2.186e+02 3.277e+02, threshold=3.648e+02, percent-clipped=0.0 +2023-03-26 10:32:10,221 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0036, 1.9935, 1.9576, 1.3724, 2.1670, 2.1423, 2.1098, 1.8217], + device='cuda:4'), covar=tensor([0.0516, 0.0540, 0.0681, 0.0837, 0.0544, 0.0703, 0.0552, 0.0960], + device='cuda:4'), in_proj_covar=tensor([0.0135, 0.0131, 0.0142, 0.0123, 0.0115, 0.0143, 0.0143, 0.0159], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:32:11,904 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47849.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:32:35,741 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47870.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 10:32:36,818 INFO [finetune.py:976] (4/7) Epoch 9, batch 2050, loss[loss=0.2275, simple_loss=0.2725, pruned_loss=0.09125, over 4706.00 frames. ], tot_loss[loss=0.1997, simple_loss=0.2628, pruned_loss=0.06831, over 958564.81 frames. ], batch size: 23, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:32:39,426 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5593, 1.4713, 1.4592, 1.5032, 1.0798, 2.9401, 1.0879, 1.5315], + device='cuda:4'), covar=tensor([0.3412, 0.2483, 0.2150, 0.2360, 0.1947, 0.0256, 0.2632, 0.1370], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0115, 0.0120, 0.0123, 0.0117, 0.0098, 0.0100, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 10:32:57,818 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=47897.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:33:15,823 INFO [finetune.py:976] (4/7) Epoch 9, batch 2100, loss[loss=0.189, simple_loss=0.2553, pruned_loss=0.06139, over 4873.00 frames. ], tot_loss[loss=0.2004, simple_loss=0.2636, pruned_loss=0.06858, over 959409.09 frames. ], batch size: 34, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:33:21,286 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.115e+02 1.604e+02 1.917e+02 2.370e+02 5.169e+02, threshold=3.834e+02, percent-clipped=3.0 +2023-03-26 10:33:29,792 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=47945.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:33:54,984 INFO [finetune.py:976] (4/7) Epoch 9, batch 2150, loss[loss=0.2675, simple_loss=0.3261, pruned_loss=0.1044, over 4817.00 frames. ], tot_loss[loss=0.2023, simple_loss=0.2663, pruned_loss=0.06915, over 960074.29 frames. ], batch size: 38, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:34:06,549 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8480, 1.4160, 1.9057, 1.7115, 1.4963, 1.5520, 1.6354, 1.6965], + device='cuda:4'), covar=tensor([0.4444, 0.5207, 0.4144, 0.4807, 0.5790, 0.4658, 0.6073, 0.4108], + device='cuda:4'), in_proj_covar=tensor([0.0233, 0.0241, 0.0254, 0.0256, 0.0249, 0.0225, 0.0274, 0.0230], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:35:00,856 INFO [finetune.py:976] (4/7) Epoch 9, batch 2200, loss[loss=0.1841, simple_loss=0.2541, pruned_loss=0.05707, over 4815.00 frames. ], tot_loss[loss=0.203, simple_loss=0.2676, pruned_loss=0.06921, over 958280.40 frames. ], batch size: 39, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:35:11,848 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.338e+02 1.754e+02 2.052e+02 2.528e+02 4.321e+02, threshold=4.105e+02, percent-clipped=1.0 +2023-03-26 10:35:18,708 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48035.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:35:29,778 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4590, 1.0035, 0.7701, 1.3492, 1.9096, 0.7302, 1.2389, 1.3684], + device='cuda:4'), covar=tensor([0.1580, 0.2246, 0.1800, 0.1285, 0.2001, 0.2275, 0.1533, 0.2146], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0112, 0.0091, 0.0121, 0.0095, 0.0098, 0.0091], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 10:35:52,145 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4549, 1.4338, 1.5237, 0.7950, 1.5490, 1.5551, 1.4590, 1.3050], + device='cuda:4'), covar=tensor([0.0632, 0.0839, 0.0764, 0.0992, 0.0789, 0.0776, 0.0701, 0.1320], + device='cuda:4'), in_proj_covar=tensor([0.0136, 0.0132, 0.0144, 0.0125, 0.0117, 0.0144, 0.0145, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:36:02,991 INFO [finetune.py:976] (4/7) Epoch 9, batch 2250, loss[loss=0.2766, simple_loss=0.3255, pruned_loss=0.1138, over 4886.00 frames. ], tot_loss[loss=0.2051, simple_loss=0.2696, pruned_loss=0.07032, over 958786.29 frames. ], batch size: 32, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:36:03,102 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48072.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:36:20,660 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48083.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:36:24,389 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-03-26 10:36:32,288 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48094.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:37:05,451 INFO [finetune.py:976] (4/7) Epoch 9, batch 2300, loss[loss=0.1819, simple_loss=0.2564, pruned_loss=0.05371, over 4870.00 frames. ], tot_loss[loss=0.204, simple_loss=0.2688, pruned_loss=0.0696, over 957420.37 frames. ], batch size: 34, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:37:12,528 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 10:37:14,981 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7840, 1.6854, 1.5477, 1.9588, 2.1245, 1.9162, 1.3065, 1.5352], + device='cuda:4'), covar=tensor([0.2366, 0.2204, 0.2005, 0.1647, 0.1676, 0.1212, 0.2817, 0.1957], + device='cuda:4'), in_proj_covar=tensor([0.0240, 0.0210, 0.0207, 0.0188, 0.0242, 0.0180, 0.0216, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:37:16,654 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.508e+01 1.652e+02 1.874e+02 2.360e+02 5.580e+02, threshold=3.748e+02, percent-clipped=1.0 +2023-03-26 10:37:23,193 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48133.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:37:34,217 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48144.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:37:54,848 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48165.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 10:38:01,014 INFO [finetune.py:976] (4/7) Epoch 9, batch 2350, loss[loss=0.165, simple_loss=0.2343, pruned_loss=0.04781, over 4751.00 frames. ], tot_loss[loss=0.2012, simple_loss=0.2654, pruned_loss=0.0685, over 956332.00 frames. ], batch size: 23, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:38:35,597 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48212.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 10:38:43,033 INFO [finetune.py:976] (4/7) Epoch 9, batch 2400, loss[loss=0.1983, simple_loss=0.2632, pruned_loss=0.06668, over 4830.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.263, pruned_loss=0.06816, over 958019.78 frames. ], batch size: 33, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:38:49,470 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.159e+02 1.602e+02 2.015e+02 2.421e+02 3.465e+02, threshold=4.031e+02, percent-clipped=0.0 +2023-03-26 10:39:18,902 INFO [finetune.py:976] (4/7) Epoch 9, batch 2450, loss[loss=0.2034, simple_loss=0.2656, pruned_loss=0.07059, over 4842.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.2597, pruned_loss=0.06692, over 957643.35 frames. ], batch size: 47, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:39:19,648 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48273.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 10:39:30,299 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.24 vs. limit=5.0 +2023-03-26 10:40:01,691 INFO [finetune.py:976] (4/7) Epoch 9, batch 2500, loss[loss=0.2337, simple_loss=0.307, pruned_loss=0.08023, over 4824.00 frames. ], tot_loss[loss=0.2005, simple_loss=0.2636, pruned_loss=0.0687, over 957093.69 frames. ], batch size: 40, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:40:03,405 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48324.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:40:09,017 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.093e+02 1.681e+02 1.962e+02 2.420e+02 5.026e+02, threshold=3.923e+02, percent-clipped=2.0 +2023-03-26 10:40:11,408 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9492, 1.8376, 1.5521, 1.7507, 1.7403, 1.7258, 1.7012, 2.4715], + device='cuda:4'), covar=tensor([0.4744, 0.5075, 0.3904, 0.4959, 0.4829, 0.2911, 0.4995, 0.1884], + device='cuda:4'), in_proj_covar=tensor([0.0287, 0.0260, 0.0222, 0.0280, 0.0243, 0.0209, 0.0245, 0.0210], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:40:24,532 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7745, 1.4720, 2.1319, 3.4638, 2.2425, 2.5403, 0.8543, 2.8284], + device='cuda:4'), covar=tensor([0.1854, 0.1710, 0.1528, 0.0866, 0.0933, 0.1377, 0.2355, 0.0610], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0116, 0.0132, 0.0164, 0.0102, 0.0138, 0.0125, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 10:40:35,357 INFO [finetune.py:976] (4/7) Epoch 9, batch 2550, loss[loss=0.1765, simple_loss=0.2575, pruned_loss=0.04778, over 4922.00 frames. ], tot_loss[loss=0.2041, simple_loss=0.268, pruned_loss=0.07009, over 956430.15 frames. ], batch size: 38, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:40:45,877 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48385.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:40:51,848 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48394.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:40:53,631 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48397.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:41:02,901 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5811, 1.4660, 1.4401, 1.5876, 1.1076, 3.0303, 1.1145, 1.6072], + device='cuda:4'), covar=tensor([0.3511, 0.2591, 0.2214, 0.2345, 0.1975, 0.0231, 0.2873, 0.1360], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0115, 0.0120, 0.0123, 0.0117, 0.0098, 0.0100, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 10:41:08,897 INFO [finetune.py:976] (4/7) Epoch 9, batch 2600, loss[loss=0.2166, simple_loss=0.2863, pruned_loss=0.07348, over 4916.00 frames. ], tot_loss[loss=0.2045, simple_loss=0.2685, pruned_loss=0.0702, over 954329.95 frames. ], batch size: 36, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:41:09,620 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48423.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:41:12,598 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48428.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:41:15,191 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.166e+02 1.754e+02 2.041e+02 2.553e+02 4.015e+02, threshold=4.083e+02, percent-clipped=1.0 +2023-03-26 10:41:23,751 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48442.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:41:25,462 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48444.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:41:34,012 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48458.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:41:38,229 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48465.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 10:41:39,454 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.8834, 3.3847, 3.5648, 3.7465, 3.6132, 3.4098, 3.9325, 1.2042], + device='cuda:4'), covar=tensor([0.0882, 0.0897, 0.0954, 0.0998, 0.1424, 0.1738, 0.0861, 0.5533], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0244, 0.0275, 0.0291, 0.0330, 0.0281, 0.0300, 0.0295], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:41:42,380 INFO [finetune.py:976] (4/7) Epoch 9, batch 2650, loss[loss=0.1871, simple_loss=0.2495, pruned_loss=0.0624, over 4302.00 frames. ], tot_loss[loss=0.2052, simple_loss=0.2694, pruned_loss=0.07048, over 953667.05 frames. ], batch size: 65, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:41:56,001 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48484.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:42:06,172 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48492.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:42:19,480 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48513.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 10:42:24,896 INFO [finetune.py:976] (4/7) Epoch 9, batch 2700, loss[loss=0.1675, simple_loss=0.2361, pruned_loss=0.04943, over 4900.00 frames. ], tot_loss[loss=0.2019, simple_loss=0.2666, pruned_loss=0.06864, over 955546.32 frames. ], batch size: 32, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:42:30,335 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.052e+02 1.590e+02 1.873e+02 2.487e+02 4.580e+02, threshold=3.745e+02, percent-clipped=2.0 +2023-03-26 10:42:47,461 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7060, 1.1910, 0.8943, 1.6467, 2.1556, 1.4934, 1.4656, 1.7717], + device='cuda:4'), covar=tensor([0.1474, 0.2129, 0.2097, 0.1168, 0.1911, 0.1983, 0.1531, 0.1927], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0092, 0.0122, 0.0095, 0.0099, 0.0091], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 10:43:07,960 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48568.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 10:43:10,318 INFO [finetune.py:976] (4/7) Epoch 9, batch 2750, loss[loss=0.1793, simple_loss=0.2497, pruned_loss=0.05446, over 4816.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.2643, pruned_loss=0.06873, over 955889.09 frames. ], batch size: 40, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:43:32,795 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9748, 1.8688, 1.9875, 1.5353, 2.1118, 2.2869, 2.0875, 1.4625], + device='cuda:4'), covar=tensor([0.0640, 0.0734, 0.0759, 0.0914, 0.0573, 0.0607, 0.0694, 0.1598], + device='cuda:4'), in_proj_covar=tensor([0.0136, 0.0132, 0.0143, 0.0124, 0.0117, 0.0143, 0.0144, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:43:40,975 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 10:43:45,668 INFO [finetune.py:976] (4/7) Epoch 9, batch 2800, loss[loss=0.1928, simple_loss=0.2562, pruned_loss=0.06468, over 4911.00 frames. ], tot_loss[loss=0.197, simple_loss=0.2602, pruned_loss=0.06684, over 954373.65 frames. ], batch size: 32, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:43:51,108 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.018e+02 1.496e+02 1.797e+02 2.211e+02 4.995e+02, threshold=3.593e+02, percent-clipped=1.0 +2023-03-26 10:43:52,458 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3640, 1.9933, 2.7804, 1.6440, 2.4530, 2.5609, 1.9290, 2.6767], + device='cuda:4'), covar=tensor([0.1211, 0.1772, 0.1372, 0.2268, 0.0797, 0.1449, 0.2291, 0.0853], + device='cuda:4'), in_proj_covar=tensor([0.0200, 0.0204, 0.0195, 0.0193, 0.0181, 0.0219, 0.0217, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:43:56,374 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-26 10:44:19,115 INFO [finetune.py:976] (4/7) Epoch 9, batch 2850, loss[loss=0.1744, simple_loss=0.2375, pruned_loss=0.05565, over 4771.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2598, pruned_loss=0.06671, over 954959.47 frames. ], batch size: 26, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:44:24,057 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48680.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:44:39,320 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.81 vs. limit=5.0 +2023-03-26 10:45:04,574 INFO [finetune.py:976] (4/7) Epoch 9, batch 2900, loss[loss=0.1771, simple_loss=0.2532, pruned_loss=0.05048, over 4759.00 frames. ], tot_loss[loss=0.1995, simple_loss=0.263, pruned_loss=0.06801, over 954301.16 frames. ], batch size: 26, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:45:08,314 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48728.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:45:10,030 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.274e+02 1.660e+02 1.926e+02 2.355e+02 4.281e+02, threshold=3.853e+02, percent-clipped=2.0 +2023-03-26 10:45:16,923 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4171, 2.2265, 1.7433, 2.2291, 2.1479, 1.8858, 2.7322, 2.3920], + device='cuda:4'), covar=tensor([0.1321, 0.2718, 0.3508, 0.3419, 0.3118, 0.1806, 0.4488, 0.1942], + device='cuda:4'), in_proj_covar=tensor([0.0174, 0.0189, 0.0234, 0.0254, 0.0237, 0.0195, 0.0212, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:45:25,495 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48753.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:45:30,934 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-26 10:45:38,485 INFO [finetune.py:976] (4/7) Epoch 9, batch 2950, loss[loss=0.2057, simple_loss=0.2833, pruned_loss=0.06406, over 4830.00 frames. ], tot_loss[loss=0.2014, simple_loss=0.2658, pruned_loss=0.06848, over 955123.51 frames. ], batch size: 33, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:45:40,980 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48776.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:45:42,826 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48779.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:45:46,114 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-26 10:46:04,007 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 10:46:10,721 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0331, 2.0021, 2.1066, 1.5820, 2.0769, 2.2854, 2.1662, 1.6496], + device='cuda:4'), covar=tensor([0.0541, 0.0519, 0.0591, 0.0795, 0.0681, 0.0520, 0.0499, 0.1019], + device='cuda:4'), in_proj_covar=tensor([0.0135, 0.0131, 0.0142, 0.0122, 0.0116, 0.0142, 0.0142, 0.0159], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:46:11,320 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48821.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:46:11,795 INFO [finetune.py:976] (4/7) Epoch 9, batch 3000, loss[loss=0.1942, simple_loss=0.2555, pruned_loss=0.06643, over 4817.00 frames. ], tot_loss[loss=0.204, simple_loss=0.2684, pruned_loss=0.0698, over 954762.40 frames. ], batch size: 38, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:46:11,795 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 10:46:22,396 INFO [finetune.py:1010] (4/7) Epoch 9, validation: loss=0.159, simple_loss=0.2302, pruned_loss=0.04393, over 2265189.00 frames. +2023-03-26 10:46:22,397 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 10:46:27,910 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.631e+02 1.894e+02 2.277e+02 3.777e+02, threshold=3.789e+02, percent-clipped=0.0 +2023-03-26 10:46:32,257 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48838.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:46:52,190 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48868.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 10:46:54,487 INFO [finetune.py:976] (4/7) Epoch 9, batch 3050, loss[loss=0.1885, simple_loss=0.2488, pruned_loss=0.06412, over 4791.00 frames. ], tot_loss[loss=0.2032, simple_loss=0.2677, pruned_loss=0.06936, over 953830.21 frames. ], batch size: 54, lr: 3.79e-03, grad_scale: 16.0 +2023-03-26 10:47:03,441 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48882.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:47:07,885 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.50 vs. limit=5.0 +2023-03-26 10:47:13,622 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48899.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:47:23,637 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6089, 1.5469, 1.3524, 1.5574, 1.8666, 1.8176, 1.6405, 1.3818], + device='cuda:4'), covar=tensor([0.0337, 0.0338, 0.0582, 0.0325, 0.0204, 0.0474, 0.0290, 0.0400], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0110, 0.0140, 0.0115, 0.0102, 0.0101, 0.0091, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.0443e-05, 8.5800e-05, 1.1145e-04, 8.9910e-05, 7.9921e-05, 7.5079e-05, + 6.8687e-05, 8.3515e-05], device='cuda:4') +2023-03-26 10:47:24,761 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=48916.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 10:47:29,283 INFO [finetune.py:976] (4/7) Epoch 9, batch 3100, loss[loss=0.2059, simple_loss=0.2595, pruned_loss=0.07616, over 4763.00 frames. ], tot_loss[loss=0.202, simple_loss=0.2661, pruned_loss=0.069, over 953296.82 frames. ], batch size: 54, lr: 3.79e-03, grad_scale: 32.0 +2023-03-26 10:47:31,146 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.99 vs. limit=5.0 +2023-03-26 10:47:36,136 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.216e+02 1.624e+02 1.916e+02 2.206e+02 4.881e+02, threshold=3.833e+02, percent-clipped=2.0 +2023-03-26 10:47:44,106 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8750, 1.8378, 1.6846, 2.1312, 2.2466, 2.2185, 1.6220, 1.5658], + device='cuda:4'), covar=tensor([0.2188, 0.1949, 0.1849, 0.1580, 0.1870, 0.1019, 0.2554, 0.1958], + device='cuda:4'), in_proj_covar=tensor([0.0237, 0.0207, 0.0205, 0.0187, 0.0240, 0.0179, 0.0214, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:48:07,638 INFO [finetune.py:976] (4/7) Epoch 9, batch 3150, loss[loss=0.1883, simple_loss=0.254, pruned_loss=0.06128, over 4895.00 frames. ], tot_loss[loss=0.2008, simple_loss=0.2642, pruned_loss=0.06874, over 955709.13 frames. ], batch size: 32, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:48:16,689 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48980.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:48:49,389 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7049, 1.6368, 2.0198, 1.3861, 1.6744, 1.8163, 1.5429, 2.0982], + device='cuda:4'), covar=tensor([0.1152, 0.1761, 0.1107, 0.1647, 0.0908, 0.1270, 0.2444, 0.0748], + device='cuda:4'), in_proj_covar=tensor([0.0198, 0.0204, 0.0193, 0.0191, 0.0179, 0.0216, 0.0215, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:48:51,004 INFO [finetune.py:976] (4/7) Epoch 9, batch 3200, loss[loss=0.1625, simple_loss=0.233, pruned_loss=0.046, over 4786.00 frames. ], tot_loss[loss=0.1974, simple_loss=0.2602, pruned_loss=0.06727, over 953970.96 frames. ], batch size: 26, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:48:55,661 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49028.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:48:57,864 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.147e+02 1.665e+02 1.973e+02 2.326e+02 6.022e+02, threshold=3.945e+02, percent-clipped=4.0 +2023-03-26 10:49:12,701 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49053.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:49:14,843 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 10:49:30,124 INFO [finetune.py:976] (4/7) Epoch 9, batch 3250, loss[loss=0.1976, simple_loss=0.2619, pruned_loss=0.0667, over 4913.00 frames. ], tot_loss[loss=0.199, simple_loss=0.2617, pruned_loss=0.06819, over 953673.16 frames. ], batch size: 36, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:49:40,140 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49079.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:50:05,719 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49101.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:50:21,990 INFO [finetune.py:976] (4/7) Epoch 9, batch 3300, loss[loss=0.2361, simple_loss=0.2989, pruned_loss=0.08662, over 4863.00 frames. ], tot_loss[loss=0.2035, simple_loss=0.2669, pruned_loss=0.07007, over 953865.66 frames. ], batch size: 31, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:50:26,597 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49127.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:50:28,553 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.55 vs. limit=5.0 +2023-03-26 10:50:28,941 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.155e+02 1.707e+02 1.914e+02 2.346e+02 3.542e+02, threshold=3.827e+02, percent-clipped=0.0 +2023-03-26 10:50:56,006 INFO [finetune.py:976] (4/7) Epoch 9, batch 3350, loss[loss=0.2131, simple_loss=0.2838, pruned_loss=0.0712, over 4897.00 frames. ], tot_loss[loss=0.2031, simple_loss=0.2674, pruned_loss=0.06938, over 954566.99 frames. ], batch size: 36, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:50:58,580 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2920, 2.1219, 2.0102, 2.3517, 2.9418, 2.3245, 2.0628, 1.7231], + device='cuda:4'), covar=tensor([0.2261, 0.2200, 0.1870, 0.1732, 0.1790, 0.1057, 0.2214, 0.2010], + device='cuda:4'), in_proj_covar=tensor([0.0238, 0.0209, 0.0207, 0.0188, 0.0241, 0.0180, 0.0214, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:50:59,114 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49177.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:51:17,578 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49194.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:51:42,061 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=5.11 vs. limit=5.0 +2023-03-26 10:51:50,174 INFO [finetune.py:976] (4/7) Epoch 9, batch 3400, loss[loss=0.1978, simple_loss=0.2739, pruned_loss=0.06081, over 4816.00 frames. ], tot_loss[loss=0.2041, simple_loss=0.2685, pruned_loss=0.06981, over 953228.68 frames. ], batch size: 33, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:51:59,826 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.353e+01 1.606e+02 1.878e+02 2.295e+02 4.525e+02, threshold=3.756e+02, percent-clipped=2.0 +2023-03-26 10:52:10,022 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49236.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:52:40,930 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8289, 1.6090, 2.1680, 1.4338, 1.7857, 2.0647, 1.5162, 2.2334], + device='cuda:4'), covar=tensor([0.1370, 0.2062, 0.1313, 0.2051, 0.0999, 0.1434, 0.2736, 0.0901], + device='cuda:4'), in_proj_covar=tensor([0.0199, 0.0205, 0.0194, 0.0193, 0.0181, 0.0219, 0.0218, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:52:55,270 INFO [finetune.py:976] (4/7) Epoch 9, batch 3450, loss[loss=0.1919, simple_loss=0.2577, pruned_loss=0.06302, over 4867.00 frames. ], tot_loss[loss=0.2016, simple_loss=0.2667, pruned_loss=0.06824, over 954729.73 frames. ], batch size: 31, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:53:32,520 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49297.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:53:32,841 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 10:53:53,350 INFO [finetune.py:976] (4/7) Epoch 9, batch 3500, loss[loss=0.1958, simple_loss=0.2617, pruned_loss=0.06499, over 4817.00 frames. ], tot_loss[loss=0.1995, simple_loss=0.2641, pruned_loss=0.06747, over 956134.83 frames. ], batch size: 40, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:53:58,772 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.331e+01 1.641e+02 1.916e+02 2.289e+02 6.335e+02, threshold=3.833e+02, percent-clipped=2.0 +2023-03-26 10:54:32,749 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-26 10:54:34,216 INFO [finetune.py:976] (4/7) Epoch 9, batch 3550, loss[loss=0.2071, simple_loss=0.2678, pruned_loss=0.07319, over 4820.00 frames. ], tot_loss[loss=0.1975, simple_loss=0.2615, pruned_loss=0.06678, over 956154.80 frames. ], batch size: 51, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:55:09,813 INFO [finetune.py:976] (4/7) Epoch 9, batch 3600, loss[loss=0.2303, simple_loss=0.291, pruned_loss=0.08482, over 4903.00 frames. ], tot_loss[loss=0.1981, simple_loss=0.2613, pruned_loss=0.06742, over 956930.04 frames. ], batch size: 36, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:55:09,933 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49422.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:55:10,543 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1344, 1.9184, 2.0930, 0.8419, 2.3264, 2.5227, 2.0959, 1.9598], + device='cuda:4'), covar=tensor([0.0825, 0.0662, 0.0449, 0.0719, 0.0419, 0.0505, 0.0399, 0.0604], + device='cuda:4'), in_proj_covar=tensor([0.0128, 0.0155, 0.0121, 0.0134, 0.0131, 0.0125, 0.0145, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.5212e-05, 1.1410e-04, 8.7053e-05, 9.7567e-05, 9.3595e-05, 9.1443e-05, + 1.0637e-04, 1.0779e-04], device='cuda:4') +2023-03-26 10:55:15,220 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.662e+02 2.002e+02 2.382e+02 4.044e+02, threshold=4.004e+02, percent-clipped=1.0 +2023-03-26 10:55:15,970 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49432.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 10:55:27,916 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0748, 1.5869, 2.3981, 3.9736, 2.7575, 2.8690, 0.6960, 3.1740], + device='cuda:4'), covar=tensor([0.1857, 0.1849, 0.1623, 0.0697, 0.0846, 0.1476, 0.2440, 0.0532], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0118, 0.0135, 0.0167, 0.0103, 0.0140, 0.0127, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 10:55:38,427 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0569, 1.9213, 1.7036, 2.0466, 1.8896, 1.8792, 1.8344, 2.7268], + device='cuda:4'), covar=tensor([0.4871, 0.6039, 0.4012, 0.5758, 0.5452, 0.3001, 0.5813, 0.1787], + device='cuda:4'), in_proj_covar=tensor([0.0286, 0.0260, 0.0221, 0.0280, 0.0242, 0.0209, 0.0245, 0.0211], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:55:43,180 INFO [finetune.py:976] (4/7) Epoch 9, batch 3650, loss[loss=0.1902, simple_loss=0.2656, pruned_loss=0.0574, over 4747.00 frames. ], tot_loss[loss=0.1995, simple_loss=0.2631, pruned_loss=0.06791, over 956650.26 frames. ], batch size: 54, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:55:46,392 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49477.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:55:50,111 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49483.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:55:56,193 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49493.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 10:55:56,752 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49494.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:56:17,056 INFO [finetune.py:976] (4/7) Epoch 9, batch 3700, loss[loss=0.1896, simple_loss=0.2603, pruned_loss=0.0595, over 4756.00 frames. ], tot_loss[loss=0.2016, simple_loss=0.2655, pruned_loss=0.06884, over 955675.15 frames. ], batch size: 27, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:56:18,953 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49525.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:56:22,513 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.782e+02 2.076e+02 2.384e+02 4.659e+02, threshold=4.152e+02, percent-clipped=5.0 +2023-03-26 10:56:29,220 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49542.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:56:50,563 INFO [finetune.py:976] (4/7) Epoch 9, batch 3750, loss[loss=0.1986, simple_loss=0.2677, pruned_loss=0.06478, over 4829.00 frames. ], tot_loss[loss=0.2033, simple_loss=0.2676, pruned_loss=0.0695, over 955799.08 frames. ], batch size: 30, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:56:54,382 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8746, 1.7328, 1.4423, 1.5602, 1.6780, 1.5835, 1.6728, 2.3312], + device='cuda:4'), covar=tensor([0.5176, 0.5629, 0.4008, 0.4986, 0.4821, 0.2958, 0.4795, 0.2025], + device='cuda:4'), in_proj_covar=tensor([0.0286, 0.0259, 0.0221, 0.0280, 0.0242, 0.0208, 0.0245, 0.0211], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:56:54,430 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 10:57:02,695 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49592.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:57:20,772 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-03-26 10:57:28,148 INFO [finetune.py:976] (4/7) Epoch 9, batch 3800, loss[loss=0.1938, simple_loss=0.2651, pruned_loss=0.06121, over 4912.00 frames. ], tot_loss[loss=0.2036, simple_loss=0.2686, pruned_loss=0.0693, over 956381.89 frames. ], batch size: 33, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:57:29,382 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3820, 2.2962, 2.8896, 1.7588, 2.5416, 2.7490, 2.1003, 2.8869], + device='cuda:4'), covar=tensor([0.1756, 0.2127, 0.1600, 0.2515, 0.1136, 0.1780, 0.2546, 0.0999], + device='cuda:4'), in_proj_covar=tensor([0.0200, 0.0206, 0.0195, 0.0193, 0.0180, 0.0218, 0.0217, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:57:39,044 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.635e+02 1.863e+02 2.259e+02 4.048e+02, threshold=3.725e+02, percent-clipped=0.0 +2023-03-26 10:58:12,409 INFO [finetune.py:976] (4/7) Epoch 9, batch 3850, loss[loss=0.1782, simple_loss=0.2476, pruned_loss=0.05443, over 4810.00 frames. ], tot_loss[loss=0.2003, simple_loss=0.2656, pruned_loss=0.06748, over 957347.66 frames. ], batch size: 33, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:58:28,255 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8239, 1.5900, 1.4355, 1.1753, 1.6492, 1.5655, 1.5714, 2.1417], + device='cuda:4'), covar=tensor([0.4514, 0.5009, 0.3472, 0.4682, 0.4245, 0.2508, 0.4093, 0.2012], + device='cuda:4'), in_proj_covar=tensor([0.0287, 0.0261, 0.0222, 0.0282, 0.0243, 0.0209, 0.0246, 0.0211], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 10:58:38,347 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4938, 1.3983, 2.0289, 3.1839, 2.0900, 2.3509, 0.9705, 2.4552], + device='cuda:4'), covar=tensor([0.1716, 0.1430, 0.1189, 0.0500, 0.0764, 0.1356, 0.1713, 0.0536], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0119, 0.0135, 0.0166, 0.0103, 0.0140, 0.0127, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 10:58:48,066 INFO [finetune.py:976] (4/7) Epoch 9, batch 3900, loss[loss=0.1854, simple_loss=0.2462, pruned_loss=0.06234, over 4911.00 frames. ], tot_loss[loss=0.1978, simple_loss=0.2625, pruned_loss=0.0665, over 959526.10 frames. ], batch size: 35, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:58:58,258 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.638e+02 1.913e+02 2.415e+02 4.821e+02, threshold=3.825e+02, percent-clipped=2.0 +2023-03-26 10:59:31,485 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49766.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:59:36,466 INFO [finetune.py:976] (4/7) Epoch 9, batch 3950, loss[loss=0.1811, simple_loss=0.2349, pruned_loss=0.06369, over 4858.00 frames. ], tot_loss[loss=0.1959, simple_loss=0.2598, pruned_loss=0.06595, over 958518.45 frames. ], batch size: 44, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 10:59:40,656 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49778.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 10:59:47,183 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49788.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 10:59:56,611 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6733, 1.4099, 0.9154, 0.2257, 1.1951, 1.4984, 1.3214, 1.3491], + device='cuda:4'), covar=tensor([0.0847, 0.0825, 0.1271, 0.1829, 0.1389, 0.2287, 0.2487, 0.0845], + device='cuda:4'), in_proj_covar=tensor([0.0167, 0.0198, 0.0199, 0.0186, 0.0214, 0.0205, 0.0221, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:00:09,535 INFO [finetune.py:976] (4/7) Epoch 9, batch 4000, loss[loss=0.1898, simple_loss=0.2608, pruned_loss=0.05942, over 4933.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2599, pruned_loss=0.06674, over 957136.76 frames. ], batch size: 33, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 11:00:10,252 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.56 vs. limit=5.0 +2023-03-26 11:00:13,721 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49827.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:00:16,514 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.178e+02 1.645e+02 2.017e+02 2.376e+02 6.319e+02, threshold=4.034e+02, percent-clipped=3.0 +2023-03-26 11:00:42,824 INFO [finetune.py:976] (4/7) Epoch 9, batch 4050, loss[loss=0.2284, simple_loss=0.2943, pruned_loss=0.08122, over 4845.00 frames. ], tot_loss[loss=0.2008, simple_loss=0.2645, pruned_loss=0.06851, over 956862.38 frames. ], batch size: 47, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 11:00:56,964 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49892.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:01:02,847 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.93 vs. limit=5.0 +2023-03-26 11:01:15,999 INFO [finetune.py:976] (4/7) Epoch 9, batch 4100, loss[loss=0.2215, simple_loss=0.2707, pruned_loss=0.08619, over 4817.00 frames. ], tot_loss[loss=0.2034, simple_loss=0.2673, pruned_loss=0.06975, over 955655.64 frames. ], batch size: 33, lr: 3.78e-03, grad_scale: 32.0 +2023-03-26 11:01:17,503 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-03-26 11:01:22,950 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.254e+02 1.718e+02 2.083e+02 2.512e+02 3.689e+02, threshold=4.166e+02, percent-clipped=0.0 +2023-03-26 11:01:27,154 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4972, 1.1240, 0.8599, 1.3755, 1.8907, 0.7768, 1.3277, 1.3538], + device='cuda:4'), covar=tensor([0.1572, 0.2224, 0.1832, 0.1277, 0.2118, 0.2184, 0.1528, 0.2199], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0092, 0.0122, 0.0095, 0.0099, 0.0092], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 11:01:28,937 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=49940.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:01:48,769 INFO [finetune.py:976] (4/7) Epoch 9, batch 4150, loss[loss=0.2427, simple_loss=0.297, pruned_loss=0.09423, over 4819.00 frames. ], tot_loss[loss=0.2044, simple_loss=0.2689, pruned_loss=0.0699, over 955524.24 frames. ], batch size: 39, lr: 3.78e-03, grad_scale: 16.0 +2023-03-26 11:01:57,851 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5952, 1.5535, 1.3748, 1.8121, 1.9249, 1.6985, 1.2320, 1.3711], + device='cuda:4'), covar=tensor([0.2285, 0.2179, 0.1987, 0.1648, 0.1770, 0.1198, 0.2635, 0.1915], + device='cuda:4'), in_proj_covar=tensor([0.0238, 0.0210, 0.0208, 0.0189, 0.0241, 0.0180, 0.0214, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:02:01,361 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0335, 0.9304, 0.9315, 1.1101, 1.1661, 1.1237, 1.0117, 0.9638], + device='cuda:4'), covar=tensor([0.0317, 0.0305, 0.0620, 0.0272, 0.0274, 0.0467, 0.0298, 0.0408], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0109, 0.0139, 0.0114, 0.0102, 0.0101, 0.0091, 0.0108], + device='cuda:4'), out_proj_covar=tensor([6.9701e-05, 8.4979e-05, 1.1062e-04, 8.9307e-05, 7.9659e-05, 7.5134e-05, + 6.8827e-05, 8.3041e-05], device='cuda:4') +2023-03-26 11:02:05,626 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5880, 2.2655, 1.9226, 0.8791, 2.0402, 1.8857, 1.5942, 1.9771], + device='cuda:4'), covar=tensor([0.0777, 0.0991, 0.1733, 0.2272, 0.1667, 0.2317, 0.2367, 0.1189], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0200, 0.0201, 0.0187, 0.0215, 0.0206, 0.0223, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:02:23,457 INFO [finetune.py:976] (4/7) Epoch 9, batch 4200, loss[loss=0.1905, simple_loss=0.2593, pruned_loss=0.06087, over 4756.00 frames. ], tot_loss[loss=0.2033, simple_loss=0.2684, pruned_loss=0.0691, over 955504.92 frames. ], batch size: 28, lr: 3.78e-03, grad_scale: 16.0 +2023-03-26 11:02:31,376 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.147e+02 1.706e+02 2.002e+02 2.506e+02 6.230e+02, threshold=4.003e+02, percent-clipped=2.0 +2023-03-26 11:02:47,171 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0424, 0.9322, 0.9168, 1.1676, 1.1634, 1.1352, 1.0093, 1.0001], + device='cuda:4'), covar=tensor([0.0328, 0.0281, 0.0641, 0.0257, 0.0275, 0.0421, 0.0305, 0.0387], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0109, 0.0139, 0.0114, 0.0102, 0.0102, 0.0091, 0.0108], + device='cuda:4'), out_proj_covar=tensor([6.9910e-05, 8.5153e-05, 1.1100e-04, 8.9453e-05, 7.9965e-05, 7.5505e-05, + 6.9028e-05, 8.3167e-05], device='cuda:4') +2023-03-26 11:03:15,796 INFO [finetune.py:976] (4/7) Epoch 9, batch 4250, loss[loss=0.239, simple_loss=0.2935, pruned_loss=0.0923, over 4884.00 frames. ], tot_loss[loss=0.201, simple_loss=0.2659, pruned_loss=0.06805, over 957269.92 frames. ], batch size: 32, lr: 3.78e-03, grad_scale: 16.0 +2023-03-26 11:03:24,164 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50078.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:03:32,347 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50088.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 11:03:53,929 INFO [finetune.py:976] (4/7) Epoch 9, batch 4300, loss[loss=0.1525, simple_loss=0.2272, pruned_loss=0.0389, over 4772.00 frames. ], tot_loss[loss=0.198, simple_loss=0.2629, pruned_loss=0.06655, over 958214.89 frames. ], batch size: 28, lr: 3.78e-03, grad_scale: 16.0 +2023-03-26 11:03:54,015 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50122.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:03:56,368 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=50126.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:03:57,615 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50128.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:04:00,431 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.608e+02 1.917e+02 2.228e+02 4.011e+02, threshold=3.835e+02, percent-clipped=1.0 +2023-03-26 11:04:03,883 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=50136.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:04:49,374 INFO [finetune.py:976] (4/7) Epoch 9, batch 4350, loss[loss=0.2433, simple_loss=0.2987, pruned_loss=0.09394, over 4855.00 frames. ], tot_loss[loss=0.195, simple_loss=0.2595, pruned_loss=0.06525, over 957876.33 frames. ], batch size: 44, lr: 3.78e-03, grad_scale: 16.0 +2023-03-26 11:04:59,030 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-26 11:04:59,865 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8795, 2.5449, 2.1019, 1.0705, 2.2197, 2.2415, 1.9083, 2.1920], + device='cuda:4'), covar=tensor([0.0824, 0.0801, 0.1554, 0.2190, 0.1533, 0.1960, 0.2050, 0.1017], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0200, 0.0200, 0.0187, 0.0215, 0.0207, 0.0223, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:05:18,364 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50189.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:05:30,449 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50197.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 11:05:32,712 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.76 vs. limit=5.0 +2023-03-26 11:05:54,102 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1701, 1.1605, 1.0974, 1.1747, 1.3849, 1.3573, 1.2132, 1.0872], + device='cuda:4'), covar=tensor([0.0332, 0.0245, 0.0542, 0.0275, 0.0217, 0.0337, 0.0311, 0.0358], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0109, 0.0138, 0.0114, 0.0102, 0.0101, 0.0091, 0.0108], + device='cuda:4'), out_proj_covar=tensor([6.9725e-05, 8.5006e-05, 1.1043e-04, 8.9419e-05, 7.9774e-05, 7.5306e-05, + 6.8731e-05, 8.3081e-05], device='cuda:4') +2023-03-26 11:06:02,457 INFO [finetune.py:976] (4/7) Epoch 9, batch 4400, loss[loss=0.2168, simple_loss=0.2872, pruned_loss=0.07318, over 4863.00 frames. ], tot_loss[loss=0.1963, simple_loss=0.2607, pruned_loss=0.06593, over 954708.47 frames. ], batch size: 34, lr: 3.78e-03, grad_scale: 16.0 +2023-03-26 11:06:14,080 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.827e+01 1.714e+02 1.989e+02 2.480e+02 5.028e+02, threshold=3.977e+02, percent-clipped=2.0 +2023-03-26 11:06:24,319 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7596, 1.4668, 2.3325, 3.5047, 2.3222, 2.5358, 0.8134, 2.7715], + device='cuda:4'), covar=tensor([0.1706, 0.1662, 0.1214, 0.0570, 0.0837, 0.1356, 0.2047, 0.0547], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0117, 0.0133, 0.0165, 0.0102, 0.0138, 0.0126, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 11:06:45,051 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4897, 2.2503, 1.8307, 0.8015, 1.9170, 1.8137, 1.6508, 1.9956], + device='cuda:4'), covar=tensor([0.0920, 0.0852, 0.1851, 0.2169, 0.1805, 0.2494, 0.2419, 0.1095], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0200, 0.0200, 0.0187, 0.0216, 0.0207, 0.0223, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:06:48,121 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50258.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 11:06:58,871 INFO [finetune.py:976] (4/7) Epoch 9, batch 4450, loss[loss=0.2335, simple_loss=0.3056, pruned_loss=0.08064, over 4768.00 frames. ], tot_loss[loss=0.2011, simple_loss=0.2657, pruned_loss=0.06819, over 954923.48 frames. ], batch size: 59, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:07:04,461 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7364, 1.6416, 2.1491, 2.0090, 1.8280, 3.5443, 1.5119, 1.9060], + device='cuda:4'), covar=tensor([0.0895, 0.1529, 0.1476, 0.0884, 0.1386, 0.0291, 0.1325, 0.1458], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0080, 0.0075, 0.0077, 0.0091, 0.0081, 0.0084, 0.0078], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 11:07:32,436 INFO [finetune.py:976] (4/7) Epoch 9, batch 4500, loss[loss=0.2137, simple_loss=0.2806, pruned_loss=0.07334, over 4856.00 frames. ], tot_loss[loss=0.2016, simple_loss=0.2667, pruned_loss=0.0683, over 955766.31 frames. ], batch size: 44, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:07:38,449 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.049e+02 1.523e+02 1.896e+02 2.480e+02 4.445e+02, threshold=3.793e+02, percent-clipped=2.0 +2023-03-26 11:07:44,030 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6926, 2.5586, 2.1447, 2.8427, 2.6443, 2.2414, 3.1751, 2.6266], + device='cuda:4'), covar=tensor([0.1355, 0.2520, 0.3283, 0.2844, 0.2811, 0.1801, 0.2999, 0.2007], + device='cuda:4'), in_proj_covar=tensor([0.0174, 0.0190, 0.0235, 0.0255, 0.0240, 0.0196, 0.0213, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:08:05,987 INFO [finetune.py:976] (4/7) Epoch 9, batch 4550, loss[loss=0.2319, simple_loss=0.288, pruned_loss=0.08793, over 4811.00 frames. ], tot_loss[loss=0.202, simple_loss=0.2669, pruned_loss=0.06855, over 952278.40 frames. ], batch size: 39, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:08:36,641 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8197, 1.6430, 1.5932, 1.8704, 2.0331, 1.9395, 1.3611, 1.6044], + device='cuda:4'), covar=tensor([0.2015, 0.1987, 0.1742, 0.1524, 0.1600, 0.0967, 0.2432, 0.1833], + device='cuda:4'), in_proj_covar=tensor([0.0240, 0.0210, 0.0208, 0.0190, 0.0243, 0.0181, 0.0216, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:08:58,492 INFO [finetune.py:976] (4/7) Epoch 9, batch 4600, loss[loss=0.1782, simple_loss=0.2437, pruned_loss=0.05638, over 4922.00 frames. ], tot_loss[loss=0.2012, simple_loss=0.2661, pruned_loss=0.06816, over 951433.35 frames. ], batch size: 33, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:08:58,585 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50422.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:09:06,261 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.671e+02 1.983e+02 2.416e+02 3.848e+02, threshold=3.965e+02, percent-clipped=1.0 +2023-03-26 11:09:39,358 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=50470.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:09:40,499 INFO [finetune.py:976] (4/7) Epoch 9, batch 4650, loss[loss=0.1581, simple_loss=0.2353, pruned_loss=0.04043, over 4779.00 frames. ], tot_loss[loss=0.1997, simple_loss=0.2642, pruned_loss=0.06755, over 954099.41 frames. ], batch size: 28, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:09:50,494 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50484.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:10:18,040 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4664, 1.6432, 1.3573, 1.5645, 1.8637, 1.7930, 1.6212, 1.5045], + device='cuda:4'), covar=tensor([0.0479, 0.0298, 0.0627, 0.0323, 0.0222, 0.0527, 0.0308, 0.0366], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0110, 0.0140, 0.0115, 0.0103, 0.0102, 0.0092, 0.0109], + device='cuda:4'), out_proj_covar=tensor([7.0414e-05, 8.6001e-05, 1.1145e-04, 9.0340e-05, 8.0456e-05, 7.6045e-05, + 6.9474e-05, 8.4101e-05], device='cuda:4') +2023-03-26 11:10:22,784 INFO [finetune.py:976] (4/7) Epoch 9, batch 4700, loss[loss=0.2047, simple_loss=0.2591, pruned_loss=0.0751, over 4899.00 frames. ], tot_loss[loss=0.1972, simple_loss=0.2611, pruned_loss=0.06666, over 955146.80 frames. ], batch size: 35, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:10:24,818 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 11:10:29,353 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.527e+02 1.850e+02 2.224e+02 3.838e+02, threshold=3.699e+02, percent-clipped=0.0 +2023-03-26 11:10:38,037 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50546.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:10:42,745 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50553.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 11:10:42,779 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5779, 1.5686, 2.0181, 1.2743, 1.6102, 1.8385, 1.5162, 2.0077], + device='cuda:4'), covar=tensor([0.1341, 0.2235, 0.1258, 0.1840, 0.1009, 0.1377, 0.2897, 0.0918], + device='cuda:4'), in_proj_covar=tensor([0.0200, 0.0206, 0.0195, 0.0193, 0.0180, 0.0220, 0.0219, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:10:56,119 INFO [finetune.py:976] (4/7) Epoch 9, batch 4750, loss[loss=0.1963, simple_loss=0.2665, pruned_loss=0.06308, over 4790.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2591, pruned_loss=0.06612, over 955860.07 frames. ], batch size: 45, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:10:59,274 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6335, 1.4760, 2.2519, 3.2807, 2.2233, 2.3124, 1.1629, 2.4803], + device='cuda:4'), covar=tensor([0.1859, 0.1590, 0.1304, 0.0626, 0.0851, 0.1513, 0.1859, 0.0680], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0118, 0.0134, 0.0165, 0.0102, 0.0139, 0.0126, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 11:11:18,567 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50607.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:11:29,494 INFO [finetune.py:976] (4/7) Epoch 9, batch 4800, loss[loss=0.2098, simple_loss=0.282, pruned_loss=0.06879, over 4838.00 frames. ], tot_loss[loss=0.1992, simple_loss=0.2627, pruned_loss=0.06788, over 955969.54 frames. ], batch size: 49, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:11:36,105 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.695e+01 1.587e+02 1.907e+02 2.189e+02 3.978e+02, threshold=3.813e+02, percent-clipped=2.0 +2023-03-26 11:11:50,423 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-26 11:12:03,073 INFO [finetune.py:976] (4/7) Epoch 9, batch 4850, loss[loss=0.202, simple_loss=0.2601, pruned_loss=0.07192, over 4733.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.2647, pruned_loss=0.06758, over 955707.48 frames. ], batch size: 23, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:12:31,047 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.53 vs. limit=5.0 +2023-03-26 11:12:36,222 INFO [finetune.py:976] (4/7) Epoch 9, batch 4900, loss[loss=0.2097, simple_loss=0.2771, pruned_loss=0.0711, over 4769.00 frames. ], tot_loss[loss=0.2013, simple_loss=0.266, pruned_loss=0.06828, over 955773.33 frames. ], batch size: 54, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:12:42,295 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.610e+02 1.915e+02 2.289e+02 4.400e+02, threshold=3.830e+02, percent-clipped=2.0 +2023-03-26 11:12:43,026 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.9879, 1.8841, 2.0344, 0.8030, 2.1293, 2.3283, 1.9649, 1.8006], + device='cuda:4'), covar=tensor([0.0992, 0.0714, 0.0393, 0.0737, 0.0473, 0.0680, 0.0485, 0.0744], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0154, 0.0121, 0.0134, 0.0131, 0.0125, 0.0145, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.5863e-05, 1.1356e-04, 8.7025e-05, 9.7160e-05, 9.3616e-05, 9.1712e-05, + 1.0648e-04, 1.0777e-04], device='cuda:4') +2023-03-26 11:13:08,691 INFO [finetune.py:976] (4/7) Epoch 9, batch 4950, loss[loss=0.2435, simple_loss=0.3037, pruned_loss=0.0917, over 4802.00 frames. ], tot_loss[loss=0.2043, simple_loss=0.2689, pruned_loss=0.06982, over 955852.32 frames. ], batch size: 40, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:13:15,369 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50782.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:13:16,533 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50784.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:13:25,450 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2978, 1.9429, 1.5947, 0.7200, 1.8176, 1.8427, 1.5349, 1.7365], + device='cuda:4'), covar=tensor([0.0871, 0.0875, 0.1382, 0.1884, 0.1310, 0.1760, 0.2069, 0.0942], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0199, 0.0199, 0.0187, 0.0215, 0.0206, 0.0223, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:13:26,103 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.90 vs. limit=5.0 +2023-03-26 11:13:33,488 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50808.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:13:44,497 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8329, 1.6644, 2.0483, 1.3982, 1.8715, 2.0654, 1.6292, 2.2266], + device='cuda:4'), covar=tensor([0.1178, 0.1969, 0.1407, 0.1828, 0.0864, 0.1164, 0.2655, 0.0755], + device='cuda:4'), in_proj_covar=tensor([0.0200, 0.0206, 0.0196, 0.0193, 0.0180, 0.0219, 0.0219, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:13:51,975 INFO [finetune.py:976] (4/7) Epoch 9, batch 5000, loss[loss=0.2072, simple_loss=0.2621, pruned_loss=0.07609, over 4862.00 frames. ], tot_loss[loss=0.2017, simple_loss=0.2662, pruned_loss=0.06861, over 956216.09 frames. ], batch size: 34, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:14:03,067 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.212e+02 1.752e+02 2.044e+02 2.444e+02 6.074e+02, threshold=4.089e+02, percent-clipped=4.0 +2023-03-26 11:14:03,135 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=50832.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:14:13,399 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50843.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:14:23,314 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-26 11:14:24,235 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50853.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:14:37,409 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50869.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:14:40,057 INFO [finetune.py:976] (4/7) Epoch 9, batch 5050, loss[loss=0.2083, simple_loss=0.2686, pruned_loss=0.07402, over 4826.00 frames. ], tot_loss[loss=0.1995, simple_loss=0.2631, pruned_loss=0.0679, over 955242.73 frames. ], batch size: 39, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:14:49,483 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2165, 1.3143, 0.6381, 2.1702, 2.5032, 1.7098, 1.8166, 1.9083], + device='cuda:4'), covar=tensor([0.1432, 0.2145, 0.2431, 0.1179, 0.1837, 0.2124, 0.1480, 0.2137], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0096, 0.0113, 0.0092, 0.0121, 0.0096, 0.0100, 0.0092], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 11:15:00,246 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=50901.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 11:15:00,837 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50902.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:15:13,734 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8893, 1.9257, 2.3382, 2.0590, 2.1925, 4.6125, 2.0037, 2.2654], + device='cuda:4'), covar=tensor([0.1030, 0.1647, 0.0986, 0.1014, 0.1362, 0.0161, 0.1307, 0.1557], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0091, 0.0082, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 11:15:21,256 INFO [finetune.py:976] (4/7) Epoch 9, batch 5100, loss[loss=0.2206, simple_loss=0.2723, pruned_loss=0.08442, over 4759.00 frames. ], tot_loss[loss=0.1941, simple_loss=0.2579, pruned_loss=0.06512, over 956039.93 frames. ], batch size: 27, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:15:29,782 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.163e+02 1.571e+02 1.989e+02 2.366e+02 5.072e+02, threshold=3.977e+02, percent-clipped=2.0 +2023-03-26 11:15:38,846 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50946.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:15:55,123 INFO [finetune.py:976] (4/7) Epoch 9, batch 5150, loss[loss=0.2324, simple_loss=0.3026, pruned_loss=0.08115, over 4843.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.259, pruned_loss=0.06615, over 954995.04 frames. ], batch size: 49, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:16:10,906 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-03-26 11:16:11,763 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-03-26 11:16:19,663 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51007.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:16:29,111 INFO [finetune.py:976] (4/7) Epoch 9, batch 5200, loss[loss=0.2342, simple_loss=0.2953, pruned_loss=0.0865, over 4898.00 frames. ], tot_loss[loss=0.1992, simple_loss=0.2634, pruned_loss=0.06753, over 954945.92 frames. ], batch size: 43, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:16:37,438 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.247e+02 1.691e+02 2.095e+02 2.506e+02 4.401e+02, threshold=4.191e+02, percent-clipped=2.0 +2023-03-26 11:17:12,601 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51062.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:17:18,570 INFO [finetune.py:976] (4/7) Epoch 9, batch 5250, loss[loss=0.1855, simple_loss=0.2591, pruned_loss=0.05592, over 4849.00 frames. ], tot_loss[loss=0.2014, simple_loss=0.266, pruned_loss=0.06842, over 954974.92 frames. ], batch size: 44, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:17:41,174 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6768, 1.7078, 1.7281, 0.9962, 1.8580, 1.9368, 1.8384, 1.4752], + device='cuda:4'), covar=tensor([0.0967, 0.0562, 0.0520, 0.0639, 0.0391, 0.0579, 0.0373, 0.0773], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0156, 0.0122, 0.0135, 0.0132, 0.0127, 0.0147, 0.0148], + device='cuda:4'), out_proj_covar=tensor([9.6881e-05, 1.1456e-04, 8.7982e-05, 9.7999e-05, 9.4450e-05, 9.2536e-05, + 1.0766e-04, 1.0866e-04], device='cuda:4') +2023-03-26 11:17:51,198 INFO [finetune.py:976] (4/7) Epoch 9, batch 5300, loss[loss=0.1522, simple_loss=0.2266, pruned_loss=0.03892, over 4857.00 frames. ], tot_loss[loss=0.2018, simple_loss=0.2667, pruned_loss=0.06841, over 954419.40 frames. ], batch size: 31, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:17:51,959 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51123.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:17:53,863 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.81 vs. limit=5.0 +2023-03-26 11:17:57,262 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.167e+02 1.689e+02 2.023e+02 2.414e+02 5.734e+02, threshold=4.045e+02, percent-clipped=1.0 +2023-03-26 11:17:59,055 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0965, 1.8876, 1.7825, 2.1175, 2.6965, 2.0403, 1.9293, 1.5843], + device='cuda:4'), covar=tensor([0.2161, 0.2101, 0.1898, 0.1642, 0.1740, 0.1248, 0.2338, 0.1950], + device='cuda:4'), in_proj_covar=tensor([0.0238, 0.0210, 0.0207, 0.0190, 0.0242, 0.0181, 0.0214, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:18:01,855 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51138.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:18:17,296 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6979, 1.2390, 0.9101, 1.6375, 2.0680, 1.2450, 1.5075, 1.6090], + device='cuda:4'), covar=tensor([0.1445, 0.1979, 0.1984, 0.1136, 0.1857, 0.2175, 0.1372, 0.1897], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0092, 0.0121, 0.0095, 0.0100, 0.0092], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 11:18:19,582 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51164.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 11:18:24,358 INFO [finetune.py:976] (4/7) Epoch 9, batch 5350, loss[loss=0.1987, simple_loss=0.2675, pruned_loss=0.06499, over 4889.00 frames. ], tot_loss[loss=0.202, simple_loss=0.267, pruned_loss=0.06854, over 953470.08 frames. ], batch size: 35, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:18:24,454 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51172.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 11:18:35,431 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-26 11:18:56,189 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51202.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:19:18,523 INFO [finetune.py:976] (4/7) Epoch 9, batch 5400, loss[loss=0.1726, simple_loss=0.2399, pruned_loss=0.05261, over 4930.00 frames. ], tot_loss[loss=0.2, simple_loss=0.2644, pruned_loss=0.06779, over 955946.58 frames. ], batch size: 38, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:19:26,735 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.913e+01 1.608e+02 1.826e+02 2.251e+02 3.272e+02, threshold=3.651e+02, percent-clipped=0.0 +2023-03-26 11:19:32,720 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51233.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:19:48,897 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51250.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:19:57,398 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1342, 1.9223, 1.7466, 1.7345, 2.1049, 1.8057, 2.2257, 2.0548], + device='cuda:4'), covar=tensor([0.1350, 0.2404, 0.3179, 0.2817, 0.2721, 0.1767, 0.2808, 0.2060], + device='cuda:4'), in_proj_covar=tensor([0.0173, 0.0188, 0.0232, 0.0253, 0.0238, 0.0195, 0.0211, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:20:03,230 INFO [finetune.py:976] (4/7) Epoch 9, batch 5450, loss[loss=0.1715, simple_loss=0.2396, pruned_loss=0.05171, over 4819.00 frames. ], tot_loss[loss=0.1963, simple_loss=0.2603, pruned_loss=0.06613, over 956120.76 frames. ], batch size: 38, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:20:31,732 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51302.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:20:50,274 INFO [finetune.py:976] (4/7) Epoch 9, batch 5500, loss[loss=0.1421, simple_loss=0.2169, pruned_loss=0.03365, over 4784.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.257, pruned_loss=0.06485, over 956074.91 frames. ], batch size: 29, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:20:56,824 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.040e+02 1.581e+02 1.869e+02 2.249e+02 3.902e+02, threshold=3.738e+02, percent-clipped=2.0 +2023-03-26 11:21:48,954 INFO [finetune.py:976] (4/7) Epoch 9, batch 5550, loss[loss=0.2288, simple_loss=0.2919, pruned_loss=0.08288, over 4921.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2605, pruned_loss=0.06641, over 954114.96 frames. ], batch size: 38, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:21:59,551 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51382.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:22:07,977 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51392.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:22:24,913 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51418.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:22:27,582 INFO [finetune.py:976] (4/7) Epoch 9, batch 5600, loss[loss=0.1815, simple_loss=0.2267, pruned_loss=0.06817, over 4223.00 frames. ], tot_loss[loss=0.2007, simple_loss=0.2654, pruned_loss=0.06797, over 954388.13 frames. ], batch size: 18, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:22:33,286 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.626e+02 2.005e+02 2.362e+02 4.096e+02, threshold=4.011e+02, percent-clipped=2.0 +2023-03-26 11:22:36,855 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51438.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:22:39,792 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51443.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:22:46,038 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51453.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:22:52,436 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51464.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:22:57,074 INFO [finetune.py:976] (4/7) Epoch 9, batch 5650, loss[loss=0.1771, simple_loss=0.2418, pruned_loss=0.05621, over 4777.00 frames. ], tot_loss[loss=0.2027, simple_loss=0.2677, pruned_loss=0.06886, over 951393.74 frames. ], batch size: 26, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:22:57,787 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0768, 1.6559, 2.0240, 1.9943, 1.7254, 1.7150, 1.9249, 1.7964], + device='cuda:4'), covar=tensor([0.4763, 0.5584, 0.4272, 0.5134, 0.6350, 0.4659, 0.6292, 0.4300], + device='cuda:4'), in_proj_covar=tensor([0.0236, 0.0242, 0.0255, 0.0258, 0.0252, 0.0227, 0.0276, 0.0231], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:23:05,307 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51486.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:23:10,714 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51495.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:23:20,780 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51512.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 11:23:25,703 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-26 11:23:26,713 INFO [finetune.py:976] (4/7) Epoch 9, batch 5700, loss[loss=0.1485, simple_loss=0.207, pruned_loss=0.04497, over 4319.00 frames. ], tot_loss[loss=0.199, simple_loss=0.2625, pruned_loss=0.06777, over 930331.13 frames. ], batch size: 18, lr: 3.77e-03, grad_scale: 16.0 +2023-03-26 11:23:30,374 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51528.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 11:23:32,893 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.987e+01 1.624e+02 1.963e+02 2.341e+02 6.572e+02, threshold=3.927e+02, percent-clipped=1.0 +2023-03-26 11:23:57,292 INFO [finetune.py:976] (4/7) Epoch 10, batch 0, loss[loss=0.1738, simple_loss=0.2378, pruned_loss=0.05495, over 4832.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2378, pruned_loss=0.05495, over 4832.00 frames. ], batch size: 49, lr: 3.76e-03, grad_scale: 16.0 +2023-03-26 11:23:57,293 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 11:24:16,171 INFO [finetune.py:1010] (4/7) Epoch 10, validation: loss=0.1604, simple_loss=0.2317, pruned_loss=0.04451, over 2265189.00 frames. +2023-03-26 11:24:16,171 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 11:24:22,493 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51556.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:24:31,153 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-26 11:24:52,583 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5264, 1.0761, 0.7157, 1.3975, 1.8845, 0.7018, 1.3180, 1.3940], + device='cuda:4'), covar=tensor([0.1590, 0.2243, 0.1970, 0.1297, 0.2118, 0.2127, 0.1530, 0.2171], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0092, 0.0121, 0.0095, 0.0100, 0.0091], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 11:24:58,291 INFO [finetune.py:976] (4/7) Epoch 10, batch 50, loss[loss=0.1769, simple_loss=0.2501, pruned_loss=0.05188, over 4806.00 frames. ], tot_loss[loss=0.202, simple_loss=0.2667, pruned_loss=0.06862, over 215135.32 frames. ], batch size: 39, lr: 3.76e-03, grad_scale: 16.0 +2023-03-26 11:25:01,630 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51602.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:25:20,208 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.916e+01 1.735e+02 2.131e+02 2.642e+02 7.480e+02, threshold=4.262e+02, percent-clipped=4.0 +2023-03-26 11:25:31,992 INFO [finetune.py:976] (4/7) Epoch 10, batch 100, loss[loss=0.1915, simple_loss=0.2575, pruned_loss=0.06278, over 4827.00 frames. ], tot_loss[loss=0.1971, simple_loss=0.2601, pruned_loss=0.06704, over 379119.60 frames. ], batch size: 38, lr: 3.76e-03, grad_scale: 16.0 +2023-03-26 11:25:33,108 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51650.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:26:04,771 INFO [finetune.py:976] (4/7) Epoch 10, batch 150, loss[loss=0.1974, simple_loss=0.2568, pruned_loss=0.06893, over 4799.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2571, pruned_loss=0.06568, over 509229.75 frames. ], batch size: 45, lr: 3.76e-03, grad_scale: 16.0 +2023-03-26 11:26:18,704 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51718.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:26:33,392 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.594e+02 1.858e+02 2.240e+02 3.308e+02, threshold=3.716e+02, percent-clipped=0.0 +2023-03-26 11:26:37,113 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51738.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:26:47,598 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51747.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:26:48,176 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51748.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:26:52,006 INFO [finetune.py:976] (4/7) Epoch 10, batch 200, loss[loss=0.2536, simple_loss=0.2985, pruned_loss=0.1043, over 3903.00 frames. ], tot_loss[loss=0.193, simple_loss=0.255, pruned_loss=0.0655, over 606865.73 frames. ], batch size: 65, lr: 3.76e-03, grad_scale: 16.0 +2023-03-26 11:27:04,853 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51766.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:27:25,425 INFO [finetune.py:976] (4/7) Epoch 10, batch 250, loss[loss=0.182, simple_loss=0.259, pruned_loss=0.05249, over 4755.00 frames. ], tot_loss[loss=0.1955, simple_loss=0.2584, pruned_loss=0.06627, over 684945.60 frames. ], batch size: 27, lr: 3.76e-03, grad_scale: 16.0 +2023-03-26 11:27:33,011 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51808.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:27:45,687 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51828.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 11:27:48,003 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.605e+02 1.930e+02 2.331e+02 5.576e+02, threshold=3.861e+02, percent-clipped=5.0 +2023-03-26 11:27:58,885 INFO [finetune.py:976] (4/7) Epoch 10, batch 300, loss[loss=0.2064, simple_loss=0.2837, pruned_loss=0.06456, over 4807.00 frames. ], tot_loss[loss=0.1969, simple_loss=0.2611, pruned_loss=0.06631, over 745719.50 frames. ], batch size: 51, lr: 3.76e-03, grad_scale: 16.0 +2023-03-26 11:28:00,161 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51851.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:28:04,960 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1031, 1.6633, 1.9883, 1.9561, 1.7429, 1.7289, 1.9112, 1.8722], + device='cuda:4'), covar=tensor([0.4894, 0.5636, 0.4507, 0.5461, 0.6248, 0.4802, 0.7040, 0.4243], + device='cuda:4'), in_proj_covar=tensor([0.0235, 0.0241, 0.0255, 0.0257, 0.0252, 0.0227, 0.0276, 0.0230], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:28:17,690 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=51876.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:28:31,953 INFO [finetune.py:976] (4/7) Epoch 10, batch 350, loss[loss=0.2497, simple_loss=0.2943, pruned_loss=0.1025, over 4806.00 frames. ], tot_loss[loss=0.2, simple_loss=0.2649, pruned_loss=0.06752, over 792908.89 frames. ], batch size: 25, lr: 3.76e-03, grad_scale: 16.0 +2023-03-26 11:28:46,752 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.42 vs. limit=5.0 +2023-03-26 11:28:54,280 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.052e+02 1.678e+02 2.044e+02 2.443e+02 3.814e+02, threshold=4.089e+02, percent-clipped=0.0 +2023-03-26 11:29:04,638 INFO [finetune.py:976] (4/7) Epoch 10, batch 400, loss[loss=0.1886, simple_loss=0.258, pruned_loss=0.05959, over 4837.00 frames. ], tot_loss[loss=0.2005, simple_loss=0.2655, pruned_loss=0.0677, over 827516.84 frames. ], batch size: 49, lr: 3.76e-03, grad_scale: 16.0 +2023-03-26 11:29:30,096 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5586, 1.4713, 1.7359, 1.7989, 1.6136, 3.4768, 1.3657, 1.6729], + device='cuda:4'), covar=tensor([0.0915, 0.1722, 0.1150, 0.0928, 0.1490, 0.0226, 0.1449, 0.1637], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0091, 0.0082, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 11:29:56,983 INFO [finetune.py:976] (4/7) Epoch 10, batch 450, loss[loss=0.2129, simple_loss=0.2739, pruned_loss=0.07594, over 4895.00 frames. ], tot_loss[loss=0.199, simple_loss=0.2638, pruned_loss=0.06708, over 855222.44 frames. ], batch size: 35, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:30:21,154 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.724e+02 2.025e+02 2.574e+02 4.346e+02, threshold=4.050e+02, percent-clipped=1.0 +2023-03-26 11:30:24,978 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52038.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:30:26,197 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52040.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:30:30,947 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52048.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:30:31,464 INFO [finetune.py:976] (4/7) Epoch 10, batch 500, loss[loss=0.192, simple_loss=0.2569, pruned_loss=0.06356, over 4935.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.2613, pruned_loss=0.06616, over 879361.97 frames. ], batch size: 33, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:30:41,529 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.87 vs. limit=5.0 +2023-03-26 11:30:56,287 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8954, 1.7151, 1.5806, 1.2948, 1.9152, 1.6493, 1.8528, 1.8286], + device='cuda:4'), covar=tensor([0.1562, 0.2362, 0.3518, 0.2869, 0.2968, 0.1932, 0.2999, 0.2100], + device='cuda:4'), in_proj_covar=tensor([0.0174, 0.0188, 0.0233, 0.0253, 0.0239, 0.0196, 0.0212, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:30:56,807 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52086.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:31:02,788 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52096.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:31:04,606 INFO [finetune.py:976] (4/7) Epoch 10, batch 550, loss[loss=0.1584, simple_loss=0.2138, pruned_loss=0.05147, over 4439.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2584, pruned_loss=0.06517, over 897379.22 frames. ], batch size: 19, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:31:05,924 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52101.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:31:07,068 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52103.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:31:16,088 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3993, 1.2624, 1.2418, 1.2560, 1.5599, 1.5170, 1.3971, 1.2171], + device='cuda:4'), covar=tensor([0.0297, 0.0278, 0.0536, 0.0310, 0.0239, 0.0475, 0.0306, 0.0364], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0108, 0.0138, 0.0114, 0.0101, 0.0100, 0.0090, 0.0107], + device='cuda:4'), out_proj_covar=tensor([6.9568e-05, 8.4209e-05, 1.0952e-04, 8.9203e-05, 7.9529e-05, 7.4213e-05, + 6.8184e-05, 8.2422e-05], device='cuda:4') +2023-03-26 11:31:27,058 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.020e+02 1.591e+02 1.822e+02 2.163e+02 6.487e+02, threshold=3.643e+02, percent-clipped=1.0 +2023-03-26 11:31:32,633 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6060, 1.4284, 1.5834, 0.9467, 1.7054, 1.8708, 1.7659, 1.3614], + device='cuda:4'), covar=tensor([0.1113, 0.0990, 0.0509, 0.0657, 0.0524, 0.0495, 0.0470, 0.0852], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0154, 0.0121, 0.0134, 0.0131, 0.0124, 0.0144, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.5626e-05, 1.1353e-04, 8.7010e-05, 9.7241e-05, 9.3711e-05, 9.0884e-05, + 1.0594e-04, 1.0754e-04], device='cuda:4') +2023-03-26 11:31:37,963 INFO [finetune.py:976] (4/7) Epoch 10, batch 600, loss[loss=0.2188, simple_loss=0.2733, pruned_loss=0.08215, over 4825.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2604, pruned_loss=0.06646, over 910956.53 frames. ], batch size: 33, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:31:39,240 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52151.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:32:15,973 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52193.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:32:19,560 INFO [finetune.py:976] (4/7) Epoch 10, batch 650, loss[loss=0.2744, simple_loss=0.3221, pruned_loss=0.1134, over 4103.00 frames. ], tot_loss[loss=0.1987, simple_loss=0.2628, pruned_loss=0.06733, over 919760.47 frames. ], batch size: 65, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:32:19,619 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52199.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:32:42,615 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.128e+02 1.681e+02 1.969e+02 2.336e+02 3.855e+02, threshold=3.938e+02, percent-clipped=2.0 +2023-03-26 11:32:46,417 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3048, 1.3806, 1.4520, 1.4510, 1.5495, 3.0175, 1.2901, 1.5654], + device='cuda:4'), covar=tensor([0.1058, 0.1842, 0.1120, 0.1058, 0.1585, 0.0283, 0.1481, 0.1689], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0082, 0.0076, 0.0078, 0.0091, 0.0082, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 11:32:53,493 INFO [finetune.py:976] (4/7) Epoch 10, batch 700, loss[loss=0.2407, simple_loss=0.3042, pruned_loss=0.08866, over 4855.00 frames. ], tot_loss[loss=0.2001, simple_loss=0.2642, pruned_loss=0.06796, over 927524.20 frames. ], batch size: 31, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:32:55,082 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 11:32:56,642 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52254.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:33:26,703 INFO [finetune.py:976] (4/7) Epoch 10, batch 750, loss[loss=0.1841, simple_loss=0.2583, pruned_loss=0.05499, over 4809.00 frames. ], tot_loss[loss=0.2008, simple_loss=0.2652, pruned_loss=0.06825, over 930980.84 frames. ], batch size: 39, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:33:43,148 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3660, 2.2560, 1.9925, 1.0343, 2.0865, 1.8285, 1.6762, 2.0713], + device='cuda:4'), covar=tensor([0.0905, 0.0804, 0.1500, 0.2052, 0.1519, 0.2053, 0.2240, 0.1036], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0203, 0.0204, 0.0191, 0.0217, 0.0211, 0.0226, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:33:45,072 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52312.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:34:02,791 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.612e+02 1.864e+02 2.364e+02 4.342e+02, threshold=3.728e+02, percent-clipped=1.0 +2023-03-26 11:34:15,210 INFO [finetune.py:976] (4/7) Epoch 10, batch 800, loss[loss=0.1371, simple_loss=0.2171, pruned_loss=0.02861, over 4813.00 frames. ], tot_loss[loss=0.2014, simple_loss=0.266, pruned_loss=0.06838, over 935991.46 frames. ], batch size: 33, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:34:18,418 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6162, 1.4857, 1.5777, 1.5657, 0.8864, 3.0303, 1.1037, 1.5884], + device='cuda:4'), covar=tensor([0.3127, 0.2384, 0.2003, 0.2215, 0.1944, 0.0230, 0.2489, 0.1251], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0115, 0.0120, 0.0123, 0.0116, 0.0099, 0.0099, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 11:34:23,011 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.23 vs. limit=5.0 +2023-03-26 11:34:30,566 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52373.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:34:49,109 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52396.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:34:50,909 INFO [finetune.py:976] (4/7) Epoch 10, batch 850, loss[loss=0.1572, simple_loss=0.2244, pruned_loss=0.04498, over 4819.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.2626, pruned_loss=0.06692, over 940197.40 frames. ], batch size: 41, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:34:54,117 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52403.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:35:14,907 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.556e+02 1.848e+02 2.239e+02 3.627e+02, threshold=3.695e+02, percent-clipped=0.0 +2023-03-26 11:35:36,905 INFO [finetune.py:976] (4/7) Epoch 10, batch 900, loss[loss=0.1943, simple_loss=0.2572, pruned_loss=0.06576, over 4819.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2596, pruned_loss=0.06553, over 943048.12 frames. ], batch size: 39, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:35:38,209 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52451.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:35:50,395 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9285, 1.9061, 1.7075, 2.0686, 2.4776, 2.1671, 1.7570, 1.5891], + device='cuda:4'), covar=tensor([0.2156, 0.2098, 0.1994, 0.1751, 0.1821, 0.1120, 0.2533, 0.1977], + device='cuda:4'), in_proj_covar=tensor([0.0235, 0.0207, 0.0205, 0.0187, 0.0238, 0.0179, 0.0212, 0.0193], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:36:19,944 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.30 vs. limit=5.0 +2023-03-26 11:36:25,209 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0364, 1.7704, 2.2892, 1.5213, 2.1501, 2.1627, 1.7137, 2.3579], + device='cuda:4'), covar=tensor([0.1259, 0.1870, 0.1385, 0.2072, 0.0920, 0.1573, 0.2706, 0.0969], + device='cuda:4'), in_proj_covar=tensor([0.0200, 0.0206, 0.0195, 0.0193, 0.0180, 0.0217, 0.0218, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:36:25,702 INFO [finetune.py:976] (4/7) Epoch 10, batch 950, loss[loss=0.1704, simple_loss=0.2353, pruned_loss=0.05278, over 4764.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2579, pruned_loss=0.06477, over 946019.48 frames. ], batch size: 26, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:36:31,255 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 11:36:46,735 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.124e+02 1.551e+02 1.918e+02 2.238e+02 5.409e+02, threshold=3.837e+02, percent-clipped=4.0 +2023-03-26 11:37:01,195 INFO [finetune.py:976] (4/7) Epoch 10, batch 1000, loss[loss=0.1995, simple_loss=0.2616, pruned_loss=0.06868, over 4784.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.261, pruned_loss=0.06632, over 948442.02 frames. ], batch size: 28, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:37:01,275 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52549.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:37:04,952 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52555.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 11:37:07,686 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.57 vs. limit=2.0 +2023-03-26 11:37:14,974 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.49 vs. limit=5.0 +2023-03-26 11:37:26,954 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-03-26 11:37:57,588 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7457, 1.7376, 1.8417, 1.1949, 1.7811, 2.0024, 1.8818, 1.4342], + device='cuda:4'), covar=tensor([0.0536, 0.0637, 0.0636, 0.0853, 0.0773, 0.0563, 0.0521, 0.1174], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0131, 0.0142, 0.0122, 0.0117, 0.0141, 0.0141, 0.0159], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:38:00,468 INFO [finetune.py:976] (4/7) Epoch 10, batch 1050, loss[loss=0.1678, simple_loss=0.241, pruned_loss=0.04732, over 4832.00 frames. ], tot_loss[loss=0.1993, simple_loss=0.2645, pruned_loss=0.06707, over 951702.65 frames. ], batch size: 30, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:38:01,707 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-26 11:38:21,477 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52616.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 11:38:31,487 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.240e+02 1.591e+02 1.928e+02 2.293e+02 3.930e+02, threshold=3.855e+02, percent-clipped=1.0 +2023-03-26 11:38:44,946 INFO [finetune.py:976] (4/7) Epoch 10, batch 1100, loss[loss=0.1776, simple_loss=0.2542, pruned_loss=0.05052, over 4795.00 frames. ], tot_loss[loss=0.1998, simple_loss=0.2653, pruned_loss=0.06713, over 951364.48 frames. ], batch size: 51, lr: 3.76e-03, grad_scale: 32.0 +2023-03-26 11:38:59,646 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52668.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:39:14,607 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-26 11:39:16,687 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9034, 1.8650, 1.8007, 2.1040, 2.4832, 2.0458, 1.7935, 1.5220], + device='cuda:4'), covar=tensor([0.2233, 0.2147, 0.1815, 0.1595, 0.1900, 0.1206, 0.2415, 0.1946], + device='cuda:4'), in_proj_covar=tensor([0.0236, 0.0207, 0.0206, 0.0187, 0.0240, 0.0180, 0.0212, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:39:17,753 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52696.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:39:19,452 INFO [finetune.py:976] (4/7) Epoch 10, batch 1150, loss[loss=0.1629, simple_loss=0.2387, pruned_loss=0.04361, over 4744.00 frames. ], tot_loss[loss=0.2004, simple_loss=0.2661, pruned_loss=0.06737, over 953181.05 frames. ], batch size: 27, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:39:19,842 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 11:39:40,747 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 11:39:40,841 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.654e+02 1.930e+02 2.314e+02 4.484e+02, threshold=3.861e+02, percent-clipped=2.0 +2023-03-26 11:39:48,712 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52744.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:39:52,581 INFO [finetune.py:976] (4/7) Epoch 10, batch 1200, loss[loss=0.201, simple_loss=0.2564, pruned_loss=0.07285, over 4906.00 frames. ], tot_loss[loss=0.1975, simple_loss=0.2631, pruned_loss=0.06597, over 953919.70 frames. ], batch size: 36, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:39:56,854 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52752.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:40:00,363 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4383, 2.2884, 1.8806, 2.5750, 2.3877, 2.0065, 2.9633, 2.4382], + device='cuda:4'), covar=tensor([0.1501, 0.2759, 0.3454, 0.3007, 0.2949, 0.1846, 0.3766, 0.1997], + device='cuda:4'), in_proj_covar=tensor([0.0173, 0.0188, 0.0232, 0.0252, 0.0238, 0.0195, 0.0211, 0.0193], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:40:23,855 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9634, 1.7632, 1.5193, 1.5153, 1.7166, 1.6914, 1.6795, 2.4222], + device='cuda:4'), covar=tensor([0.4463, 0.5092, 0.3745, 0.4559, 0.4357, 0.2678, 0.4591, 0.1868], + device='cuda:4'), in_proj_covar=tensor([0.0286, 0.0260, 0.0222, 0.0279, 0.0243, 0.0209, 0.0246, 0.0213], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:40:35,761 INFO [finetune.py:976] (4/7) Epoch 10, batch 1250, loss[loss=0.1579, simple_loss=0.2189, pruned_loss=0.04842, over 4151.00 frames. ], tot_loss[loss=0.1964, simple_loss=0.2612, pruned_loss=0.06585, over 953008.23 frames. ], batch size: 65, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:40:47,081 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52813.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:40:49,488 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6898, 1.4573, 2.1221, 3.3800, 2.3765, 2.3746, 0.9630, 2.6498], + device='cuda:4'), covar=tensor([0.1665, 0.1478, 0.1270, 0.0561, 0.0765, 0.1527, 0.1957, 0.0531], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0117, 0.0135, 0.0165, 0.0102, 0.0139, 0.0126, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 11:41:05,426 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.514e+02 1.794e+02 2.223e+02 4.744e+02, threshold=3.588e+02, percent-clipped=2.0 +2023-03-26 11:41:19,445 INFO [finetune.py:976] (4/7) Epoch 10, batch 1300, loss[loss=0.1807, simple_loss=0.2505, pruned_loss=0.05544, over 4849.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2578, pruned_loss=0.06434, over 953509.07 frames. ], batch size: 47, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:41:19,556 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52849.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:41:20,255 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-03-26 11:41:51,930 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=52897.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:41:53,102 INFO [finetune.py:976] (4/7) Epoch 10, batch 1350, loss[loss=0.1753, simple_loss=0.2508, pruned_loss=0.04985, over 4798.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.2575, pruned_loss=0.06461, over 952442.83 frames. ], batch size: 45, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:42:02,447 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52911.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 11:42:15,926 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.950e+01 1.660e+02 2.003e+02 2.564e+02 3.985e+02, threshold=4.006e+02, percent-clipped=2.0 +2023-03-26 11:42:16,057 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4371, 2.1825, 1.9408, 0.9988, 2.1841, 1.8556, 1.6515, 2.0202], + device='cuda:4'), covar=tensor([0.0811, 0.0811, 0.1614, 0.1973, 0.1517, 0.2000, 0.2160, 0.1048], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0202, 0.0203, 0.0189, 0.0216, 0.0208, 0.0224, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:42:30,880 INFO [finetune.py:976] (4/7) Epoch 10, batch 1400, loss[loss=0.2594, simple_loss=0.3217, pruned_loss=0.09861, over 4850.00 frames. ], tot_loss[loss=0.197, simple_loss=0.2613, pruned_loss=0.0663, over 951512.13 frames. ], batch size: 47, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:42:38,088 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-26 11:42:48,553 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52968.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:42:49,478 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 11:43:11,423 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5456, 3.5032, 3.5032, 1.8381, 3.7308, 2.6961, 0.8132, 2.4636], + device='cuda:4'), covar=tensor([0.2856, 0.1983, 0.1447, 0.3010, 0.0953, 0.0977, 0.4250, 0.1390], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0175, 0.0161, 0.0129, 0.0157, 0.0123, 0.0146, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 11:43:14,419 INFO [finetune.py:976] (4/7) Epoch 10, batch 1450, loss[loss=0.2198, simple_loss=0.291, pruned_loss=0.07433, over 4811.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.2635, pruned_loss=0.06646, over 953217.45 frames. ], batch size: 38, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:43:22,111 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4702, 2.2034, 1.8012, 2.4245, 2.3737, 1.9711, 2.7754, 2.2833], + device='cuda:4'), covar=tensor([0.1378, 0.2658, 0.3566, 0.3325, 0.2900, 0.1905, 0.3844, 0.2280], + device='cuda:4'), in_proj_covar=tensor([0.0173, 0.0188, 0.0232, 0.0252, 0.0238, 0.0195, 0.0211, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:43:35,022 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53016.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:43:45,115 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.022e+02 1.605e+02 1.913e+02 2.318e+02 4.347e+02, threshold=3.826e+02, percent-clipped=3.0 +2023-03-26 11:43:55,916 INFO [finetune.py:976] (4/7) Epoch 10, batch 1500, loss[loss=0.2186, simple_loss=0.2774, pruned_loss=0.0799, over 4169.00 frames. ], tot_loss[loss=0.2011, simple_loss=0.2662, pruned_loss=0.068, over 953152.42 frames. ], batch size: 65, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:44:00,679 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53056.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:44:29,473 INFO [finetune.py:976] (4/7) Epoch 10, batch 1550, loss[loss=0.1692, simple_loss=0.2348, pruned_loss=0.05182, over 4727.00 frames. ], tot_loss[loss=0.201, simple_loss=0.2661, pruned_loss=0.06791, over 955075.85 frames. ], batch size: 23, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:44:35,501 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53108.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:44:41,523 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53117.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:44:52,483 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.115e+02 1.652e+02 2.005e+02 2.543e+02 4.651e+02, threshold=4.009e+02, percent-clipped=4.0 +2023-03-26 11:45:03,285 INFO [finetune.py:976] (4/7) Epoch 10, batch 1600, loss[loss=0.1934, simple_loss=0.2663, pruned_loss=0.06024, over 4830.00 frames. ], tot_loss[loss=0.1972, simple_loss=0.2623, pruned_loss=0.06606, over 955088.16 frames. ], batch size: 33, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:45:48,101 INFO [finetune.py:976] (4/7) Epoch 10, batch 1650, loss[loss=0.1581, simple_loss=0.2203, pruned_loss=0.048, over 4908.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.2598, pruned_loss=0.06521, over 954816.77 frames. ], batch size: 46, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:45:50,824 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0881, 1.8904, 1.6627, 1.8783, 1.8057, 1.8102, 1.8197, 2.6738], + device='cuda:4'), covar=tensor([0.3999, 0.4676, 0.3546, 0.4349, 0.4409, 0.2506, 0.4521, 0.1599], + device='cuda:4'), in_proj_covar=tensor([0.0286, 0.0259, 0.0222, 0.0279, 0.0243, 0.0208, 0.0244, 0.0212], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:45:52,555 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1233, 3.5883, 3.8284, 3.8712, 3.9145, 3.7130, 4.1628, 1.8311], + device='cuda:4'), covar=tensor([0.0696, 0.0783, 0.0687, 0.0967, 0.1007, 0.1080, 0.0618, 0.4468], + device='cuda:4'), in_proj_covar=tensor([0.0346, 0.0242, 0.0274, 0.0289, 0.0326, 0.0279, 0.0298, 0.0292], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:45:56,035 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53211.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:46:10,716 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.364e+01 1.592e+02 1.774e+02 2.189e+02 3.836e+02, threshold=3.549e+02, percent-clipped=0.0 +2023-03-26 11:46:16,366 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53241.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:46:23,558 INFO [finetune.py:976] (4/7) Epoch 10, batch 1700, loss[loss=0.1579, simple_loss=0.227, pruned_loss=0.04436, over 4747.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2573, pruned_loss=0.0641, over 954276.12 frames. ], batch size: 27, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:46:29,716 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53259.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 11:46:36,982 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3982, 1.4460, 1.8225, 1.7378, 1.6314, 3.2216, 1.2712, 1.5484], + device='cuda:4'), covar=tensor([0.1010, 0.1646, 0.1112, 0.0959, 0.1413, 0.0262, 0.1453, 0.1608], + device='cuda:4'), in_proj_covar=tensor([0.0077, 0.0081, 0.0075, 0.0078, 0.0091, 0.0083, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 11:46:37,006 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4207, 1.5645, 1.2000, 1.5175, 1.7852, 1.5887, 1.4020, 1.3511], + device='cuda:4'), covar=tensor([0.0390, 0.0302, 0.0632, 0.0272, 0.0180, 0.0590, 0.0335, 0.0411], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0108, 0.0138, 0.0114, 0.0101, 0.0102, 0.0091, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.0149e-05, 8.4498e-05, 1.1012e-04, 8.9359e-05, 7.9097e-05, 7.5320e-05, + 6.8914e-05, 8.2951e-05], device='cuda:4') +2023-03-26 11:46:44,549 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9617, 1.8475, 1.8642, 1.9195, 1.4201, 3.6035, 1.6844, 2.2075], + device='cuda:4'), covar=tensor([0.2952, 0.2130, 0.1926, 0.2099, 0.1759, 0.0213, 0.2409, 0.1118], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0115, 0.0120, 0.0123, 0.0116, 0.0099, 0.0099, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 11:46:56,426 INFO [finetune.py:976] (4/7) Epoch 10, batch 1750, loss[loss=0.2046, simple_loss=0.2819, pruned_loss=0.06366, over 4903.00 frames. ], tot_loss[loss=0.1958, simple_loss=0.2603, pruned_loss=0.06563, over 955631.18 frames. ], batch size: 43, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:46:58,854 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53302.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:47:00,070 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53304.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:47:07,266 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53315.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:47:18,948 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.172e+02 1.605e+02 1.832e+02 2.176e+02 4.638e+02, threshold=3.664e+02, percent-clipped=2.0 +2023-03-26 11:47:29,901 INFO [finetune.py:976] (4/7) Epoch 10, batch 1800, loss[loss=0.1496, simple_loss=0.2135, pruned_loss=0.0428, over 4738.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2617, pruned_loss=0.06579, over 953831.62 frames. ], batch size: 23, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:47:45,418 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53365.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:47:46,659 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53367.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:47:56,026 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53376.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:48:26,357 INFO [finetune.py:976] (4/7) Epoch 10, batch 1850, loss[loss=0.2369, simple_loss=0.2953, pruned_loss=0.08931, over 4164.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.264, pruned_loss=0.06713, over 952646.99 frames. ], batch size: 66, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:48:32,716 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53408.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:48:35,096 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53412.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:48:51,321 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53428.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 11:48:58,641 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.752e+02 2.111e+02 2.637e+02 7.323e+02, threshold=4.222e+02, percent-clipped=6.0 +2023-03-26 11:49:10,459 INFO [finetune.py:976] (4/7) Epoch 10, batch 1900, loss[loss=0.2091, simple_loss=0.2802, pruned_loss=0.06907, over 4907.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.265, pruned_loss=0.06689, over 952074.82 frames. ], batch size: 46, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:49:14,807 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53456.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:49:43,867 INFO [finetune.py:976] (4/7) Epoch 10, batch 1950, loss[loss=0.2258, simple_loss=0.2763, pruned_loss=0.08763, over 4934.00 frames. ], tot_loss[loss=0.1979, simple_loss=0.2637, pruned_loss=0.06606, over 952728.80 frames. ], batch size: 38, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:49:58,808 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7742, 3.7392, 3.6279, 1.7280, 3.8460, 2.7448, 1.2633, 2.5046], + device='cuda:4'), covar=tensor([0.2229, 0.2206, 0.1442, 0.3758, 0.1044, 0.1170, 0.4101, 0.1793], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0174, 0.0161, 0.0129, 0.0157, 0.0123, 0.0146, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 11:50:09,884 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.065e+02 1.535e+02 1.778e+02 2.101e+02 3.650e+02, threshold=3.555e+02, percent-clipped=0.0 +2023-03-26 11:50:29,330 INFO [finetune.py:976] (4/7) Epoch 10, batch 2000, loss[loss=0.1797, simple_loss=0.2436, pruned_loss=0.05791, over 4758.00 frames. ], tot_loss[loss=0.1969, simple_loss=0.2618, pruned_loss=0.06595, over 952636.13 frames. ], batch size: 26, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:50:53,155 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-03-26 11:51:22,349 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53597.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:51:23,506 INFO [finetune.py:976] (4/7) Epoch 10, batch 2050, loss[loss=0.1425, simple_loss=0.2127, pruned_loss=0.0361, over 4803.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2575, pruned_loss=0.06434, over 954302.41 frames. ], batch size: 25, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:51:30,503 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1985, 2.0745, 1.6548, 1.9909, 2.0968, 1.7740, 2.4118, 2.1806], + device='cuda:4'), covar=tensor([0.1306, 0.2269, 0.3245, 0.2935, 0.2592, 0.1655, 0.3406, 0.1784], + device='cuda:4'), in_proj_covar=tensor([0.0175, 0.0189, 0.0233, 0.0254, 0.0239, 0.0196, 0.0212, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:51:44,833 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.628e+02 1.908e+02 2.249e+02 5.707e+02, threshold=3.816e+02, percent-clipped=1.0 +2023-03-26 11:51:56,176 INFO [finetune.py:976] (4/7) Epoch 10, batch 2100, loss[loss=0.1766, simple_loss=0.2583, pruned_loss=0.04743, over 4854.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2583, pruned_loss=0.06455, over 955966.33 frames. ], batch size: 44, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:52:03,970 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53660.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:52:13,508 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53671.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:52:37,542 INFO [finetune.py:976] (4/7) Epoch 10, batch 2150, loss[loss=0.2037, simple_loss=0.2548, pruned_loss=0.07627, over 4208.00 frames. ], tot_loss[loss=0.197, simple_loss=0.262, pruned_loss=0.06607, over 954134.35 frames. ], batch size: 18, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:52:51,228 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 11:52:51,493 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5320, 1.4127, 1.6219, 1.7255, 1.5097, 3.2313, 1.2742, 1.4847], + device='cuda:4'), covar=tensor([0.0978, 0.1911, 0.1204, 0.1010, 0.1712, 0.0271, 0.1627, 0.1761], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0091, 0.0082, 0.0084, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 11:52:52,732 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53712.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:53:04,001 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53723.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 11:53:11,747 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7782, 1.9558, 1.1812, 2.6991, 3.0407, 2.1764, 2.5652, 2.5839], + device='cuda:4'), covar=tensor([0.1141, 0.1823, 0.2014, 0.0954, 0.1443, 0.1736, 0.1151, 0.1603], + device='cuda:4'), in_proj_covar=tensor([0.0088, 0.0094, 0.0111, 0.0092, 0.0121, 0.0095, 0.0098, 0.0091], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 11:53:19,766 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.764e+02 2.057e+02 2.459e+02 5.535e+02, threshold=4.114e+02, percent-clipped=2.0 +2023-03-26 11:53:34,458 INFO [finetune.py:976] (4/7) Epoch 10, batch 2200, loss[loss=0.1809, simple_loss=0.2489, pruned_loss=0.05639, over 4839.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2645, pruned_loss=0.06613, over 956264.43 frames. ], batch size: 49, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:53:38,690 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 11:53:43,153 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53760.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:53:46,315 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5127, 1.5929, 1.2565, 1.4787, 1.8329, 1.7373, 1.5235, 1.3827], + device='cuda:4'), covar=tensor([0.0332, 0.0276, 0.0563, 0.0304, 0.0208, 0.0462, 0.0356, 0.0379], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0108, 0.0139, 0.0114, 0.0101, 0.0102, 0.0091, 0.0107], + device='cuda:4'), out_proj_covar=tensor([7.0195e-05, 8.4707e-05, 1.1027e-04, 8.9436e-05, 7.9320e-05, 7.5306e-05, + 6.9031e-05, 8.2653e-05], device='cuda:4') +2023-03-26 11:53:51,205 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.9354, 0.8114, 0.7099, 0.9957, 1.0910, 1.0372, 0.8950, 0.8168], + device='cuda:4'), covar=tensor([0.0321, 0.0286, 0.0547, 0.0279, 0.0280, 0.0358, 0.0269, 0.0394], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0108, 0.0139, 0.0114, 0.0101, 0.0102, 0.0091, 0.0107], + device='cuda:4'), out_proj_covar=tensor([7.0232e-05, 8.4737e-05, 1.1033e-04, 8.9480e-05, 7.9376e-05, 7.5411e-05, + 6.9040e-05, 8.2702e-05], device='cuda:4') +2023-03-26 11:54:00,937 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-26 11:54:07,986 INFO [finetune.py:976] (4/7) Epoch 10, batch 2250, loss[loss=0.2629, simple_loss=0.3233, pruned_loss=0.1013, over 4758.00 frames. ], tot_loss[loss=0.1992, simple_loss=0.2653, pruned_loss=0.06654, over 955156.80 frames. ], batch size: 54, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:54:12,572 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.07 vs. limit=5.0 +2023-03-26 11:54:19,270 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6312, 1.6019, 1.8913, 1.8822, 1.6232, 3.6743, 1.4085, 1.7012], + device='cuda:4'), covar=tensor([0.0945, 0.1760, 0.1093, 0.0996, 0.1657, 0.0217, 0.1497, 0.1619], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0091, 0.0083, 0.0084, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 11:54:30,210 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.658e+02 1.958e+02 2.430e+02 3.560e+02, threshold=3.915e+02, percent-clipped=0.0 +2023-03-26 11:54:41,561 INFO [finetune.py:976] (4/7) Epoch 10, batch 2300, loss[loss=0.2161, simple_loss=0.2729, pruned_loss=0.0796, over 4747.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2649, pruned_loss=0.06597, over 956429.53 frames. ], batch size: 54, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:54:42,842 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0244, 1.9680, 2.1051, 1.3354, 2.0463, 2.3468, 2.1758, 1.6692], + device='cuda:4'), covar=tensor([0.0505, 0.0561, 0.0583, 0.0862, 0.0689, 0.0424, 0.0467, 0.0968], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0132, 0.0142, 0.0123, 0.0118, 0.0141, 0.0141, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:55:15,989 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53897.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:55:17,109 INFO [finetune.py:976] (4/7) Epoch 10, batch 2350, loss[loss=0.1665, simple_loss=0.2337, pruned_loss=0.04972, over 4824.00 frames. ], tot_loss[loss=0.1973, simple_loss=0.2631, pruned_loss=0.06571, over 954094.00 frames. ], batch size: 33, lr: 3.75e-03, grad_scale: 32.0 +2023-03-26 11:55:47,268 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.820e+01 1.627e+02 1.969e+02 2.442e+02 4.599e+02, threshold=3.938e+02, percent-clipped=2.0 +2023-03-26 11:55:58,287 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=53945.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:56:05,529 INFO [finetune.py:976] (4/7) Epoch 10, batch 2400, loss[loss=0.1815, simple_loss=0.244, pruned_loss=0.05951, over 4865.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2593, pruned_loss=0.06468, over 953050.74 frames. ], batch size: 31, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 11:56:15,968 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53960.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:56:24,609 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53971.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:56:31,198 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5154, 1.3950, 1.8840, 2.8514, 1.9375, 2.0548, 0.8236, 2.2389], + device='cuda:4'), covar=tensor([0.1723, 0.1501, 0.1229, 0.0655, 0.0834, 0.1489, 0.1896, 0.0725], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0117, 0.0134, 0.0165, 0.0101, 0.0138, 0.0127, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 11:56:35,434 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2474, 1.6720, 2.1402, 2.0556, 1.7903, 1.8324, 2.0438, 1.9309], + device='cuda:4'), covar=tensor([0.3983, 0.5085, 0.3881, 0.4951, 0.5848, 0.4549, 0.5862, 0.3816], + device='cuda:4'), in_proj_covar=tensor([0.0233, 0.0239, 0.0251, 0.0255, 0.0249, 0.0225, 0.0272, 0.0228], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:56:41,931 INFO [finetune.py:976] (4/7) Epoch 10, batch 2450, loss[loss=0.123, simple_loss=0.2004, pruned_loss=0.0228, over 4763.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.257, pruned_loss=0.06393, over 954056.70 frames. ], batch size: 27, lr: 3.74e-03, grad_scale: 64.0 +2023-03-26 11:56:49,145 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=54008.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:56:56,921 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=54019.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:56:59,868 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=54023.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 11:57:04,735 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1218, 2.0824, 1.6532, 2.1173, 2.0752, 1.7534, 2.5035, 2.1482], + device='cuda:4'), covar=tensor([0.1416, 0.2562, 0.3240, 0.3060, 0.2651, 0.1722, 0.3581, 0.2023], + device='cuda:4'), in_proj_covar=tensor([0.0175, 0.0188, 0.0233, 0.0253, 0.0240, 0.0196, 0.0213, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:57:05,184 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.615e+02 1.942e+02 2.268e+02 4.833e+02, threshold=3.884e+02, percent-clipped=2.0 +2023-03-26 11:57:14,116 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-26 11:57:16,020 INFO [finetune.py:976] (4/7) Epoch 10, batch 2500, loss[loss=0.1915, simple_loss=0.2608, pruned_loss=0.06103, over 4777.00 frames. ], tot_loss[loss=0.1945, simple_loss=0.2589, pruned_loss=0.06507, over 953642.70 frames. ], batch size: 29, lr: 3.74e-03, grad_scale: 64.0 +2023-03-26 11:57:42,141 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=54071.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 11:58:00,118 INFO [finetune.py:976] (4/7) Epoch 10, batch 2550, loss[loss=0.2034, simple_loss=0.2825, pruned_loss=0.06214, over 4912.00 frames. ], tot_loss[loss=0.1965, simple_loss=0.2618, pruned_loss=0.06565, over 953502.09 frames. ], batch size: 37, lr: 3.74e-03, grad_scale: 64.0 +2023-03-26 11:58:35,815 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.936e+01 1.671e+02 2.051e+02 2.356e+02 3.900e+02, threshold=4.103e+02, percent-clipped=1.0 +2023-03-26 11:58:46,749 INFO [finetune.py:976] (4/7) Epoch 10, batch 2600, loss[loss=0.1952, simple_loss=0.2654, pruned_loss=0.06247, over 4833.00 frames. ], tot_loss[loss=0.1975, simple_loss=0.263, pruned_loss=0.06596, over 952086.84 frames. ], batch size: 44, lr: 3.74e-03, grad_scale: 64.0 +2023-03-26 11:59:19,473 INFO [finetune.py:976] (4/7) Epoch 10, batch 2650, loss[loss=0.2148, simple_loss=0.281, pruned_loss=0.07435, over 4861.00 frames. ], tot_loss[loss=0.199, simple_loss=0.2646, pruned_loss=0.06673, over 953709.01 frames. ], batch size: 34, lr: 3.74e-03, grad_scale: 64.0 +2023-03-26 11:59:43,597 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.87 vs. limit=5.0 +2023-03-26 11:59:43,742 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.585e+01 1.566e+02 1.779e+02 2.159e+02 3.883e+02, threshold=3.557e+02, percent-clipped=0.0 +2023-03-26 11:59:46,355 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1898, 2.0908, 1.8802, 2.3438, 2.6534, 2.1978, 2.0163, 1.7427], + device='cuda:4'), covar=tensor([0.2092, 0.1934, 0.1804, 0.1567, 0.1830, 0.1135, 0.2167, 0.1813], + device='cuda:4'), in_proj_covar=tensor([0.0237, 0.0208, 0.0208, 0.0189, 0.0241, 0.0180, 0.0214, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 11:59:53,473 INFO [finetune.py:976] (4/7) Epoch 10, batch 2700, loss[loss=0.202, simple_loss=0.2665, pruned_loss=0.06872, over 4710.00 frames. ], tot_loss[loss=0.1986, simple_loss=0.2643, pruned_loss=0.0665, over 953859.71 frames. ], batch size: 59, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 11:59:55,911 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7348, 1.6380, 1.6772, 1.7319, 1.2069, 3.8415, 1.4934, 2.0467], + device='cuda:4'), covar=tensor([0.3400, 0.2540, 0.2093, 0.2314, 0.1808, 0.0158, 0.2618, 0.1340], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0116, 0.0120, 0.0124, 0.0116, 0.0099, 0.0099, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 12:00:08,948 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 12:00:19,409 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3942, 1.4052, 1.6035, 1.7177, 1.4710, 3.1298, 1.2074, 1.5216], + device='cuda:4'), covar=tensor([0.1042, 0.1828, 0.1196, 0.0980, 0.1648, 0.0282, 0.1597, 0.1720], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0080, 0.0075, 0.0077, 0.0090, 0.0082, 0.0084, 0.0078], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 12:00:26,576 INFO [finetune.py:976] (4/7) Epoch 10, batch 2750, loss[loss=0.2038, simple_loss=0.2686, pruned_loss=0.06946, over 4768.00 frames. ], tot_loss[loss=0.1959, simple_loss=0.2611, pruned_loss=0.06532, over 954642.35 frames. ], batch size: 59, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:00:50,902 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.545e+02 1.929e+02 2.415e+02 3.548e+02, threshold=3.859e+02, percent-clipped=0.0 +2023-03-26 12:01:01,566 INFO [finetune.py:976] (4/7) Epoch 10, batch 2800, loss[loss=0.234, simple_loss=0.2961, pruned_loss=0.086, over 4819.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2589, pruned_loss=0.06481, over 954715.39 frames. ], batch size: 41, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:01:02,301 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4239, 1.4882, 1.8499, 1.8215, 1.5394, 3.5484, 1.2730, 1.6299], + device='cuda:4'), covar=tensor([0.1056, 0.1819, 0.1096, 0.0981, 0.1626, 0.0229, 0.1549, 0.1707], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0080, 0.0075, 0.0077, 0.0090, 0.0082, 0.0084, 0.0078], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:4') +2023-03-26 12:01:10,065 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1739, 1.9945, 1.6817, 1.9616, 1.8725, 1.8059, 1.8431, 2.6807], + device='cuda:4'), covar=tensor([0.4382, 0.5063, 0.3751, 0.4741, 0.4535, 0.2716, 0.4910, 0.1793], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0259, 0.0221, 0.0278, 0.0242, 0.0208, 0.0245, 0.0211], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:01:48,151 INFO [finetune.py:976] (4/7) Epoch 10, batch 2850, loss[loss=0.1683, simple_loss=0.2196, pruned_loss=0.05848, over 4726.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2562, pruned_loss=0.06416, over 953800.35 frames. ], batch size: 23, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:02:08,218 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4470, 1.2768, 1.8306, 2.7529, 1.8924, 1.9966, 0.8121, 2.2004], + device='cuda:4'), covar=tensor([0.1962, 0.1813, 0.1510, 0.0950, 0.0969, 0.1711, 0.2133, 0.0855], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0116, 0.0134, 0.0164, 0.0101, 0.0137, 0.0126, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 12:02:10,453 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.958e+01 1.628e+02 1.894e+02 2.190e+02 3.699e+02, threshold=3.787e+02, percent-clipped=0.0 +2023-03-26 12:02:11,850 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1506, 1.8851, 1.6728, 2.0005, 1.8166, 1.7857, 1.7751, 2.5957], + device='cuda:4'), covar=tensor([0.4890, 0.5636, 0.4242, 0.5107, 0.5255, 0.3059, 0.5221, 0.2106], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0259, 0.0222, 0.0279, 0.0243, 0.0208, 0.0245, 0.0211], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:02:22,200 INFO [finetune.py:976] (4/7) Epoch 10, batch 2900, loss[loss=0.1928, simple_loss=0.2561, pruned_loss=0.06481, over 4833.00 frames. ], tot_loss[loss=0.196, simple_loss=0.2602, pruned_loss=0.06596, over 951353.32 frames. ], batch size: 30, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:02:23,619 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-03-26 12:02:39,666 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4245, 1.4052, 1.6275, 1.6715, 1.4762, 3.2618, 1.2442, 1.5039], + device='cuda:4'), covar=tensor([0.1062, 0.1886, 0.1337, 0.1022, 0.1660, 0.0251, 0.1556, 0.1768], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0091, 0.0082, 0.0084, 0.0078], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 12:02:57,316 INFO [finetune.py:976] (4/7) Epoch 10, batch 2950, loss[loss=0.1995, simple_loss=0.2736, pruned_loss=0.06267, over 4844.00 frames. ], tot_loss[loss=0.1987, simple_loss=0.2638, pruned_loss=0.06678, over 952327.77 frames. ], batch size: 47, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:03:15,846 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6709, 1.6530, 1.3141, 1.6605, 1.9670, 1.8695, 1.5848, 1.4465], + device='cuda:4'), covar=tensor([0.0276, 0.0295, 0.0552, 0.0276, 0.0188, 0.0405, 0.0335, 0.0367], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0108, 0.0137, 0.0113, 0.0100, 0.0102, 0.0090, 0.0106], + device='cuda:4'), out_proj_covar=tensor([7.0367e-05, 8.4240e-05, 1.0917e-04, 8.8402e-05, 7.8580e-05, 7.5335e-05, + 6.8439e-05, 8.1886e-05], device='cuda:4') +2023-03-26 12:03:18,744 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.146e+02 1.694e+02 2.010e+02 2.318e+02 4.609e+02, threshold=4.019e+02, percent-clipped=2.0 +2023-03-26 12:03:30,486 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9511, 1.7662, 1.5471, 1.7889, 1.6732, 1.6689, 1.6571, 2.4395], + device='cuda:4'), covar=tensor([0.5068, 0.5310, 0.4341, 0.4917, 0.4985, 0.2902, 0.4868, 0.1960], + device='cuda:4'), in_proj_covar=tensor([0.0286, 0.0259, 0.0222, 0.0279, 0.0243, 0.0208, 0.0246, 0.0212], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:03:39,999 INFO [finetune.py:976] (4/7) Epoch 10, batch 3000, loss[loss=0.1809, simple_loss=0.2549, pruned_loss=0.05347, over 4927.00 frames. ], tot_loss[loss=0.2006, simple_loss=0.2659, pruned_loss=0.06765, over 953845.49 frames. ], batch size: 38, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:03:39,999 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 12:03:48,272 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4966, 1.3198, 1.3754, 1.4140, 1.6904, 1.6108, 1.4712, 1.3805], + device='cuda:4'), covar=tensor([0.0395, 0.0245, 0.0520, 0.0296, 0.0292, 0.0421, 0.0301, 0.0350], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0109, 0.0138, 0.0113, 0.0101, 0.0102, 0.0091, 0.0107], + device='cuda:4'), out_proj_covar=tensor([7.0816e-05, 8.4765e-05, 1.0989e-04, 8.8848e-05, 7.9031e-05, 7.5771e-05, + 6.8884e-05, 8.2397e-05], device='cuda:4') +2023-03-26 12:03:56,629 INFO [finetune.py:1010] (4/7) Epoch 10, validation: loss=0.1584, simple_loss=0.2295, pruned_loss=0.04366, over 2265189.00 frames. +2023-03-26 12:03:56,629 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 12:04:00,711 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3463, 2.2943, 1.8816, 0.8716, 2.0404, 1.8414, 1.6059, 2.1952], + device='cuda:4'), covar=tensor([0.0852, 0.0829, 0.1533, 0.2070, 0.1415, 0.2063, 0.2182, 0.0829], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0200, 0.0201, 0.0186, 0.0215, 0.0207, 0.0223, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:04:25,521 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.2201, 1.3327, 1.4137, 0.6405, 1.2714, 1.5883, 1.6193, 1.2739], + device='cuda:4'), covar=tensor([0.0873, 0.0522, 0.0455, 0.0492, 0.0452, 0.0488, 0.0274, 0.0624], + device='cuda:4'), in_proj_covar=tensor([0.0127, 0.0153, 0.0120, 0.0132, 0.0130, 0.0123, 0.0143, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.4098e-05, 1.1214e-04, 8.6482e-05, 9.5470e-05, 9.3060e-05, 8.9932e-05, + 1.0474e-04, 1.0689e-04], device='cuda:4') +2023-03-26 12:04:29,066 INFO [finetune.py:976] (4/7) Epoch 10, batch 3050, loss[loss=0.2308, simple_loss=0.2935, pruned_loss=0.08405, over 4847.00 frames. ], tot_loss[loss=0.2008, simple_loss=0.2664, pruned_loss=0.06757, over 954006.65 frames. ], batch size: 44, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:04:52,093 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.606e+02 1.839e+02 2.259e+02 4.011e+02, threshold=3.679e+02, percent-clipped=0.0 +2023-03-26 12:05:02,813 INFO [finetune.py:976] (4/7) Epoch 10, batch 3100, loss[loss=0.2074, simple_loss=0.2636, pruned_loss=0.07555, over 4742.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.2649, pruned_loss=0.06715, over 955234.74 frames. ], batch size: 27, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:05:09,039 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 12:05:14,405 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=54664.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 12:05:20,177 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 12:05:36,457 INFO [finetune.py:976] (4/7) Epoch 10, batch 3150, loss[loss=0.2054, simple_loss=0.2616, pruned_loss=0.07456, over 4925.00 frames. ], tot_loss[loss=0.1974, simple_loss=0.2621, pruned_loss=0.06638, over 956841.98 frames. ], batch size: 37, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:05:37,745 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6651, 1.5042, 2.0278, 3.1101, 2.1495, 2.2369, 1.0671, 2.4807], + device='cuda:4'), covar=tensor([0.1704, 0.1392, 0.1307, 0.0629, 0.0771, 0.1416, 0.1701, 0.0626], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0118, 0.0135, 0.0166, 0.0101, 0.0138, 0.0127, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 12:05:42,905 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6656, 1.5387, 1.9871, 1.2274, 1.7470, 1.7939, 1.5127, 2.1436], + device='cuda:4'), covar=tensor([0.1269, 0.2142, 0.1073, 0.1923, 0.0943, 0.1483, 0.2728, 0.0797], + device='cuda:4'), in_proj_covar=tensor([0.0196, 0.0203, 0.0191, 0.0189, 0.0176, 0.0213, 0.0214, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:05:54,666 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=54725.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 12:05:59,383 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.653e+02 1.993e+02 2.298e+02 5.311e+02, threshold=3.986e+02, percent-clipped=2.0 +2023-03-26 12:06:10,112 INFO [finetune.py:976] (4/7) Epoch 10, batch 3200, loss[loss=0.2169, simple_loss=0.28, pruned_loss=0.07685, over 4828.00 frames. ], tot_loss[loss=0.1947, simple_loss=0.2586, pruned_loss=0.0654, over 956862.39 frames. ], batch size: 40, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:06:53,304 INFO [finetune.py:976] (4/7) Epoch 10, batch 3250, loss[loss=0.1971, simple_loss=0.2642, pruned_loss=0.06494, over 4862.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.2589, pruned_loss=0.06566, over 954968.57 frames. ], batch size: 34, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:07:26,173 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.711e+02 2.094e+02 2.546e+02 5.601e+02, threshold=4.189e+02, percent-clipped=2.0 +2023-03-26 12:07:28,433 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4949, 1.4306, 1.8619, 2.8118, 1.9852, 2.0127, 0.8299, 2.2632], + device='cuda:4'), covar=tensor([0.1715, 0.1410, 0.1248, 0.0608, 0.0757, 0.1586, 0.1786, 0.0588], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0117, 0.0135, 0.0166, 0.0101, 0.0138, 0.0127, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 12:07:46,362 INFO [finetune.py:976] (4/7) Epoch 10, batch 3300, loss[loss=0.2058, simple_loss=0.275, pruned_loss=0.06829, over 4854.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.2627, pruned_loss=0.06683, over 955274.32 frames. ], batch size: 44, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:07:55,833 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=54862.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:07:57,402 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 12:08:06,190 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3381, 1.3434, 1.6019, 2.4364, 1.6996, 2.0935, 1.0378, 1.9693], + device='cuda:4'), covar=tensor([0.1821, 0.1500, 0.1203, 0.0847, 0.0885, 0.1145, 0.1505, 0.0759], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0117, 0.0134, 0.0165, 0.0101, 0.0137, 0.0126, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 12:08:20,113 INFO [finetune.py:976] (4/7) Epoch 10, batch 3350, loss[loss=0.2343, simple_loss=0.2912, pruned_loss=0.08864, over 4843.00 frames. ], tot_loss[loss=0.2005, simple_loss=0.2654, pruned_loss=0.06782, over 953502.09 frames. ], batch size: 49, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:08:24,421 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0327, 1.8339, 1.6108, 1.8824, 1.7964, 1.7347, 1.7816, 2.4976], + device='cuda:4'), covar=tensor([0.4407, 0.5569, 0.3894, 0.4724, 0.4855, 0.2691, 0.4793, 0.1793], + device='cuda:4'), in_proj_covar=tensor([0.0284, 0.0259, 0.0222, 0.0278, 0.0242, 0.0208, 0.0245, 0.0211], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:08:47,114 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0610, 1.9483, 1.7736, 2.1565, 2.6247, 2.2595, 1.7774, 1.6779], + device='cuda:4'), covar=tensor([0.2084, 0.2008, 0.1855, 0.1600, 0.1569, 0.1081, 0.2323, 0.1938], + device='cuda:4'), in_proj_covar=tensor([0.0236, 0.0207, 0.0207, 0.0189, 0.0239, 0.0180, 0.0213, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:08:48,197 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=54923.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 12:08:57,771 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.164e+02 1.693e+02 1.965e+02 2.442e+02 4.084e+02, threshold=3.930e+02, percent-clipped=0.0 +2023-03-26 12:09:07,544 INFO [finetune.py:976] (4/7) Epoch 10, batch 3400, loss[loss=0.1969, simple_loss=0.2692, pruned_loss=0.06229, over 4856.00 frames. ], tot_loss[loss=0.2014, simple_loss=0.2665, pruned_loss=0.06815, over 955191.44 frames. ], batch size: 44, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:09:10,673 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4306, 2.8716, 2.5046, 1.9029, 2.6394, 2.7842, 2.8012, 2.4776], + device='cuda:4'), covar=tensor([0.0681, 0.0574, 0.0735, 0.0907, 0.0611, 0.0789, 0.0678, 0.0963], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0133, 0.0143, 0.0123, 0.0119, 0.0142, 0.0142, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:09:38,064 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7568, 1.6049, 1.4819, 1.8311, 2.2044, 1.8497, 1.2174, 1.4101], + device='cuda:4'), covar=tensor([0.2214, 0.2158, 0.2038, 0.1633, 0.1679, 0.1309, 0.2804, 0.2069], + device='cuda:4'), in_proj_covar=tensor([0.0235, 0.0207, 0.0206, 0.0188, 0.0239, 0.0180, 0.0213, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:09:56,947 INFO [finetune.py:976] (4/7) Epoch 10, batch 3450, loss[loss=0.1942, simple_loss=0.253, pruned_loss=0.06773, over 4813.00 frames. ], tot_loss[loss=0.2, simple_loss=0.2655, pruned_loss=0.06727, over 956118.39 frames. ], batch size: 30, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:10:16,050 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55020.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 12:10:18,274 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-26 12:10:30,757 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.349e+01 1.534e+02 1.957e+02 2.350e+02 5.428e+02, threshold=3.914e+02, percent-clipped=3.0 +2023-03-26 12:10:50,316 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4582, 1.5226, 1.9102, 1.8242, 1.6306, 3.5676, 1.4427, 1.6778], + device='cuda:4'), covar=tensor([0.1007, 0.1839, 0.1069, 0.0973, 0.1584, 0.0229, 0.1425, 0.1685], + device='cuda:4'), in_proj_covar=tensor([0.0077, 0.0082, 0.0076, 0.0078, 0.0092, 0.0082, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 12:10:51,433 INFO [finetune.py:976] (4/7) Epoch 10, batch 3500, loss[loss=0.1713, simple_loss=0.236, pruned_loss=0.05329, over 4778.00 frames. ], tot_loss[loss=0.1987, simple_loss=0.2638, pruned_loss=0.06679, over 955958.18 frames. ], batch size: 26, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:11:08,985 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-26 12:11:22,120 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.8565, 3.3344, 3.5352, 3.7306, 3.6403, 3.4249, 3.9320, 1.1750], + device='cuda:4'), covar=tensor([0.0853, 0.0900, 0.0902, 0.0945, 0.1233, 0.1567, 0.0822, 0.5116], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0243, 0.0274, 0.0289, 0.0328, 0.0282, 0.0300, 0.0294], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:11:27,360 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.8807, 3.4351, 3.0221, 2.0469, 3.2166, 2.9000, 2.7006, 3.0744], + device='cuda:4'), covar=tensor([0.0734, 0.0613, 0.1287, 0.1735, 0.1357, 0.1430, 0.1644, 0.0913], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0202, 0.0202, 0.0187, 0.0216, 0.0207, 0.0224, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:11:36,219 INFO [finetune.py:976] (4/7) Epoch 10, batch 3550, loss[loss=0.1773, simple_loss=0.2432, pruned_loss=0.05565, over 4829.00 frames. ], tot_loss[loss=0.1963, simple_loss=0.2608, pruned_loss=0.06591, over 955989.47 frames. ], batch size: 40, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:11:37,526 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5236, 1.2959, 1.1436, 1.1674, 1.7200, 1.6866, 1.5306, 1.2790], + device='cuda:4'), covar=tensor([0.0283, 0.0367, 0.0902, 0.0421, 0.0222, 0.0412, 0.0294, 0.0433], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0109, 0.0139, 0.0114, 0.0101, 0.0103, 0.0092, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.0927e-05, 8.5410e-05, 1.1063e-04, 8.9626e-05, 7.9314e-05, 7.6471e-05, + 6.9713e-05, 8.2996e-05], device='cuda:4') +2023-03-26 12:11:58,143 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2031, 2.0390, 1.6323, 1.9816, 2.0192, 1.9291, 1.9593, 2.7513], + device='cuda:4'), covar=tensor([0.4465, 0.5078, 0.4174, 0.5043, 0.4663, 0.2776, 0.4525, 0.1976], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0259, 0.0222, 0.0278, 0.0242, 0.0209, 0.0244, 0.0212], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:11:58,573 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.369e+01 1.530e+02 1.896e+02 2.280e+02 4.793e+02, threshold=3.791e+02, percent-clipped=5.0 +2023-03-26 12:12:09,221 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=55148.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:12:09,742 INFO [finetune.py:976] (4/7) Epoch 10, batch 3600, loss[loss=0.151, simple_loss=0.221, pruned_loss=0.04051, over 4837.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2584, pruned_loss=0.06514, over 957914.62 frames. ], batch size: 30, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:12:43,417 INFO [finetune.py:976] (4/7) Epoch 10, batch 3650, loss[loss=0.2479, simple_loss=0.3092, pruned_loss=0.09336, over 4739.00 frames. ], tot_loss[loss=0.196, simple_loss=0.2602, pruned_loss=0.06588, over 955741.23 frames. ], batch size: 59, lr: 3.74e-03, grad_scale: 32.0 +2023-03-26 12:12:49,699 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55209.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:12:52,573 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=55211.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:12:56,772 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55218.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 12:13:05,136 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5656, 1.9371, 1.4913, 1.5104, 2.0997, 2.0413, 1.7527, 1.7600], + device='cuda:4'), covar=tensor([0.0407, 0.0310, 0.0536, 0.0360, 0.0277, 0.0608, 0.0502, 0.0392], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0109, 0.0138, 0.0114, 0.0101, 0.0102, 0.0091, 0.0107], + device='cuda:4'), out_proj_covar=tensor([7.0504e-05, 8.4845e-05, 1.0998e-04, 8.9048e-05, 7.8876e-05, 7.5876e-05, + 6.9269e-05, 8.2553e-05], device='cuda:4') +2023-03-26 12:13:14,908 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.143e+02 1.616e+02 1.938e+02 2.270e+02 4.700e+02, threshold=3.875e+02, percent-clipped=1.0 +2023-03-26 12:13:26,510 INFO [finetune.py:976] (4/7) Epoch 10, batch 3700, loss[loss=0.2208, simple_loss=0.2886, pruned_loss=0.0765, over 4800.00 frames. ], tot_loss[loss=0.1995, simple_loss=0.2642, pruned_loss=0.06734, over 956506.94 frames. ], batch size: 45, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:13:40,335 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55272.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:14:00,010 INFO [finetune.py:976] (4/7) Epoch 10, batch 3750, loss[loss=0.2025, simple_loss=0.2694, pruned_loss=0.06783, over 4889.00 frames. ], tot_loss[loss=0.2008, simple_loss=0.2658, pruned_loss=0.06792, over 955280.64 frames. ], batch size: 37, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:14:16,777 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55320.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 12:14:33,808 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.583e+02 1.835e+02 2.150e+02 3.880e+02, threshold=3.669e+02, percent-clipped=1.0 +2023-03-26 12:14:45,562 INFO [finetune.py:976] (4/7) Epoch 10, batch 3800, loss[loss=0.1972, simple_loss=0.2569, pruned_loss=0.06878, over 4765.00 frames. ], tot_loss[loss=0.2008, simple_loss=0.266, pruned_loss=0.06783, over 954384.62 frames. ], batch size: 28, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:14:57,817 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=55368.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 12:15:27,045 INFO [finetune.py:976] (4/7) Epoch 10, batch 3850, loss[loss=0.1476, simple_loss=0.2226, pruned_loss=0.03628, over 4752.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2637, pruned_loss=0.0666, over 954525.14 frames. ], batch size: 26, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:15:38,108 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0637, 1.0504, 1.0403, 0.4748, 0.8736, 1.1552, 1.2226, 1.0037], + device='cuda:4'), covar=tensor([0.1112, 0.0765, 0.0573, 0.0596, 0.0661, 0.0694, 0.0496, 0.0817], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0156, 0.0123, 0.0134, 0.0132, 0.0125, 0.0146, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.5531e-05, 1.1425e-04, 8.8418e-05, 9.7162e-05, 9.4680e-05, 9.0995e-05, + 1.0681e-04, 1.0779e-04], device='cuda:4') +2023-03-26 12:15:49,874 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.578e+02 1.920e+02 2.344e+02 4.809e+02, threshold=3.839e+02, percent-clipped=3.0 +2023-03-26 12:16:01,517 INFO [finetune.py:976] (4/7) Epoch 10, batch 3900, loss[loss=0.2253, simple_loss=0.2797, pruned_loss=0.08543, over 4772.00 frames. ], tot_loss[loss=0.1962, simple_loss=0.2609, pruned_loss=0.06573, over 955428.18 frames. ], batch size: 28, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:16:14,920 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9660, 1.1706, 1.8410, 1.8438, 1.6458, 1.6161, 1.7067, 1.6938], + device='cuda:4'), covar=tensor([0.3947, 0.4704, 0.4024, 0.4073, 0.5492, 0.3978, 0.4946, 0.3946], + device='cuda:4'), in_proj_covar=tensor([0.0235, 0.0239, 0.0253, 0.0257, 0.0252, 0.0229, 0.0274, 0.0230], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:16:15,511 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2577, 2.2179, 2.2768, 1.5768, 2.3517, 2.3862, 2.3272, 1.9012], + device='cuda:4'), covar=tensor([0.0545, 0.0566, 0.0654, 0.0881, 0.0518, 0.0702, 0.0588, 0.0959], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0133, 0.0143, 0.0124, 0.0120, 0.0142, 0.0143, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:16:44,730 INFO [finetune.py:976] (4/7) Epoch 10, batch 3950, loss[loss=0.201, simple_loss=0.2633, pruned_loss=0.0694, over 4780.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2572, pruned_loss=0.06443, over 954345.72 frames. ], batch size: 26, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:16:48,766 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55504.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:16:58,364 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55518.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 12:17:13,739 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.106e+02 1.580e+02 1.857e+02 2.217e+02 3.906e+02, threshold=3.714e+02, percent-clipped=1.0 +2023-03-26 12:17:35,772 INFO [finetune.py:976] (4/7) Epoch 10, batch 4000, loss[loss=0.2727, simple_loss=0.3321, pruned_loss=0.1067, over 4725.00 frames. ], tot_loss[loss=0.1939, simple_loss=0.2578, pruned_loss=0.06496, over 952782.29 frames. ], batch size: 59, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:17:48,291 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=55566.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:17:48,916 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55567.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:17:50,758 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6703, 1.1553, 0.8419, 1.5891, 2.0365, 1.3042, 1.4508, 1.5518], + device='cuda:4'), covar=tensor([0.1458, 0.2117, 0.2112, 0.1221, 0.2003, 0.1942, 0.1467, 0.2029], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0097, 0.0114, 0.0093, 0.0122, 0.0096, 0.0100, 0.0092], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 12:18:09,102 INFO [finetune.py:976] (4/7) Epoch 10, batch 4050, loss[loss=0.2735, simple_loss=0.3244, pruned_loss=0.1113, over 4908.00 frames. ], tot_loss[loss=0.2003, simple_loss=0.2641, pruned_loss=0.06821, over 952090.02 frames. ], batch size: 37, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:18:34,628 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.736e+02 2.138e+02 2.510e+02 4.140e+02, threshold=4.276e+02, percent-clipped=4.0 +2023-03-26 12:18:44,820 INFO [finetune.py:976] (4/7) Epoch 10, batch 4100, loss[loss=0.2022, simple_loss=0.2618, pruned_loss=0.07129, over 4921.00 frames. ], tot_loss[loss=0.2002, simple_loss=0.2649, pruned_loss=0.06776, over 951731.63 frames. ], batch size: 42, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:19:12,858 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0080, 1.8940, 1.5617, 1.8225, 1.8233, 1.7811, 1.8041, 2.5693], + device='cuda:4'), covar=tensor([0.4695, 0.5080, 0.3872, 0.5175, 0.5087, 0.2844, 0.4545, 0.1845], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0259, 0.0222, 0.0278, 0.0242, 0.0209, 0.0245, 0.0212], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:19:17,497 INFO [finetune.py:976] (4/7) Epoch 10, batch 4150, loss[loss=0.2228, simple_loss=0.295, pruned_loss=0.07524, over 4905.00 frames. ], tot_loss[loss=0.2016, simple_loss=0.2666, pruned_loss=0.06832, over 952258.45 frames. ], batch size: 37, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:19:49,995 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.152e+02 1.688e+02 2.035e+02 2.462e+02 3.895e+02, threshold=4.069e+02, percent-clipped=0.0 +2023-03-26 12:19:59,098 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=55748.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:19:59,614 INFO [finetune.py:976] (4/7) Epoch 10, batch 4200, loss[loss=0.1644, simple_loss=0.2249, pruned_loss=0.05193, over 4817.00 frames. ], tot_loss[loss=0.201, simple_loss=0.2665, pruned_loss=0.06775, over 955079.71 frames. ], batch size: 30, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:20:53,375 INFO [finetune.py:976] (4/7) Epoch 10, batch 4250, loss[loss=0.1984, simple_loss=0.2586, pruned_loss=0.06916, over 4868.00 frames. ], tot_loss[loss=0.2001, simple_loss=0.2646, pruned_loss=0.06776, over 953130.13 frames. ], batch size: 31, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:20:57,635 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55804.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:21:06,141 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55809.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:21:38,838 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.929e+01 1.631e+02 1.908e+02 2.201e+02 4.056e+02, threshold=3.816e+02, percent-clipped=0.0 +2023-03-26 12:21:48,079 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=55848.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:21:48,566 INFO [finetune.py:976] (4/7) Epoch 10, batch 4300, loss[loss=0.1889, simple_loss=0.2438, pruned_loss=0.06697, over 4904.00 frames. ], tot_loss[loss=0.1964, simple_loss=0.2606, pruned_loss=0.06611, over 951566.03 frames. ], batch size: 37, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:21:49,301 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1986, 2.0655, 1.6904, 2.0582, 2.1049, 1.8271, 2.4638, 2.1541], + device='cuda:4'), covar=tensor([0.1225, 0.2139, 0.2933, 0.2641, 0.2430, 0.1576, 0.2902, 0.1708], + device='cuda:4'), in_proj_covar=tensor([0.0175, 0.0188, 0.0234, 0.0254, 0.0240, 0.0197, 0.0213, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:21:50,383 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=55852.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:22:10,868 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55867.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:22:36,507 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6652, 2.4864, 1.8572, 2.6411, 2.4959, 2.2041, 2.9703, 2.5549], + device='cuda:4'), covar=tensor([0.1262, 0.2437, 0.3623, 0.3040, 0.2906, 0.1723, 0.3734, 0.1947], + device='cuda:4'), in_proj_covar=tensor([0.0175, 0.0188, 0.0233, 0.0253, 0.0239, 0.0196, 0.0213, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:22:36,982 INFO [finetune.py:976] (4/7) Epoch 10, batch 4350, loss[loss=0.1537, simple_loss=0.2219, pruned_loss=0.04279, over 4798.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.258, pruned_loss=0.06541, over 952893.33 frames. ], batch size: 29, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:22:48,854 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55909.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:22:58,948 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=55915.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:23:23,255 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.622e+02 1.928e+02 2.410e+02 3.855e+02, threshold=3.856e+02, percent-clipped=1.0 +2023-03-26 12:23:37,719 INFO [finetune.py:976] (4/7) Epoch 10, batch 4400, loss[loss=0.1773, simple_loss=0.2643, pruned_loss=0.04519, over 4141.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2583, pruned_loss=0.06505, over 954788.87 frames. ], batch size: 65, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:23:43,341 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1800, 3.6619, 3.8360, 4.0725, 3.9049, 3.6260, 4.2460, 1.4253], + device='cuda:4'), covar=tensor([0.0778, 0.0904, 0.0792, 0.0923, 0.1247, 0.1732, 0.0820, 0.5187], + device='cuda:4'), in_proj_covar=tensor([0.0349, 0.0246, 0.0278, 0.0292, 0.0332, 0.0285, 0.0302, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:24:05,745 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5028, 1.4598, 1.5354, 0.8344, 1.5880, 1.5757, 1.5021, 1.4004], + device='cuda:4'), covar=tensor([0.0622, 0.0689, 0.0641, 0.0929, 0.0733, 0.0690, 0.0625, 0.1111], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0133, 0.0142, 0.0124, 0.0119, 0.0142, 0.0142, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:24:11,804 INFO [finetune.py:976] (4/7) Epoch 10, batch 4450, loss[loss=0.2024, simple_loss=0.2517, pruned_loss=0.07654, over 4229.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2615, pruned_loss=0.06595, over 953424.93 frames. ], batch size: 18, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:24:36,680 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.054e+02 1.562e+02 1.840e+02 2.258e+02 4.729e+02, threshold=3.681e+02, percent-clipped=2.0 +2023-03-26 12:24:40,383 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3693, 2.1115, 1.6198, 0.6801, 1.8102, 1.8230, 1.6602, 1.9205], + device='cuda:4'), covar=tensor([0.0792, 0.0902, 0.1676, 0.2085, 0.1476, 0.2246, 0.2228, 0.0943], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0201, 0.0203, 0.0188, 0.0218, 0.0209, 0.0225, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:24:46,910 INFO [finetune.py:976] (4/7) Epoch 10, batch 4500, loss[loss=0.2109, simple_loss=0.2899, pruned_loss=0.06595, over 4815.00 frames. ], tot_loss[loss=0.197, simple_loss=0.2623, pruned_loss=0.06586, over 954105.58 frames. ], batch size: 39, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:25:31,192 INFO [finetune.py:976] (4/7) Epoch 10, batch 4550, loss[loss=0.2444, simple_loss=0.2969, pruned_loss=0.09595, over 4813.00 frames. ], tot_loss[loss=0.1993, simple_loss=0.2646, pruned_loss=0.067, over 953442.00 frames. ], batch size: 30, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:25:34,313 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56104.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:25:53,265 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.089e+02 1.686e+02 1.941e+02 2.447e+02 3.858e+02, threshold=3.882e+02, percent-clipped=3.0 +2023-03-26 12:26:02,545 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56145.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:26:04,873 INFO [finetune.py:976] (4/7) Epoch 10, batch 4600, loss[loss=0.1999, simple_loss=0.2705, pruned_loss=0.06463, over 4795.00 frames. ], tot_loss[loss=0.199, simple_loss=0.2644, pruned_loss=0.06682, over 955656.74 frames. ], batch size: 51, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:26:40,469 INFO [finetune.py:976] (4/7) Epoch 10, batch 4650, loss[loss=0.1764, simple_loss=0.2424, pruned_loss=0.05523, over 4759.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.2618, pruned_loss=0.06592, over 955469.22 frames. ], batch size: 54, lr: 3.73e-03, grad_scale: 32.0 +2023-03-26 12:26:43,730 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56204.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:26:44,970 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56206.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:27:11,331 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2891, 1.4267, 1.4156, 1.4827, 1.5944, 2.9384, 1.3317, 1.5794], + device='cuda:4'), covar=tensor([0.1011, 0.1705, 0.1069, 0.0962, 0.1407, 0.0283, 0.1365, 0.1540], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0075, 0.0078, 0.0091, 0.0082, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 12:27:11,820 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.027e+02 1.561e+02 1.851e+02 2.355e+02 3.865e+02, threshold=3.702e+02, percent-clipped=0.0 +2023-03-26 12:27:23,137 INFO [finetune.py:976] (4/7) Epoch 10, batch 4700, loss[loss=0.163, simple_loss=0.2285, pruned_loss=0.04874, over 4830.00 frames. ], tot_loss[loss=0.1935, simple_loss=0.2584, pruned_loss=0.06433, over 956617.38 frames. ], batch size: 33, lr: 3.73e-03, grad_scale: 64.0 +2023-03-26 12:27:23,323 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 12:28:09,017 INFO [finetune.py:976] (4/7) Epoch 10, batch 4750, loss[loss=0.1904, simple_loss=0.2596, pruned_loss=0.06062, over 4857.00 frames. ], tot_loss[loss=0.1922, simple_loss=0.2571, pruned_loss=0.06366, over 955664.40 frames. ], batch size: 44, lr: 3.73e-03, grad_scale: 64.0 +2023-03-26 12:28:10,753 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5233, 1.3774, 2.0197, 2.8616, 1.8829, 2.2919, 1.0953, 2.3199], + device='cuda:4'), covar=tensor([0.1736, 0.1439, 0.1147, 0.0650, 0.0873, 0.1175, 0.1712, 0.0629], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0163, 0.0100, 0.0137, 0.0125, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 12:28:30,218 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.456e+02 1.800e+02 2.273e+02 6.888e+02, threshold=3.601e+02, percent-clipped=2.0 +2023-03-26 12:28:36,741 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-26 12:28:42,336 INFO [finetune.py:976] (4/7) Epoch 10, batch 4800, loss[loss=0.2328, simple_loss=0.2981, pruned_loss=0.08379, over 4807.00 frames. ], tot_loss[loss=0.195, simple_loss=0.2602, pruned_loss=0.06484, over 954953.71 frames. ], batch size: 41, lr: 3.73e-03, grad_scale: 64.0 +2023-03-26 12:29:14,947 INFO [finetune.py:976] (4/7) Epoch 10, batch 4850, loss[loss=0.1801, simple_loss=0.2508, pruned_loss=0.05465, over 4741.00 frames. ], tot_loss[loss=0.1976, simple_loss=0.2633, pruned_loss=0.0659, over 954632.64 frames. ], batch size: 27, lr: 3.73e-03, grad_scale: 64.0 +2023-03-26 12:29:19,700 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=56404.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:29:37,025 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.167e+02 1.738e+02 2.004e+02 2.451e+02 5.164e+02, threshold=4.009e+02, percent-clipped=2.0 +2023-03-26 12:29:48,226 INFO [finetune.py:976] (4/7) Epoch 10, batch 4900, loss[loss=0.206, simple_loss=0.2663, pruned_loss=0.07286, over 4693.00 frames. ], tot_loss[loss=0.2001, simple_loss=0.2655, pruned_loss=0.06738, over 953777.17 frames. ], batch size: 23, lr: 3.73e-03, grad_scale: 64.0 +2023-03-26 12:29:50,492 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=56452.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:30:15,911 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-26 12:30:26,421 INFO [finetune.py:976] (4/7) Epoch 10, batch 4950, loss[loss=0.2091, simple_loss=0.275, pruned_loss=0.07163, over 4912.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.2653, pruned_loss=0.06726, over 951126.55 frames. ], batch size: 46, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:30:32,461 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56501.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:30:33,126 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1359, 1.3719, 0.9604, 2.0623, 2.4889, 1.7780, 1.7606, 1.9638], + device='cuda:4'), covar=tensor([0.1432, 0.1973, 0.2086, 0.1182, 0.1728, 0.1951, 0.1394, 0.1882], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0092, 0.0121, 0.0095, 0.0099, 0.0091], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 12:30:34,893 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=56504.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:30:49,776 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56523.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 12:30:55,720 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.612e+02 1.965e+02 2.275e+02 4.231e+02, threshold=3.931e+02, percent-clipped=1.0 +2023-03-26 12:31:06,810 INFO [finetune.py:976] (4/7) Epoch 10, batch 5000, loss[loss=0.2089, simple_loss=0.2701, pruned_loss=0.07385, over 4881.00 frames. ], tot_loss[loss=0.1981, simple_loss=0.2631, pruned_loss=0.06657, over 950966.00 frames. ], batch size: 35, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:31:08,676 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=56552.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:31:29,657 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56584.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 12:31:39,223 INFO [finetune.py:976] (4/7) Epoch 10, batch 5050, loss[loss=0.1831, simple_loss=0.2356, pruned_loss=0.06527, over 4867.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2599, pruned_loss=0.06574, over 953002.12 frames. ], batch size: 31, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:32:00,862 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.32 vs. limit=5.0 +2023-03-26 12:32:04,809 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.243e+02 1.580e+02 1.790e+02 2.049e+02 5.062e+02, threshold=3.579e+02, percent-clipped=1.0 +2023-03-26 12:32:14,687 INFO [finetune.py:976] (4/7) Epoch 10, batch 5100, loss[loss=0.1706, simple_loss=0.2398, pruned_loss=0.05071, over 4824.00 frames. ], tot_loss[loss=0.1918, simple_loss=0.2558, pruned_loss=0.06388, over 954101.98 frames. ], batch size: 40, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:32:55,113 INFO [finetune.py:976] (4/7) Epoch 10, batch 5150, loss[loss=0.2447, simple_loss=0.3094, pruned_loss=0.09001, over 4043.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2586, pruned_loss=0.0656, over 953841.60 frames. ], batch size: 65, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:33:27,186 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.919e+01 1.632e+02 1.974e+02 2.331e+02 5.610e+02, threshold=3.948e+02, percent-clipped=3.0 +2023-03-26 12:33:36,883 INFO [finetune.py:976] (4/7) Epoch 10, batch 5200, loss[loss=0.1644, simple_loss=0.2394, pruned_loss=0.04469, over 4900.00 frames. ], tot_loss[loss=0.1981, simple_loss=0.2627, pruned_loss=0.06674, over 953594.15 frames. ], batch size: 35, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:34:10,225 INFO [finetune.py:976] (4/7) Epoch 10, batch 5250, loss[loss=0.1843, simple_loss=0.2505, pruned_loss=0.05911, over 3925.00 frames. ], tot_loss[loss=0.1979, simple_loss=0.2634, pruned_loss=0.06618, over 953188.69 frames. ], batch size: 17, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:34:11,648 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=56801.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:34:12,887 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56803.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:34:34,258 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.164e+02 1.706e+02 2.047e+02 2.503e+02 5.084e+02, threshold=4.093e+02, percent-clipped=2.0 +2023-03-26 12:34:43,965 INFO [finetune.py:976] (4/7) Epoch 10, batch 5300, loss[loss=0.2132, simple_loss=0.2765, pruned_loss=0.07494, over 4756.00 frames. ], tot_loss[loss=0.2004, simple_loss=0.2662, pruned_loss=0.06734, over 954391.44 frames. ], batch size: 59, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:34:44,027 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=56849.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:34:47,691 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3705, 1.3937, 1.8693, 1.9075, 1.5622, 3.3602, 1.2458, 1.5431], + device='cuda:4'), covar=tensor([0.1024, 0.1811, 0.1318, 0.0931, 0.1536, 0.0271, 0.1519, 0.1771], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0082, 0.0076, 0.0079, 0.0092, 0.0082, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 12:34:53,656 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56864.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:35:04,621 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56879.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 12:35:05,721 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6242, 3.2893, 3.1547, 1.4288, 3.4247, 2.5190, 0.8738, 2.1601], + device='cuda:4'), covar=tensor([0.2386, 0.1854, 0.1656, 0.3292, 0.1163, 0.1137, 0.4044, 0.1545], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0175, 0.0160, 0.0129, 0.0156, 0.0122, 0.0145, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 12:35:09,881 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1751, 1.3660, 1.4065, 0.6620, 1.3489, 1.5807, 1.6065, 1.2865], + device='cuda:4'), covar=tensor([0.0983, 0.0633, 0.0484, 0.0638, 0.0539, 0.0668, 0.0393, 0.0798], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0155, 0.0122, 0.0134, 0.0132, 0.0125, 0.0145, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.5951e-05, 1.1354e-04, 8.8318e-05, 9.6667e-05, 9.4578e-05, 9.1452e-05, + 1.0609e-04, 1.0787e-04], device='cuda:4') +2023-03-26 12:35:17,627 INFO [finetune.py:976] (4/7) Epoch 10, batch 5350, loss[loss=0.1816, simple_loss=0.2451, pruned_loss=0.05905, over 4893.00 frames. ], tot_loss[loss=0.1992, simple_loss=0.2655, pruned_loss=0.0665, over 956018.21 frames. ], batch size: 32, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:35:23,317 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56908.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:35:49,117 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.924e+01 1.584e+02 1.842e+02 2.192e+02 3.665e+02, threshold=3.684e+02, percent-clipped=0.0 +2023-03-26 12:36:01,255 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 12:36:01,777 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1487, 2.1756, 2.1986, 1.6062, 2.2122, 2.3496, 2.2837, 1.9451], + device='cuda:4'), covar=tensor([0.0570, 0.0585, 0.0714, 0.0857, 0.0632, 0.0656, 0.0566, 0.0957], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0134, 0.0144, 0.0125, 0.0120, 0.0144, 0.0144, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:36:02,276 INFO [finetune.py:976] (4/7) Epoch 10, batch 5400, loss[loss=0.1915, simple_loss=0.2525, pruned_loss=0.06525, over 4802.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2623, pruned_loss=0.06555, over 956877.90 frames. ], batch size: 45, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:36:15,032 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56969.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:36:15,039 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2882, 2.2155, 2.0425, 2.2355, 2.8093, 2.3414, 2.0625, 1.7834], + device='cuda:4'), covar=tensor([0.2082, 0.2039, 0.1847, 0.1712, 0.1801, 0.1161, 0.2268, 0.1971], + device='cuda:4'), in_proj_covar=tensor([0.0234, 0.0206, 0.0207, 0.0187, 0.0238, 0.0179, 0.0213, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:36:24,719 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-26 12:36:32,453 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.8367, 3.3308, 3.5149, 3.6805, 3.5888, 3.3805, 3.8706, 1.2583], + device='cuda:4'), covar=tensor([0.0792, 0.0813, 0.0838, 0.0943, 0.1253, 0.1420, 0.0819, 0.5058], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0247, 0.0279, 0.0292, 0.0335, 0.0288, 0.0303, 0.0297], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:36:32,486 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56993.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:36:33,666 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56995.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:36:35,991 INFO [finetune.py:976] (4/7) Epoch 10, batch 5450, loss[loss=0.1731, simple_loss=0.2492, pruned_loss=0.04848, over 4918.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.2599, pruned_loss=0.06521, over 958029.02 frames. ], batch size: 37, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:36:57,712 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.491e+02 1.807e+02 2.344e+02 4.842e+02, threshold=3.613e+02, percent-clipped=5.0 +2023-03-26 12:36:57,794 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7396, 3.5649, 3.5101, 1.6589, 3.6561, 2.7527, 0.8468, 2.4724], + device='cuda:4'), covar=tensor([0.2895, 0.2228, 0.1442, 0.3466, 0.1230, 0.1061, 0.4507, 0.1583], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0175, 0.0160, 0.0129, 0.0156, 0.0122, 0.0146, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 12:37:09,489 INFO [finetune.py:976] (4/7) Epoch 10, batch 5500, loss[loss=0.1851, simple_loss=0.244, pruned_loss=0.06311, over 4763.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.2573, pruned_loss=0.06462, over 955983.85 frames. ], batch size: 23, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:37:12,658 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57054.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:37:13,857 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57056.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:37:43,351 INFO [finetune.py:976] (4/7) Epoch 10, batch 5550, loss[loss=0.1474, simple_loss=0.224, pruned_loss=0.03539, over 4762.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.2577, pruned_loss=0.06455, over 954519.96 frames. ], batch size: 28, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:37:49,987 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 12:37:53,577 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-26 12:38:06,742 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.869e+01 1.587e+02 1.788e+02 2.090e+02 3.209e+02, threshold=3.576e+02, percent-clipped=0.0 +2023-03-26 12:38:25,600 INFO [finetune.py:976] (4/7) Epoch 10, batch 5600, loss[loss=0.2907, simple_loss=0.3436, pruned_loss=0.1189, over 4931.00 frames. ], tot_loss[loss=0.197, simple_loss=0.2623, pruned_loss=0.06583, over 953646.20 frames. ], batch size: 38, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:38:35,044 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57159.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:38:46,660 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57179.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 12:38:50,749 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57186.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:38:55,858 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5637, 2.5622, 2.4045, 1.9120, 2.4890, 2.8071, 2.7599, 2.1604], + device='cuda:4'), covar=tensor([0.0563, 0.0632, 0.0765, 0.0927, 0.0730, 0.0731, 0.0656, 0.1105], + device='cuda:4'), in_proj_covar=tensor([0.0136, 0.0134, 0.0145, 0.0126, 0.0121, 0.0145, 0.0145, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:38:57,578 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57197.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:38:58,653 INFO [finetune.py:976] (4/7) Epoch 10, batch 5650, loss[loss=0.2084, simple_loss=0.2503, pruned_loss=0.0832, over 4116.00 frames. ], tot_loss[loss=0.1998, simple_loss=0.2654, pruned_loss=0.06705, over 951850.02 frames. ], batch size: 18, lr: 3.72e-03, grad_scale: 64.0 +2023-03-26 12:39:15,288 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57227.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 12:39:19,289 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.551e+02 1.804e+02 2.162e+02 3.713e+02, threshold=3.608e+02, percent-clipped=1.0 +2023-03-26 12:39:25,353 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57244.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:39:27,096 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57247.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:39:28,225 INFO [finetune.py:976] (4/7) Epoch 10, batch 5700, loss[loss=0.1733, simple_loss=0.2254, pruned_loss=0.06063, over 4196.00 frames. ], tot_loss[loss=0.196, simple_loss=0.2605, pruned_loss=0.06575, over 932858.91 frames. ], batch size: 18, lr: 3.72e-03, grad_scale: 32.0 +2023-03-26 12:39:33,985 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57258.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 12:39:37,771 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57264.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:40:00,692 INFO [finetune.py:976] (4/7) Epoch 11, batch 0, loss[loss=0.2043, simple_loss=0.2703, pruned_loss=0.06912, over 4824.00 frames. ], tot_loss[loss=0.2043, simple_loss=0.2703, pruned_loss=0.06912, over 4824.00 frames. ], batch size: 33, lr: 3.72e-03, grad_scale: 16.0 +2023-03-26 12:40:00,692 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 12:40:16,056 INFO [finetune.py:1010] (4/7) Epoch 11, validation: loss=0.1597, simple_loss=0.2306, pruned_loss=0.04438, over 2265189.00 frames. +2023-03-26 12:40:16,056 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 12:40:37,192 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57305.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:40:59,552 INFO [finetune.py:976] (4/7) Epoch 11, batch 50, loss[loss=0.1716, simple_loss=0.233, pruned_loss=0.05511, over 4762.00 frames. ], tot_loss[loss=0.2016, simple_loss=0.2655, pruned_loss=0.06885, over 215947.55 frames. ], batch size: 26, lr: 3.72e-03, grad_scale: 16.0 +2023-03-26 12:41:10,025 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.036e+02 1.580e+02 1.868e+02 2.535e+02 4.204e+02, threshold=3.735e+02, percent-clipped=3.0 +2023-03-26 12:41:18,571 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57349.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:41:19,795 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57351.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:41:38,107 INFO [finetune.py:976] (4/7) Epoch 11, batch 100, loss[loss=0.1813, simple_loss=0.2384, pruned_loss=0.06206, over 4746.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2566, pruned_loss=0.06406, over 380182.62 frames. ], batch size: 54, lr: 3.72e-03, grad_scale: 16.0 +2023-03-26 12:41:50,246 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8590, 1.4206, 2.2527, 1.4368, 1.8802, 1.9753, 1.4004, 2.1719], + device='cuda:4'), covar=tensor([0.1384, 0.2354, 0.1223, 0.1935, 0.1155, 0.1661, 0.2976, 0.1071], + device='cuda:4'), in_proj_covar=tensor([0.0199, 0.0206, 0.0194, 0.0191, 0.0178, 0.0216, 0.0219, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:41:51,427 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57398.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:42:11,570 INFO [finetune.py:976] (4/7) Epoch 11, batch 150, loss[loss=0.1875, simple_loss=0.2471, pruned_loss=0.0639, over 4861.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2539, pruned_loss=0.06438, over 507997.96 frames. ], batch size: 49, lr: 3.72e-03, grad_scale: 16.0 +2023-03-26 12:42:13,624 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.29 vs. limit=5.0 +2023-03-26 12:42:16,968 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.728e+02 2.070e+02 2.489e+02 4.280e+02, threshold=4.140e+02, percent-clipped=3.0 +2023-03-26 12:42:31,654 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57459.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:42:31,682 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57459.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:42:33,494 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4860, 1.0898, 0.7712, 1.3966, 1.9790, 0.7932, 1.2658, 1.4833], + device='cuda:4'), covar=tensor([0.1502, 0.2184, 0.1842, 0.1232, 0.1921, 0.2026, 0.1563, 0.1931], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0092, 0.0120, 0.0095, 0.0099, 0.0091], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 12:42:33,514 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57462.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:42:40,895 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-26 12:42:44,009 INFO [finetune.py:976] (4/7) Epoch 11, batch 200, loss[loss=0.1649, simple_loss=0.2355, pruned_loss=0.04714, over 4762.00 frames. ], tot_loss[loss=0.1898, simple_loss=0.2526, pruned_loss=0.06352, over 606277.95 frames. ], batch size: 28, lr: 3.72e-03, grad_scale: 16.0 +2023-03-26 12:42:56,401 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57495.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:43:03,708 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57507.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:43:14,488 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57523.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:43:17,309 INFO [finetune.py:976] (4/7) Epoch 11, batch 250, loss[loss=0.1661, simple_loss=0.2231, pruned_loss=0.05456, over 4776.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2563, pruned_loss=0.06482, over 681866.86 frames. ], batch size: 26, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:43:22,637 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.287e+02 1.584e+02 1.966e+02 2.356e+02 4.681e+02, threshold=3.932e+02, percent-clipped=1.0 +2023-03-26 12:43:27,976 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57542.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:43:34,243 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 12:43:43,774 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57553.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 12:43:44,475 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-03-26 12:43:45,618 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57556.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:43:55,244 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57564.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:44:04,604 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-03-26 12:44:08,380 INFO [finetune.py:976] (4/7) Epoch 11, batch 300, loss[loss=0.1854, simple_loss=0.2611, pruned_loss=0.05489, over 4924.00 frames. ], tot_loss[loss=0.1947, simple_loss=0.2588, pruned_loss=0.06534, over 742723.70 frames. ], batch size: 38, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:44:24,432 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57600.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:44:31,737 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57612.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:44:40,866 INFO [finetune.py:976] (4/7) Epoch 11, batch 350, loss[loss=0.1969, simple_loss=0.2688, pruned_loss=0.0625, over 4839.00 frames. ], tot_loss[loss=0.1954, simple_loss=0.2606, pruned_loss=0.06512, over 790946.51 frames. ], batch size: 44, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:44:46,738 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.573e+02 1.819e+02 2.403e+02 4.156e+02, threshold=3.639e+02, percent-clipped=1.0 +2023-03-26 12:44:56,809 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57649.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:44:58,451 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57651.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:45:14,039 INFO [finetune.py:976] (4/7) Epoch 11, batch 400, loss[loss=0.2121, simple_loss=0.2734, pruned_loss=0.07536, over 4731.00 frames. ], tot_loss[loss=0.1965, simple_loss=0.2623, pruned_loss=0.06531, over 827041.29 frames. ], batch size: 27, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:45:30,903 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57697.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:45:32,136 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57699.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:45:35,177 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 12:45:49,608 INFO [finetune.py:976] (4/7) Epoch 11, batch 450, loss[loss=0.2095, simple_loss=0.2642, pruned_loss=0.07745, over 4864.00 frames. ], tot_loss[loss=0.1952, simple_loss=0.2608, pruned_loss=0.06479, over 854639.40 frames. ], batch size: 31, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:45:53,878 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57733.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:45:55,475 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.602e+02 1.902e+02 2.220e+02 3.989e+02, threshold=3.804e+02, percent-clipped=2.0 +2023-03-26 12:46:15,574 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57754.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:46:32,835 INFO [finetune.py:976] (4/7) Epoch 11, batch 500, loss[loss=0.2062, simple_loss=0.2528, pruned_loss=0.07985, over 4827.00 frames. ], tot_loss[loss=0.1945, simple_loss=0.2596, pruned_loss=0.06466, over 877412.23 frames. ], batch size: 30, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:46:39,379 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6814, 1.5804, 2.3513, 3.6063, 2.3546, 2.4632, 1.0770, 2.8152], + device='cuda:4'), covar=tensor([0.1796, 0.1459, 0.1246, 0.0528, 0.0791, 0.1335, 0.1843, 0.0550], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0116, 0.0134, 0.0162, 0.0101, 0.0137, 0.0125, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 12:46:45,704 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57794.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:47:01,357 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57818.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:47:06,701 INFO [finetune.py:976] (4/7) Epoch 11, batch 550, loss[loss=0.2104, simple_loss=0.2751, pruned_loss=0.07279, over 4818.00 frames. ], tot_loss[loss=0.1929, simple_loss=0.2575, pruned_loss=0.06417, over 895604.24 frames. ], batch size: 40, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:47:11,532 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.185e+02 1.635e+02 1.936e+02 2.160e+02 3.511e+02, threshold=3.871e+02, percent-clipped=0.0 +2023-03-26 12:47:16,790 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57842.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:47:20,236 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.2441, 1.4241, 1.4930, 0.7771, 1.2708, 1.6195, 1.6759, 1.3260], + device='cuda:4'), covar=tensor([0.0835, 0.0491, 0.0343, 0.0485, 0.0378, 0.0446, 0.0245, 0.0569], + device='cuda:4'), in_proj_covar=tensor([0.0127, 0.0153, 0.0120, 0.0132, 0.0130, 0.0123, 0.0142, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.4172e-05, 1.1240e-04, 8.6302e-05, 9.5122e-05, 9.3206e-05, 8.9835e-05, + 1.0412e-04, 1.0622e-04], device='cuda:4') +2023-03-26 12:47:23,739 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57851.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:47:24,986 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57853.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 12:47:40,102 INFO [finetune.py:976] (4/7) Epoch 11, batch 600, loss[loss=0.175, simple_loss=0.2518, pruned_loss=0.04915, over 4833.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2588, pruned_loss=0.0649, over 907514.55 frames. ], batch size: 30, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:47:48,477 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57890.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:47:56,177 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57900.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:47:56,760 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57901.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:48:13,625 INFO [finetune.py:976] (4/7) Epoch 11, batch 650, loss[loss=0.1938, simple_loss=0.2611, pruned_loss=0.06321, over 4750.00 frames. ], tot_loss[loss=0.1971, simple_loss=0.2621, pruned_loss=0.06606, over 918255.46 frames. ], batch size: 27, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:48:18,498 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.178e+02 1.568e+02 1.897e+02 2.360e+02 4.682e+02, threshold=3.793e+02, percent-clipped=3.0 +2023-03-26 12:48:27,448 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=57948.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:48:30,323 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7283, 1.5968, 1.5071, 1.7758, 2.2466, 1.8658, 1.4919, 1.3987], + device='cuda:4'), covar=tensor([0.2172, 0.2142, 0.1983, 0.1655, 0.1722, 0.1215, 0.2448, 0.1967], + device='cuda:4'), in_proj_covar=tensor([0.0236, 0.0208, 0.0208, 0.0189, 0.0242, 0.0181, 0.0213, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:48:48,746 INFO [finetune.py:976] (4/7) Epoch 11, batch 700, loss[loss=0.1767, simple_loss=0.2549, pruned_loss=0.04926, over 4817.00 frames. ], tot_loss[loss=0.198, simple_loss=0.2638, pruned_loss=0.06605, over 926989.43 frames. ], batch size: 39, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:49:07,415 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.0099, 4.1441, 3.9532, 2.0620, 4.2036, 2.9718, 0.8802, 2.8932], + device='cuda:4'), covar=tensor([0.2044, 0.1941, 0.1432, 0.3024, 0.1007, 0.1036, 0.4434, 0.1289], + device='cuda:4'), in_proj_covar=tensor([0.0149, 0.0171, 0.0156, 0.0126, 0.0154, 0.0120, 0.0143, 0.0119], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 12:49:44,565 INFO [finetune.py:976] (4/7) Epoch 11, batch 750, loss[loss=0.1493, simple_loss=0.2067, pruned_loss=0.046, over 4047.00 frames. ], tot_loss[loss=0.2008, simple_loss=0.2661, pruned_loss=0.06778, over 933877.22 frames. ], batch size: 17, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:49:49,411 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.116e+01 1.579e+02 1.894e+02 2.321e+02 4.436e+02, threshold=3.789e+02, percent-clipped=3.0 +2023-03-26 12:49:52,695 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.51 vs. limit=5.0 +2023-03-26 12:50:02,178 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58054.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:50:18,110 INFO [finetune.py:976] (4/7) Epoch 11, batch 800, loss[loss=0.2077, simple_loss=0.2824, pruned_loss=0.06643, over 4823.00 frames. ], tot_loss[loss=0.1981, simple_loss=0.2638, pruned_loss=0.0662, over 936808.70 frames. ], batch size: 41, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:50:21,514 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 12:50:25,465 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=58089.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:50:33,870 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58102.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:50:45,548 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58118.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:50:51,411 INFO [finetune.py:976] (4/7) Epoch 11, batch 850, loss[loss=0.1943, simple_loss=0.2552, pruned_loss=0.06671, over 4832.00 frames. ], tot_loss[loss=0.1983, simple_loss=0.263, pruned_loss=0.06681, over 939036.76 frames. ], batch size: 47, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:50:56,225 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.394e+01 1.505e+02 1.749e+02 2.082e+02 4.545e+02, threshold=3.498e+02, percent-clipped=2.0 +2023-03-26 12:50:59,952 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58141.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:51:01,784 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58144.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:51:05,917 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58151.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:51:23,203 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58166.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:51:35,956 INFO [finetune.py:976] (4/7) Epoch 11, batch 900, loss[loss=0.1767, simple_loss=0.2393, pruned_loss=0.05703, over 4717.00 frames. ], tot_loss[loss=0.1959, simple_loss=0.2606, pruned_loss=0.06564, over 944373.53 frames. ], batch size: 23, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:51:41,968 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.69 vs. limit=5.0 +2023-03-26 12:51:56,292 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 12:51:57,411 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58199.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:51:59,456 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58202.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:52:01,296 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58205.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:52:04,283 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=5.04 vs. limit=5.0 +2023-03-26 12:52:12,171 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5253, 1.4271, 1.4057, 1.4338, 1.0508, 2.8125, 1.1321, 1.5653], + device='cuda:4'), covar=tensor([0.3468, 0.2398, 0.2290, 0.2459, 0.1844, 0.0273, 0.2996, 0.1374], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0116, 0.0121, 0.0123, 0.0115, 0.0098, 0.0099, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 12:52:17,040 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5707, 2.4083, 1.9462, 2.6152, 2.5251, 2.1406, 2.9870, 2.5397], + device='cuda:4'), covar=tensor([0.1404, 0.2841, 0.3404, 0.2952, 0.2820, 0.1702, 0.3047, 0.2058], + device='cuda:4'), in_proj_covar=tensor([0.0174, 0.0187, 0.0232, 0.0254, 0.0238, 0.0196, 0.0212, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:52:17,500 INFO [finetune.py:976] (4/7) Epoch 11, batch 950, loss[loss=0.1502, simple_loss=0.2267, pruned_loss=0.03684, over 4875.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2578, pruned_loss=0.06407, over 948723.94 frames. ], batch size: 31, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:52:22,883 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.516e+02 1.975e+02 2.310e+02 4.008e+02, threshold=3.950e+02, percent-clipped=1.0 +2023-03-26 12:52:51,452 INFO [finetune.py:976] (4/7) Epoch 11, batch 1000, loss[loss=0.2334, simple_loss=0.2976, pruned_loss=0.08459, over 4751.00 frames. ], tot_loss[loss=0.1945, simple_loss=0.2599, pruned_loss=0.06458, over 950809.19 frames. ], batch size: 59, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:53:46,405 INFO [finetune.py:976] (4/7) Epoch 11, batch 1050, loss[loss=0.1833, simple_loss=0.2585, pruned_loss=0.05408, over 4896.00 frames. ], tot_loss[loss=0.1964, simple_loss=0.2625, pruned_loss=0.06517, over 953176.13 frames. ], batch size: 36, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:53:51,319 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.197e+02 1.617e+02 2.003e+02 2.375e+02 3.670e+02, threshold=4.006e+02, percent-clipped=0.0 +2023-03-26 12:54:02,543 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.18 vs. limit=5.0 +2023-03-26 12:54:11,827 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.2560, 1.3351, 1.3190, 0.7462, 1.1868, 1.4952, 1.5273, 1.2354], + device='cuda:4'), covar=tensor([0.0828, 0.0473, 0.0497, 0.0416, 0.0497, 0.0543, 0.0261, 0.0589], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0152, 0.0119, 0.0131, 0.0129, 0.0122, 0.0141, 0.0144], + device='cuda:4'), out_proj_covar=tensor([9.3441e-05, 1.1144e-04, 8.5891e-05, 9.4605e-05, 9.1906e-05, 8.9067e-05, + 1.0343e-04, 1.0546e-04], device='cuda:4') +2023-03-26 12:54:42,548 INFO [finetune.py:976] (4/7) Epoch 11, batch 1100, loss[loss=0.2057, simple_loss=0.2787, pruned_loss=0.06635, over 4890.00 frames. ], tot_loss[loss=0.1975, simple_loss=0.2635, pruned_loss=0.06582, over 952138.90 frames. ], batch size: 43, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:54:55,677 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58389.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:55:15,875 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 12:55:23,532 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8542, 1.7205, 1.4751, 1.9437, 2.4961, 2.0856, 1.8143, 1.4904], + device='cuda:4'), covar=tensor([0.2081, 0.2005, 0.1877, 0.1468, 0.1590, 0.1020, 0.2164, 0.1829], + device='cuda:4'), in_proj_covar=tensor([0.0236, 0.0207, 0.0207, 0.0189, 0.0241, 0.0181, 0.0213, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:55:35,370 INFO [finetune.py:976] (4/7) Epoch 11, batch 1150, loss[loss=0.1864, simple_loss=0.2586, pruned_loss=0.05714, over 4807.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2647, pruned_loss=0.06608, over 954565.11 frames. ], batch size: 40, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:55:40,650 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.672e+02 1.870e+02 2.321e+02 4.403e+02, threshold=3.740e+02, percent-clipped=1.0 +2023-03-26 12:55:41,942 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58437.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:55:42,013 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3697, 1.5026, 1.6515, 0.9613, 1.5691, 1.6766, 1.8256, 1.3603], + device='cuda:4'), covar=tensor([0.0814, 0.0530, 0.0375, 0.0416, 0.0364, 0.0539, 0.0268, 0.0620], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0152, 0.0119, 0.0131, 0.0129, 0.0122, 0.0141, 0.0144], + device='cuda:4'), out_proj_covar=tensor([9.3444e-05, 1.1168e-04, 8.5887e-05, 9.4596e-05, 9.1801e-05, 8.9122e-05, + 1.0344e-04, 1.0542e-04], device='cuda:4') +2023-03-26 12:55:54,579 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5100, 1.4141, 1.3511, 1.4142, 1.0808, 2.9201, 1.1746, 1.5972], + device='cuda:4'), covar=tensor([0.4253, 0.3111, 0.2522, 0.3053, 0.1963, 0.0318, 0.2750, 0.1324], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0115, 0.0119, 0.0122, 0.0115, 0.0097, 0.0098, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 12:56:08,438 INFO [finetune.py:976] (4/7) Epoch 11, batch 1200, loss[loss=0.1841, simple_loss=0.2496, pruned_loss=0.05935, over 4812.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2625, pruned_loss=0.06548, over 953594.18 frames. ], batch size: 38, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:56:08,574 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5595, 1.4366, 1.2858, 1.6064, 1.6256, 1.6053, 0.9311, 1.3288], + device='cuda:4'), covar=tensor([0.2285, 0.2213, 0.1958, 0.1604, 0.1781, 0.1256, 0.2608, 0.1938], + device='cuda:4'), in_proj_covar=tensor([0.0236, 0.0208, 0.0207, 0.0189, 0.0242, 0.0181, 0.0213, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:56:15,633 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2633, 2.0913, 1.7765, 2.1275, 2.1794, 1.9011, 2.5158, 2.2018], + device='cuda:4'), covar=tensor([0.1356, 0.2359, 0.3096, 0.2889, 0.2888, 0.1722, 0.3332, 0.1957], + device='cuda:4'), in_proj_covar=tensor([0.0174, 0.0187, 0.0231, 0.0253, 0.0238, 0.0195, 0.0211, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:56:21,458 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=58497.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:56:23,266 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=58500.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:56:37,013 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1740, 1.8300, 1.3178, 0.5383, 1.6766, 1.7809, 1.3876, 1.6342], + device='cuda:4'), covar=tensor([0.0633, 0.0856, 0.1259, 0.1720, 0.1142, 0.1815, 0.2132, 0.0816], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0202, 0.0203, 0.0189, 0.0216, 0.0209, 0.0224, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:56:40,496 INFO [finetune.py:976] (4/7) Epoch 11, batch 1250, loss[loss=0.2297, simple_loss=0.2869, pruned_loss=0.08622, over 4121.00 frames. ], tot_loss[loss=0.1956, simple_loss=0.2607, pruned_loss=0.0653, over 954021.77 frames. ], batch size: 65, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:56:46,790 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.332e+01 1.581e+02 1.822e+02 2.261e+02 4.369e+02, threshold=3.644e+02, percent-clipped=3.0 +2023-03-26 12:57:08,445 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0378, 1.4900, 2.0029, 1.8961, 1.7614, 1.6717, 1.8358, 1.8312], + device='cuda:4'), covar=tensor([0.4043, 0.4683, 0.3949, 0.4306, 0.5511, 0.4439, 0.5436, 0.3683], + device='cuda:4'), in_proj_covar=tensor([0.0236, 0.0238, 0.0252, 0.0256, 0.0252, 0.0228, 0.0273, 0.0229], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 12:57:15,450 INFO [finetune.py:976] (4/7) Epoch 11, batch 1300, loss[loss=0.2254, simple_loss=0.2848, pruned_loss=0.08295, over 4823.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.2573, pruned_loss=0.06388, over 954523.08 frames. ], batch size: 40, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:57:30,716 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6917, 1.5018, 2.2138, 3.3237, 2.2168, 2.3639, 0.9274, 2.5478], + device='cuda:4'), covar=tensor([0.1738, 0.1443, 0.1263, 0.0603, 0.0877, 0.1630, 0.1931, 0.0654], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0163, 0.0100, 0.0137, 0.0124, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 12:57:48,895 INFO [finetune.py:976] (4/7) Epoch 11, batch 1350, loss[loss=0.2081, simple_loss=0.2741, pruned_loss=0.07108, over 4823.00 frames. ], tot_loss[loss=0.1921, simple_loss=0.2571, pruned_loss=0.06357, over 956300.48 frames. ], batch size: 39, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:57:54,734 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.484e+01 1.581e+02 1.914e+02 2.266e+02 4.857e+02, threshold=3.829e+02, percent-clipped=2.0 +2023-03-26 12:58:20,829 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-26 12:58:23,950 INFO [finetune.py:976] (4/7) Epoch 11, batch 1400, loss[loss=0.2293, simple_loss=0.303, pruned_loss=0.07783, over 4905.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2604, pruned_loss=0.06458, over 957074.00 frames. ], batch size: 35, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:58:27,003 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58681.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:58:56,044 INFO [finetune.py:976] (4/7) Epoch 11, batch 1450, loss[loss=0.2878, simple_loss=0.3337, pruned_loss=0.1209, over 4907.00 frames. ], tot_loss[loss=0.1977, simple_loss=0.2634, pruned_loss=0.06596, over 957368.32 frames. ], batch size: 37, lr: 3.71e-03, grad_scale: 16.0 +2023-03-26 12:59:00,205 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1284, 2.0819, 2.0179, 2.2150, 1.7321, 3.9569, 1.9324, 2.5246], + device='cuda:4'), covar=tensor([0.2763, 0.2164, 0.1790, 0.1958, 0.1432, 0.0157, 0.2293, 0.1036], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0116, 0.0120, 0.0123, 0.0115, 0.0098, 0.0099, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 12:59:01,960 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.669e+02 2.009e+02 2.318e+02 4.324e+02, threshold=4.017e+02, percent-clipped=1.0 +2023-03-26 12:59:07,256 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58742.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:59:35,109 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58775.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 12:59:36,228 INFO [finetune.py:976] (4/7) Epoch 11, batch 1500, loss[loss=0.1959, simple_loss=0.2607, pruned_loss=0.06553, over 4179.00 frames. ], tot_loss[loss=0.1976, simple_loss=0.2637, pruned_loss=0.06576, over 956023.16 frames. ], batch size: 65, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 12:59:58,338 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58797.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:00:04,324 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58800.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:00:17,087 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8208, 3.6626, 3.5426, 1.7998, 3.8512, 2.7381, 0.9185, 2.5035], + device='cuda:4'), covar=tensor([0.2061, 0.1633, 0.1386, 0.3016, 0.0938, 0.0986, 0.4016, 0.1287], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0174, 0.0159, 0.0129, 0.0156, 0.0121, 0.0146, 0.0121], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 13:00:33,695 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1425, 2.2300, 2.2189, 1.5733, 2.2443, 2.3793, 2.2862, 1.9098], + device='cuda:4'), covar=tensor([0.0606, 0.0602, 0.0630, 0.0921, 0.0665, 0.0641, 0.0626, 0.0990], + device='cuda:4'), in_proj_covar=tensor([0.0135, 0.0134, 0.0143, 0.0125, 0.0121, 0.0144, 0.0144, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:00:34,194 INFO [finetune.py:976] (4/7) Epoch 11, batch 1550, loss[loss=0.1692, simple_loss=0.2404, pruned_loss=0.04897, over 4806.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2611, pruned_loss=0.06387, over 955889.24 frames. ], batch size: 41, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 13:00:39,949 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.185e+02 1.569e+02 1.959e+02 2.197e+02 4.059e+02, threshold=3.918e+02, percent-clipped=1.0 +2023-03-26 13:00:41,225 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58836.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:00:47,625 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58845.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:00:49,961 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=58848.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:01:03,783 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5658, 1.7127, 1.3736, 1.6353, 2.0878, 1.8231, 1.6113, 1.4082], + device='cuda:4'), covar=tensor([0.0351, 0.0287, 0.0647, 0.0311, 0.0162, 0.0525, 0.0367, 0.0446], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0109, 0.0140, 0.0115, 0.0101, 0.0104, 0.0093, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.0917e-05, 8.4964e-05, 1.1157e-04, 8.9720e-05, 7.9245e-05, 7.7208e-05, + 7.0476e-05, 8.3321e-05], device='cuda:4') +2023-03-26 13:01:07,921 INFO [finetune.py:976] (4/7) Epoch 11, batch 1600, loss[loss=0.172, simple_loss=0.241, pruned_loss=0.05148, over 4809.00 frames. ], tot_loss[loss=0.1929, simple_loss=0.2591, pruned_loss=0.06333, over 955340.91 frames. ], batch size: 29, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 13:01:28,380 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58906.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:01:50,385 INFO [finetune.py:976] (4/7) Epoch 11, batch 1650, loss[loss=0.1462, simple_loss=0.2105, pruned_loss=0.04097, over 4803.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2563, pruned_loss=0.06269, over 957002.19 frames. ], batch size: 25, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 13:01:51,202 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.67 vs. limit=2.0 +2023-03-26 13:01:55,257 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.149e+02 1.664e+02 1.923e+02 2.390e+02 4.121e+02, threshold=3.846e+02, percent-clipped=1.0 +2023-03-26 13:02:06,924 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0936, 1.7514, 2.4359, 1.6022, 2.2915, 2.3051, 1.7558, 2.4007], + device='cuda:4'), covar=tensor([0.1323, 0.2038, 0.1332, 0.2047, 0.0891, 0.1615, 0.2581, 0.0931], + device='cuda:4'), in_proj_covar=tensor([0.0198, 0.0205, 0.0193, 0.0191, 0.0179, 0.0215, 0.0218, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:02:16,528 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0616, 1.4910, 2.0497, 1.9871, 1.7618, 1.7112, 1.8361, 1.8364], + device='cuda:4'), covar=tensor([0.3971, 0.4751, 0.3780, 0.4091, 0.5671, 0.4100, 0.5008, 0.3738], + device='cuda:4'), in_proj_covar=tensor([0.0235, 0.0236, 0.0251, 0.0254, 0.0252, 0.0227, 0.0271, 0.0228], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:02:18,255 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58967.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:02:24,167 INFO [finetune.py:976] (4/7) Epoch 11, batch 1700, loss[loss=0.2439, simple_loss=0.2955, pruned_loss=0.09611, over 4821.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2545, pruned_loss=0.06215, over 957095.82 frames. ], batch size: 40, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 13:02:51,454 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.3727, 4.6625, 4.9184, 5.2055, 5.0482, 4.7732, 5.5051, 1.7008], + device='cuda:4'), covar=tensor([0.0681, 0.0778, 0.0734, 0.0866, 0.1193, 0.1507, 0.0459, 0.5728], + device='cuda:4'), in_proj_covar=tensor([0.0351, 0.0245, 0.0276, 0.0293, 0.0334, 0.0285, 0.0302, 0.0297], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:02:51,466 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2205, 1.9680, 2.5985, 4.2332, 2.9419, 2.8369, 0.8930, 3.4970], + device='cuda:4'), covar=tensor([0.1675, 0.1477, 0.1387, 0.0615, 0.0754, 0.1425, 0.2055, 0.0479], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0163, 0.0100, 0.0137, 0.0125, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 13:02:57,899 INFO [finetune.py:976] (4/7) Epoch 11, batch 1750, loss[loss=0.2421, simple_loss=0.3026, pruned_loss=0.09082, over 4858.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2562, pruned_loss=0.06316, over 954884.82 frames. ], batch size: 44, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 13:03:02,756 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.140e+02 1.620e+02 1.895e+02 2.249e+02 5.052e+02, threshold=3.790e+02, percent-clipped=2.0 +2023-03-26 13:03:04,076 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59037.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:03:19,364 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7403, 1.7309, 1.7784, 1.1413, 1.8121, 1.8202, 1.8165, 1.5244], + device='cuda:4'), covar=tensor([0.0598, 0.0680, 0.0715, 0.0970, 0.0673, 0.0745, 0.0613, 0.1094], + device='cuda:4'), in_proj_covar=tensor([0.0136, 0.0135, 0.0144, 0.0126, 0.0122, 0.0145, 0.0146, 0.0164], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:03:33,672 INFO [finetune.py:976] (4/7) Epoch 11, batch 1800, loss[loss=0.223, simple_loss=0.2797, pruned_loss=0.08318, over 4186.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2609, pruned_loss=0.0644, over 954391.09 frames. ], batch size: 65, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 13:03:56,881 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 13:04:19,595 INFO [finetune.py:976] (4/7) Epoch 11, batch 1850, loss[loss=0.1722, simple_loss=0.2469, pruned_loss=0.04875, over 4791.00 frames. ], tot_loss[loss=0.1972, simple_loss=0.2635, pruned_loss=0.06549, over 955994.34 frames. ], batch size: 51, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 13:04:22,096 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59131.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:04:24,431 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.037e+02 1.668e+02 2.065e+02 2.636e+02 4.490e+02, threshold=4.130e+02, percent-clipped=5.0 +2023-03-26 13:04:32,969 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8604, 1.3016, 1.8807, 1.7829, 1.5313, 1.5058, 1.7126, 1.6463], + device='cuda:4'), covar=tensor([0.3768, 0.4333, 0.3362, 0.3896, 0.5199, 0.3977, 0.4849, 0.3430], + device='cuda:4'), in_proj_covar=tensor([0.0235, 0.0237, 0.0252, 0.0256, 0.0252, 0.0228, 0.0272, 0.0229], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:04:33,300 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 13:04:43,984 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59164.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:04:57,367 INFO [finetune.py:976] (4/7) Epoch 11, batch 1900, loss[loss=0.2489, simple_loss=0.2855, pruned_loss=0.1061, over 4138.00 frames. ], tot_loss[loss=0.1988, simple_loss=0.2655, pruned_loss=0.06603, over 957437.96 frames. ], batch size: 65, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 13:05:46,825 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59225.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:05:47,936 INFO [finetune.py:976] (4/7) Epoch 11, batch 1950, loss[loss=0.1777, simple_loss=0.2452, pruned_loss=0.05516, over 4886.00 frames. ], tot_loss[loss=0.1963, simple_loss=0.263, pruned_loss=0.06485, over 957216.20 frames. ], batch size: 32, lr: 3.70e-03, grad_scale: 16.0 +2023-03-26 13:05:59,308 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.090e+02 1.570e+02 1.817e+02 2.294e+02 4.310e+02, threshold=3.633e+02, percent-clipped=1.0 +2023-03-26 13:06:29,864 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59262.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:06:42,211 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.20 vs. limit=5.0 +2023-03-26 13:06:51,936 INFO [finetune.py:976] (4/7) Epoch 11, batch 2000, loss[loss=0.1925, simple_loss=0.2482, pruned_loss=0.06843, over 4827.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2605, pruned_loss=0.06392, over 957282.00 frames. ], batch size: 33, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:06:52,040 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3324, 1.4827, 1.3675, 1.6214, 1.4905, 3.0045, 1.3503, 1.5526], + device='cuda:4'), covar=tensor([0.1012, 0.1768, 0.1130, 0.0967, 0.1612, 0.0288, 0.1485, 0.1743], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0080, 0.0074, 0.0077, 0.0091, 0.0081, 0.0084, 0.0078], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 13:07:31,105 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9900, 1.3924, 1.9944, 1.9474, 1.6800, 1.6413, 1.7936, 1.7531], + device='cuda:4'), covar=tensor([0.4124, 0.4790, 0.3969, 0.4069, 0.5615, 0.4236, 0.5609, 0.3835], + device='cuda:4'), in_proj_covar=tensor([0.0237, 0.0239, 0.0253, 0.0258, 0.0254, 0.0229, 0.0274, 0.0231], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:07:37,498 INFO [finetune.py:976] (4/7) Epoch 11, batch 2050, loss[loss=0.1752, simple_loss=0.2352, pruned_loss=0.05753, over 4145.00 frames. ], tot_loss[loss=0.1919, simple_loss=0.2572, pruned_loss=0.06332, over 955775.81 frames. ], batch size: 65, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:07:42,280 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.122e+01 1.513e+02 1.843e+02 2.174e+02 3.611e+02, threshold=3.686e+02, percent-clipped=0.0 +2023-03-26 13:07:50,361 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=59337.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:08:17,291 INFO [finetune.py:976] (4/7) Epoch 11, batch 2100, loss[loss=0.2697, simple_loss=0.3057, pruned_loss=0.1169, over 4856.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.2571, pruned_loss=0.06393, over 955469.75 frames. ], batch size: 49, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:08:27,934 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=59385.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:08:32,799 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1910, 2.0340, 1.7189, 2.2049, 2.0398, 1.9891, 1.9098, 2.9437], + device='cuda:4'), covar=tensor([0.4573, 0.6174, 0.4112, 0.5351, 0.5320, 0.2809, 0.5331, 0.1855], + device='cuda:4'), in_proj_covar=tensor([0.0287, 0.0261, 0.0225, 0.0280, 0.0243, 0.0210, 0.0246, 0.0216], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:08:45,140 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2281, 1.1429, 1.1074, 1.2545, 1.5782, 1.4116, 1.2749, 1.1274], + device='cuda:4'), covar=tensor([0.0367, 0.0331, 0.0697, 0.0306, 0.0220, 0.0488, 0.0355, 0.0424], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0109, 0.0139, 0.0114, 0.0101, 0.0103, 0.0092, 0.0107], + device='cuda:4'), out_proj_covar=tensor([7.0362e-05, 8.4643e-05, 1.1079e-04, 8.9077e-05, 7.9076e-05, 7.6633e-05, + 7.0121e-05, 8.2697e-05], device='cuda:4') +2023-03-26 13:09:08,920 INFO [finetune.py:976] (4/7) Epoch 11, batch 2150, loss[loss=0.1866, simple_loss=0.2674, pruned_loss=0.05289, over 4908.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2603, pruned_loss=0.06511, over 953708.37 frames. ], batch size: 35, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:09:13,348 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=59431.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:09:15,688 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.234e+02 1.596e+02 1.893e+02 2.254e+02 5.168e+02, threshold=3.786e+02, percent-clipped=3.0 +2023-03-26 13:09:54,868 INFO [finetune.py:976] (4/7) Epoch 11, batch 2200, loss[loss=0.231, simple_loss=0.2893, pruned_loss=0.08635, over 4912.00 frames. ], tot_loss[loss=0.1979, simple_loss=0.2631, pruned_loss=0.06631, over 952419.71 frames. ], batch size: 37, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:09:56,664 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=59479.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:10:13,066 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59502.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:10:15,547 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8496, 1.7955, 1.6633, 1.8958, 1.2838, 4.4410, 1.6704, 2.2912], + device='cuda:4'), covar=tensor([0.3338, 0.2360, 0.2102, 0.2296, 0.1839, 0.0105, 0.2534, 0.1207], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0115, 0.0119, 0.0122, 0.0114, 0.0098, 0.0098, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 13:10:25,070 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59520.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:10:30,585 INFO [finetune.py:976] (4/7) Epoch 11, batch 2250, loss[loss=0.2128, simple_loss=0.2789, pruned_loss=0.07339, over 4830.00 frames. ], tot_loss[loss=0.1983, simple_loss=0.2635, pruned_loss=0.06648, over 954342.95 frames. ], batch size: 30, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:10:37,559 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.128e+02 1.729e+02 2.023e+02 2.518e+02 3.990e+02, threshold=4.047e+02, percent-clipped=2.0 +2023-03-26 13:11:00,967 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.40 vs. limit=5.0 +2023-03-26 13:11:02,694 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=59562.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:11:03,341 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59563.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:11:13,471 INFO [finetune.py:976] (4/7) Epoch 11, batch 2300, loss[loss=0.1991, simple_loss=0.265, pruned_loss=0.06658, over 4889.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2627, pruned_loss=0.06523, over 953831.61 frames. ], batch size: 43, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:11:35,296 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=59610.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:11:47,098 INFO [finetune.py:976] (4/7) Epoch 11, batch 2350, loss[loss=0.1677, simple_loss=0.2435, pruned_loss=0.04591, over 4900.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2607, pruned_loss=0.06443, over 954544.94 frames. ], batch size: 46, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:11:52,463 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.309e+01 1.451e+02 1.728e+02 2.097e+02 4.600e+02, threshold=3.455e+02, percent-clipped=1.0 +2023-03-26 13:12:19,962 INFO [finetune.py:976] (4/7) Epoch 11, batch 2400, loss[loss=0.1918, simple_loss=0.2418, pruned_loss=0.07091, over 4665.00 frames. ], tot_loss[loss=0.1921, simple_loss=0.2574, pruned_loss=0.06342, over 954013.37 frames. ], batch size: 23, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:12:53,270 INFO [finetune.py:976] (4/7) Epoch 11, batch 2450, loss[loss=0.1862, simple_loss=0.2523, pruned_loss=0.06004, over 4769.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2539, pruned_loss=0.06217, over 951994.21 frames. ], batch size: 26, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:12:57,637 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-26 13:13:01,211 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.102e+02 1.641e+02 1.877e+02 2.149e+02 5.374e+02, threshold=3.753e+02, percent-clipped=2.0 +2023-03-26 13:13:37,053 INFO [finetune.py:976] (4/7) Epoch 11, batch 2500, loss[loss=0.2277, simple_loss=0.2589, pruned_loss=0.09821, over 4116.00 frames. ], tot_loss[loss=0.194, simple_loss=0.2587, pruned_loss=0.06462, over 952162.73 frames. ], batch size: 18, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:14:29,697 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=59820.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:14:33,891 INFO [finetune.py:976] (4/7) Epoch 11, batch 2550, loss[loss=0.188, simple_loss=0.2561, pruned_loss=0.05996, over 4745.00 frames. ], tot_loss[loss=0.197, simple_loss=0.2628, pruned_loss=0.06557, over 953260.25 frames. ], batch size: 59, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:14:40,182 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.636e+02 1.885e+02 2.323e+02 4.849e+02, threshold=3.771e+02, percent-clipped=2.0 +2023-03-26 13:14:42,128 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7395, 3.6697, 3.5622, 1.8113, 3.8149, 2.8888, 0.8518, 2.5063], + device='cuda:4'), covar=tensor([0.2220, 0.2059, 0.1391, 0.3144, 0.0988, 0.0929, 0.4044, 0.1375], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0174, 0.0160, 0.0128, 0.0155, 0.0121, 0.0145, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 13:14:49,223 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59848.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:14:55,388 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6585, 1.0681, 0.9922, 1.5794, 2.0317, 1.3985, 1.2667, 1.5358], + device='cuda:4'), covar=tensor([0.1597, 0.2364, 0.1889, 0.1287, 0.1958, 0.1937, 0.1649, 0.2030], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0097, 0.0114, 0.0094, 0.0121, 0.0096, 0.0100, 0.0092], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 13:14:57,187 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59858.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:15:03,805 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=59868.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:15:09,218 INFO [finetune.py:976] (4/7) Epoch 11, batch 2600, loss[loss=0.2116, simple_loss=0.2699, pruned_loss=0.0766, over 4823.00 frames. ], tot_loss[loss=0.1987, simple_loss=0.2648, pruned_loss=0.06631, over 952638.97 frames. ], batch size: 33, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:15:17,490 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7571, 1.2406, 1.0006, 1.6425, 2.1939, 1.2105, 1.5453, 1.7126], + device='cuda:4'), covar=tensor([0.1480, 0.2123, 0.1899, 0.1165, 0.1878, 0.1904, 0.1463, 0.1847], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0098, 0.0115, 0.0094, 0.0121, 0.0096, 0.0101, 0.0092], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 13:15:18,110 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59889.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:15:31,229 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59909.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:15:36,126 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9428, 1.7716, 1.5148, 1.7250, 1.6866, 1.7118, 1.7658, 2.4835], + device='cuda:4'), covar=tensor([0.4495, 0.4791, 0.3669, 0.4454, 0.4634, 0.2573, 0.4223, 0.1779], + device='cuda:4'), in_proj_covar=tensor([0.0286, 0.0259, 0.0223, 0.0279, 0.0242, 0.0210, 0.0246, 0.0215], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:15:42,443 INFO [finetune.py:976] (4/7) Epoch 11, batch 2650, loss[loss=0.2297, simple_loss=0.3006, pruned_loss=0.07942, over 4705.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.2649, pruned_loss=0.0657, over 954590.29 frames. ], batch size: 59, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:15:46,233 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1450, 1.3142, 1.4380, 0.6034, 1.2795, 1.5428, 1.5998, 1.2487], + device='cuda:4'), covar=tensor([0.1009, 0.0619, 0.0520, 0.0539, 0.0480, 0.0598, 0.0342, 0.0746], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0154, 0.0122, 0.0133, 0.0130, 0.0125, 0.0144, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.5024e-05, 1.1295e-04, 8.7883e-05, 9.6035e-05, 9.2916e-05, 9.0741e-05, + 1.0529e-04, 1.0734e-04], device='cuda:4') +2023-03-26 13:15:47,332 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.166e+02 1.549e+02 1.976e+02 2.444e+02 3.877e+02, threshold=3.952e+02, percent-clipped=1.0 +2023-03-26 13:16:03,051 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59950.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:16:29,562 INFO [finetune.py:976] (4/7) Epoch 11, batch 2700, loss[loss=0.163, simple_loss=0.2357, pruned_loss=0.04519, over 4911.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2625, pruned_loss=0.06441, over 955390.89 frames. ], batch size: 36, lr: 3.70e-03, grad_scale: 32.0 +2023-03-26 13:16:43,933 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3220, 2.0740, 2.7739, 1.6906, 2.5022, 2.7299, 1.9362, 2.7522], + device='cuda:4'), covar=tensor([0.1687, 0.2149, 0.1557, 0.2659, 0.0969, 0.1797, 0.2970, 0.1151], + device='cuda:4'), in_proj_covar=tensor([0.0199, 0.0206, 0.0194, 0.0190, 0.0178, 0.0216, 0.0218, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:16:46,793 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.5783, 5.1855, 4.9144, 3.3225, 5.2496, 4.2146, 1.5823, 4.0236], + device='cuda:4'), covar=tensor([0.1607, 0.1389, 0.1080, 0.2293, 0.0730, 0.0651, 0.3675, 0.0979], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0176, 0.0161, 0.0129, 0.0157, 0.0122, 0.0147, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 13:16:57,163 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5132, 1.4463, 1.2795, 1.4488, 1.8066, 1.5833, 1.5212, 1.2908], + device='cuda:4'), covar=tensor([0.0274, 0.0268, 0.0585, 0.0286, 0.0188, 0.0538, 0.0273, 0.0362], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0109, 0.0140, 0.0115, 0.0102, 0.0104, 0.0093, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.1390e-05, 8.5021e-05, 1.1155e-04, 8.9683e-05, 8.0096e-05, 7.7482e-05, + 7.0435e-05, 8.3393e-05], device='cuda:4') +2023-03-26 13:17:04,325 INFO [finetune.py:976] (4/7) Epoch 11, batch 2750, loss[loss=0.1946, simple_loss=0.2567, pruned_loss=0.06629, over 4829.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.2596, pruned_loss=0.06351, over 954487.06 frames. ], batch size: 25, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:17:09,210 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.603e+02 1.823e+02 2.284e+02 4.397e+02, threshold=3.646e+02, percent-clipped=1.0 +2023-03-26 13:17:12,362 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60040.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:17:21,588 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60052.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:17:30,377 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60065.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:17:37,359 INFO [finetune.py:976] (4/7) Epoch 11, batch 2800, loss[loss=0.1828, simple_loss=0.2461, pruned_loss=0.05978, over 4836.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2576, pruned_loss=0.06364, over 955768.56 frames. ], batch size: 30, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:17:38,100 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60078.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:17:54,527 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60101.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 13:18:02,756 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60113.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 13:18:10,607 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60126.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:18:11,111 INFO [finetune.py:976] (4/7) Epoch 11, batch 2850, loss[loss=0.1806, simple_loss=0.2394, pruned_loss=0.06089, over 4869.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.257, pruned_loss=0.06398, over 956077.34 frames. ], batch size: 31, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:18:17,976 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.190e+02 1.579e+02 1.818e+02 2.348e+02 4.165e+02, threshold=3.636e+02, percent-clipped=3.0 +2023-03-26 13:18:21,051 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60139.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:18:26,663 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6424, 1.1385, 0.8524, 1.5323, 1.9873, 1.2025, 1.4287, 1.5405], + device='cuda:4'), covar=tensor([0.1489, 0.2222, 0.2100, 0.1182, 0.2030, 0.2076, 0.1493, 0.1970], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0098, 0.0115, 0.0094, 0.0122, 0.0096, 0.0101, 0.0092], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 13:18:39,221 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60158.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:18:51,669 INFO [finetune.py:976] (4/7) Epoch 11, batch 2900, loss[loss=0.2691, simple_loss=0.3235, pruned_loss=0.1074, over 4799.00 frames. ], tot_loss[loss=0.1961, simple_loss=0.261, pruned_loss=0.06558, over 955513.23 frames. ], batch size: 45, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:19:12,014 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60198.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:19:21,318 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60204.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:19:27,581 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60206.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:19:51,300 INFO [finetune.py:976] (4/7) Epoch 11, batch 2950, loss[loss=0.1674, simple_loss=0.2389, pruned_loss=0.04801, over 4890.00 frames. ], tot_loss[loss=0.1978, simple_loss=0.2635, pruned_loss=0.0661, over 956178.78 frames. ], batch size: 32, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:20:00,139 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.723e+02 2.035e+02 2.444e+02 4.360e+02, threshold=4.070e+02, percent-clipped=6.0 +2023-03-26 13:20:06,144 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7688, 1.6533, 2.1811, 1.4524, 1.7882, 1.9746, 1.5782, 2.2193], + device='cuda:4'), covar=tensor([0.1506, 0.2124, 0.1380, 0.2087, 0.0935, 0.1647, 0.2792, 0.0846], + device='cuda:4'), in_proj_covar=tensor([0.0196, 0.0203, 0.0191, 0.0188, 0.0176, 0.0213, 0.0217, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:20:06,686 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60245.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:20:16,636 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60259.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:20:28,430 INFO [finetune.py:976] (4/7) Epoch 11, batch 3000, loss[loss=0.2226, simple_loss=0.2885, pruned_loss=0.07838, over 4920.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.2651, pruned_loss=0.06681, over 953575.17 frames. ], batch size: 33, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:20:28,430 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 13:20:37,598 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8222, 1.8652, 1.9647, 1.1451, 2.0064, 1.9402, 1.9612, 1.6562], + device='cuda:4'), covar=tensor([0.0596, 0.0671, 0.0581, 0.0865, 0.0662, 0.0681, 0.0553, 0.1081], + device='cuda:4'), in_proj_covar=tensor([0.0137, 0.0136, 0.0144, 0.0127, 0.0121, 0.0145, 0.0146, 0.0166], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:20:38,898 INFO [finetune.py:1010] (4/7) Epoch 11, validation: loss=0.1572, simple_loss=0.2284, pruned_loss=0.04301, over 2265189.00 frames. +2023-03-26 13:20:38,899 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 13:20:48,015 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-26 13:21:13,698 INFO [finetune.py:976] (4/7) Epoch 11, batch 3050, loss[loss=0.1633, simple_loss=0.2268, pruned_loss=0.04986, over 4918.00 frames. ], tot_loss[loss=0.1987, simple_loss=0.2647, pruned_loss=0.06637, over 952672.23 frames. ], batch size: 33, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:21:19,480 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.587e+02 1.939e+02 2.482e+02 4.597e+02, threshold=3.877e+02, percent-clipped=2.0 +2023-03-26 13:21:56,081 INFO [finetune.py:976] (4/7) Epoch 11, batch 3100, loss[loss=0.19, simple_loss=0.2555, pruned_loss=0.06218, over 4861.00 frames. ], tot_loss[loss=0.1975, simple_loss=0.2627, pruned_loss=0.06615, over 953353.28 frames. ], batch size: 34, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:21:56,165 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1241, 3.5292, 3.7329, 3.9497, 3.8923, 3.6198, 4.1773, 1.4851], + device='cuda:4'), covar=tensor([0.0742, 0.0855, 0.0831, 0.0995, 0.1140, 0.1522, 0.0702, 0.5065], + device='cuda:4'), in_proj_covar=tensor([0.0348, 0.0243, 0.0275, 0.0291, 0.0330, 0.0284, 0.0301, 0.0294], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:21:58,585 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1836, 2.0452, 1.6845, 2.0568, 2.2027, 1.9061, 2.4719, 2.1575], + device='cuda:4'), covar=tensor([0.1318, 0.2180, 0.3139, 0.2717, 0.2541, 0.1598, 0.3289, 0.1918], + device='cuda:4'), in_proj_covar=tensor([0.0175, 0.0186, 0.0231, 0.0253, 0.0237, 0.0195, 0.0211, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:22:08,707 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60396.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 13:22:12,633 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5634, 2.3296, 1.8875, 2.7442, 2.6295, 2.2168, 3.0645, 2.5178], + device='cuda:4'), covar=tensor([0.1473, 0.2735, 0.3637, 0.3132, 0.2718, 0.1827, 0.3178, 0.2135], + device='cuda:4'), in_proj_covar=tensor([0.0175, 0.0187, 0.0232, 0.0254, 0.0238, 0.0196, 0.0212, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:22:16,691 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60408.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 13:22:25,059 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60421.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:22:29,560 INFO [finetune.py:976] (4/7) Epoch 11, batch 3150, loss[loss=0.1687, simple_loss=0.2438, pruned_loss=0.04683, over 4762.00 frames. ], tot_loss[loss=0.1946, simple_loss=0.2597, pruned_loss=0.06477, over 955295.77 frames. ], batch size: 27, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:22:34,351 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60434.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:22:34,874 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.624e+02 1.838e+02 2.200e+02 4.980e+02, threshold=3.676e+02, percent-clipped=1.0 +2023-03-26 13:22:59,671 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.64 vs. limit=2.0 +2023-03-26 13:23:01,694 INFO [finetune.py:976] (4/7) Epoch 11, batch 3200, loss[loss=0.1622, simple_loss=0.2287, pruned_loss=0.04787, over 4913.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2563, pruned_loss=0.06354, over 956167.17 frames. ], batch size: 43, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:23:20,606 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60504.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:23:36,825 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6540, 1.5496, 1.3932, 1.7573, 1.9858, 1.8075, 1.2634, 1.4050], + device='cuda:4'), covar=tensor([0.2063, 0.1978, 0.1827, 0.1585, 0.1607, 0.1121, 0.2617, 0.1850], + device='cuda:4'), in_proj_covar=tensor([0.0238, 0.0207, 0.0208, 0.0189, 0.0241, 0.0182, 0.0212, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:23:37,314 INFO [finetune.py:976] (4/7) Epoch 11, batch 3250, loss[loss=0.2141, simple_loss=0.2829, pruned_loss=0.07263, over 4894.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2566, pruned_loss=0.06329, over 957108.34 frames. ], batch size: 35, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:23:48,950 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.174e+02 1.626e+02 1.982e+02 2.397e+02 3.737e+02, threshold=3.964e+02, percent-clipped=1.0 +2023-03-26 13:23:59,839 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60545.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:24:01,678 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2991, 2.1760, 1.7268, 0.7974, 1.8904, 1.7719, 1.6008, 1.9408], + device='cuda:4'), covar=tensor([0.0839, 0.0854, 0.1563, 0.1974, 0.1386, 0.2339, 0.2324, 0.1000], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0200, 0.0202, 0.0187, 0.0216, 0.0210, 0.0224, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:24:04,034 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60552.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:24:05,273 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60554.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:24:13,140 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5867, 2.2637, 2.5638, 2.4619, 2.2148, 2.1607, 2.3862, 2.2935], + device='cuda:4'), covar=tensor([0.4312, 0.4705, 0.3729, 0.4828, 0.5441, 0.4235, 0.5336, 0.3681], + device='cuda:4'), in_proj_covar=tensor([0.0238, 0.0239, 0.0255, 0.0260, 0.0255, 0.0231, 0.0275, 0.0232], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:24:27,367 INFO [finetune.py:976] (4/7) Epoch 11, batch 3300, loss[loss=0.221, simple_loss=0.2864, pruned_loss=0.07779, over 4744.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2618, pruned_loss=0.0657, over 955675.03 frames. ], batch size: 54, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:24:45,616 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60593.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:25:28,937 INFO [finetune.py:976] (4/7) Epoch 11, batch 3350, loss[loss=0.203, simple_loss=0.2658, pruned_loss=0.07008, over 4921.00 frames. ], tot_loss[loss=0.198, simple_loss=0.2635, pruned_loss=0.06623, over 956455.24 frames. ], batch size: 33, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:25:34,912 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.073e+02 1.701e+02 2.036e+02 2.469e+02 3.577e+02, threshold=4.071e+02, percent-clipped=0.0 +2023-03-26 13:25:39,289 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 13:26:02,927 INFO [finetune.py:976] (4/7) Epoch 11, batch 3400, loss[loss=0.2316, simple_loss=0.2981, pruned_loss=0.08253, over 4893.00 frames. ], tot_loss[loss=0.1998, simple_loss=0.2652, pruned_loss=0.06722, over 954776.30 frames. ], batch size: 35, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:26:03,057 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3321, 1.5101, 1.1852, 1.4466, 1.7809, 1.5540, 1.4492, 1.2509], + device='cuda:4'), covar=tensor([0.0440, 0.0313, 0.0671, 0.0325, 0.0234, 0.0579, 0.0305, 0.0425], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0109, 0.0141, 0.0114, 0.0102, 0.0104, 0.0093, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.1097e-05, 8.4865e-05, 1.1210e-04, 8.9487e-05, 7.9976e-05, 7.7287e-05, + 7.0604e-05, 8.3417e-05], device='cuda:4') +2023-03-26 13:26:04,076 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.68 vs. limit=5.0 +2023-03-26 13:26:16,529 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60696.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:26:24,737 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60708.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:26:30,788 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60718.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:26:32,566 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60721.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:26:36,599 INFO [finetune.py:976] (4/7) Epoch 11, batch 3450, loss[loss=0.1567, simple_loss=0.2235, pruned_loss=0.04498, over 4282.00 frames. ], tot_loss[loss=0.1987, simple_loss=0.2644, pruned_loss=0.06654, over 955656.33 frames. ], batch size: 66, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:26:41,004 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60734.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:26:41,511 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.902e+01 1.594e+02 1.892e+02 2.253e+02 3.493e+02, threshold=3.783e+02, percent-clipped=0.0 +2023-03-26 13:26:52,729 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60744.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:26:55,212 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7121, 1.4592, 2.2614, 3.3149, 2.4163, 2.4490, 1.2303, 2.6787], + device='cuda:4'), covar=tensor([0.1780, 0.1632, 0.1249, 0.0625, 0.0752, 0.1376, 0.1857, 0.0590], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0117, 0.0134, 0.0165, 0.0101, 0.0138, 0.0126, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 13:27:06,421 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60756.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:27:25,459 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60769.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:27:36,414 INFO [finetune.py:976] (4/7) Epoch 11, batch 3500, loss[loss=0.1777, simple_loss=0.2415, pruned_loss=0.05692, over 4843.00 frames. ], tot_loss[loss=0.1959, simple_loss=0.2615, pruned_loss=0.06516, over 956802.25 frames. ], batch size: 44, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:27:37,765 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60779.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:27:45,018 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60782.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:28:15,228 INFO [finetune.py:976] (4/7) Epoch 11, batch 3550, loss[loss=0.1992, simple_loss=0.2544, pruned_loss=0.07196, over 4193.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2586, pruned_loss=0.06429, over 955371.21 frames. ], batch size: 18, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:28:20,663 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.181e+02 1.566e+02 1.863e+02 2.348e+02 4.575e+02, threshold=3.726e+02, percent-clipped=4.0 +2023-03-26 13:28:30,115 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2136, 2.0072, 1.7675, 2.1191, 1.9502, 1.9110, 1.9488, 2.6935], + device='cuda:4'), covar=tensor([0.4476, 0.5153, 0.3736, 0.4108, 0.4338, 0.2735, 0.4622, 0.1806], + device='cuda:4'), in_proj_covar=tensor([0.0286, 0.0261, 0.0224, 0.0278, 0.0245, 0.0210, 0.0246, 0.0216], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:28:30,985 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-03-26 13:28:34,188 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60854.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:28:49,077 INFO [finetune.py:976] (4/7) Epoch 11, batch 3600, loss[loss=0.2274, simple_loss=0.2872, pruned_loss=0.08376, over 4836.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2569, pruned_loss=0.06394, over 956874.50 frames. ], batch size: 47, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:29:17,743 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=60902.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:29:30,749 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9662, 1.7730, 1.5161, 1.7315, 1.6996, 1.6772, 1.7619, 2.4825], + device='cuda:4'), covar=tensor([0.4206, 0.4537, 0.3514, 0.4190, 0.4503, 0.2464, 0.4277, 0.1715], + device='cuda:4'), in_proj_covar=tensor([0.0282, 0.0258, 0.0222, 0.0276, 0.0242, 0.0208, 0.0244, 0.0214], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:29:39,498 INFO [finetune.py:976] (4/7) Epoch 11, batch 3650, loss[loss=0.1794, simple_loss=0.2459, pruned_loss=0.05647, over 4757.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2592, pruned_loss=0.06477, over 958490.89 frames. ], batch size: 26, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:29:40,887 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4346, 2.2612, 1.7969, 0.8612, 2.0086, 1.8225, 1.7536, 1.9523], + device='cuda:4'), covar=tensor([0.0818, 0.0863, 0.1762, 0.2214, 0.1496, 0.2243, 0.2118, 0.1090], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0201, 0.0204, 0.0187, 0.0216, 0.0210, 0.0224, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:29:44,366 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.638e+02 1.962e+02 2.312e+02 3.604e+02, threshold=3.924e+02, percent-clipped=0.0 +2023-03-26 13:30:18,696 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.05 vs. limit=5.0 +2023-03-26 13:30:33,792 INFO [finetune.py:976] (4/7) Epoch 11, batch 3700, loss[loss=0.2015, simple_loss=0.267, pruned_loss=0.06803, over 4189.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.2625, pruned_loss=0.06552, over 956439.67 frames. ], batch size: 65, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:31:12,862 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3946, 1.5450, 1.5639, 0.7759, 1.5420, 1.7541, 1.8228, 1.3302], + device='cuda:4'), covar=tensor([0.0783, 0.0491, 0.0452, 0.0525, 0.0429, 0.0497, 0.0272, 0.0731], + device='cuda:4'), in_proj_covar=tensor([0.0127, 0.0153, 0.0122, 0.0132, 0.0130, 0.0125, 0.0143, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.3998e-05, 1.1233e-04, 8.7485e-05, 9.5841e-05, 9.2782e-05, 9.0688e-05, + 1.0472e-04, 1.0708e-04], device='cuda:4') +2023-03-26 13:31:15,838 INFO [finetune.py:976] (4/7) Epoch 11, batch 3750, loss[loss=0.1542, simple_loss=0.2389, pruned_loss=0.03478, over 4810.00 frames. ], tot_loss[loss=0.1973, simple_loss=0.2634, pruned_loss=0.06561, over 955600.05 frames. ], batch size: 39, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:31:20,654 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.587e+02 1.819e+02 2.276e+02 4.586e+02, threshold=3.638e+02, percent-clipped=1.0 +2023-03-26 13:31:38,968 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61061.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:31:47,664 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=61074.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:31:49,399 INFO [finetune.py:976] (4/7) Epoch 11, batch 3800, loss[loss=0.1787, simple_loss=0.2474, pruned_loss=0.05502, over 4812.00 frames. ], tot_loss[loss=0.1981, simple_loss=0.2645, pruned_loss=0.06585, over 954368.19 frames. ], batch size: 33, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:32:29,699 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61122.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:32:32,631 INFO [finetune.py:976] (4/7) Epoch 11, batch 3850, loss[loss=0.2209, simple_loss=0.2818, pruned_loss=0.07999, over 4720.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2629, pruned_loss=0.06524, over 955589.19 frames. ], batch size: 54, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:32:37,921 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.075e+02 1.518e+02 1.864e+02 2.279e+02 4.215e+02, threshold=3.727e+02, percent-clipped=1.0 +2023-03-26 13:33:05,949 INFO [finetune.py:976] (4/7) Epoch 11, batch 3900, loss[loss=0.1543, simple_loss=0.2217, pruned_loss=0.04343, over 4756.00 frames. ], tot_loss[loss=0.1946, simple_loss=0.26, pruned_loss=0.06464, over 955604.49 frames. ], batch size: 28, lr: 3.69e-03, grad_scale: 32.0 +2023-03-26 13:33:12,643 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1810, 1.3354, 1.4208, 0.6129, 1.2956, 1.5937, 1.5997, 1.3252], + device='cuda:4'), covar=tensor([0.0916, 0.0642, 0.0451, 0.0601, 0.0497, 0.0696, 0.0349, 0.0754], + device='cuda:4'), in_proj_covar=tensor([0.0128, 0.0154, 0.0122, 0.0132, 0.0130, 0.0125, 0.0144, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.4402e-05, 1.1271e-04, 8.7590e-05, 9.5694e-05, 9.2968e-05, 9.1001e-05, + 1.0518e-04, 1.0708e-04], device='cuda:4') +2023-03-26 13:33:39,745 INFO [finetune.py:976] (4/7) Epoch 11, batch 3950, loss[loss=0.1943, simple_loss=0.257, pruned_loss=0.06579, over 4310.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2574, pruned_loss=0.06409, over 955339.73 frames. ], batch size: 65, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:33:45,060 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.129e+02 1.570e+02 1.907e+02 2.309e+02 4.377e+02, threshold=3.813e+02, percent-clipped=3.0 +2023-03-26 13:34:12,380 INFO [finetune.py:976] (4/7) Epoch 11, batch 4000, loss[loss=0.1831, simple_loss=0.2296, pruned_loss=0.06825, over 3785.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2579, pruned_loss=0.06467, over 955498.51 frames. ], batch size: 16, lr: 3.68e-03, grad_scale: 64.0 +2023-03-26 13:34:55,752 INFO [finetune.py:976] (4/7) Epoch 11, batch 4050, loss[loss=0.2341, simple_loss=0.3107, pruned_loss=0.07875, over 4746.00 frames. ], tot_loss[loss=0.1959, simple_loss=0.2613, pruned_loss=0.06524, over 955309.35 frames. ], batch size: 54, lr: 3.68e-03, grad_scale: 64.0 +2023-03-26 13:35:04,888 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.652e+02 2.086e+02 2.571e+02 4.987e+02, threshold=4.171e+02, percent-clipped=6.0 +2023-03-26 13:35:41,833 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=61374.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:35:43,532 INFO [finetune.py:976] (4/7) Epoch 11, batch 4100, loss[loss=0.1892, simple_loss=0.262, pruned_loss=0.05826, over 4796.00 frames. ], tot_loss[loss=0.197, simple_loss=0.2633, pruned_loss=0.06528, over 954156.00 frames. ], batch size: 25, lr: 3.68e-03, grad_scale: 64.0 +2023-03-26 13:35:48,353 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5803, 1.4177, 2.0924, 3.2096, 2.2621, 2.2715, 1.0134, 2.5929], + device='cuda:4'), covar=tensor([0.1697, 0.1499, 0.1262, 0.0628, 0.0763, 0.1504, 0.1799, 0.0507], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0116, 0.0134, 0.0165, 0.0101, 0.0138, 0.0126, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 13:36:16,505 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=61417.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:36:18,922 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9713, 1.9096, 1.5611, 1.8049, 1.7693, 1.6928, 1.8542, 2.4591], + device='cuda:4'), covar=tensor([0.4588, 0.4973, 0.3776, 0.4685, 0.4684, 0.2729, 0.4536, 0.1891], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0262, 0.0224, 0.0279, 0.0245, 0.0211, 0.0247, 0.0217], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:36:20,076 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=61422.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:36:26,638 INFO [finetune.py:976] (4/7) Epoch 11, batch 4150, loss[loss=0.1961, simple_loss=0.2719, pruned_loss=0.06021, over 4819.00 frames. ], tot_loss[loss=0.1978, simple_loss=0.2641, pruned_loss=0.06573, over 953449.68 frames. ], batch size: 33, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:36:32,503 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.161e+02 1.629e+02 1.982e+02 2.519e+02 5.426e+02, threshold=3.964e+02, percent-clipped=4.0 +2023-03-26 13:36:59,823 INFO [finetune.py:976] (4/7) Epoch 11, batch 4200, loss[loss=0.1848, simple_loss=0.2438, pruned_loss=0.06287, over 4796.00 frames. ], tot_loss[loss=0.1977, simple_loss=0.2643, pruned_loss=0.06558, over 954928.13 frames. ], batch size: 25, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:37:35,243 INFO [finetune.py:976] (4/7) Epoch 11, batch 4250, loss[loss=0.1693, simple_loss=0.2393, pruned_loss=0.04969, over 4758.00 frames. ], tot_loss[loss=0.1946, simple_loss=0.2607, pruned_loss=0.06419, over 955084.27 frames. ], batch size: 28, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:37:44,261 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.5919, 1.6130, 1.5996, 0.9177, 1.6307, 1.8099, 1.8716, 1.3943], + device='cuda:4'), covar=tensor([0.0737, 0.0466, 0.0508, 0.0520, 0.0431, 0.0521, 0.0266, 0.0595], + device='cuda:4'), in_proj_covar=tensor([0.0128, 0.0154, 0.0122, 0.0133, 0.0131, 0.0126, 0.0144, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.4627e-05, 1.1312e-04, 8.8174e-05, 9.5866e-05, 9.3420e-05, 9.1856e-05, + 1.0528e-04, 1.0734e-04], device='cuda:4') +2023-03-26 13:37:45,939 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.771e+01 1.547e+02 1.858e+02 2.245e+02 5.805e+02, threshold=3.715e+02, percent-clipped=2.0 +2023-03-26 13:38:07,716 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0387, 1.9594, 1.7889, 1.9899, 1.4451, 4.4913, 1.7482, 2.2526], + device='cuda:4'), covar=tensor([0.2975, 0.2234, 0.1923, 0.2054, 0.1599, 0.0092, 0.2306, 0.1154], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0115, 0.0119, 0.0123, 0.0115, 0.0098, 0.0098, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 13:38:10,188 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1238, 1.9139, 2.1329, 0.9165, 2.3555, 2.3538, 2.0596, 1.8771], + device='cuda:4'), covar=tensor([0.1075, 0.0951, 0.0517, 0.0825, 0.0570, 0.0884, 0.0634, 0.0805], + device='cuda:4'), in_proj_covar=tensor([0.0128, 0.0155, 0.0122, 0.0133, 0.0131, 0.0126, 0.0144, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.4850e-05, 1.1334e-04, 8.8183e-05, 9.6092e-05, 9.3600e-05, 9.1969e-05, + 1.0534e-04, 1.0743e-04], device='cuda:4') +2023-03-26 13:38:15,491 INFO [finetune.py:976] (4/7) Epoch 11, batch 4300, loss[loss=0.1898, simple_loss=0.2293, pruned_loss=0.07516, over 4202.00 frames. ], tot_loss[loss=0.1922, simple_loss=0.2578, pruned_loss=0.06326, over 954601.59 frames. ], batch size: 18, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:38:35,747 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0974, 1.3398, 1.2044, 1.3563, 1.4786, 2.5029, 1.2804, 1.5005], + device='cuda:4'), covar=tensor([0.0991, 0.1717, 0.1017, 0.0923, 0.1571, 0.0370, 0.1482, 0.1608], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0084, 0.0078], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 13:38:38,756 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6816, 1.1998, 0.8219, 1.6062, 2.0944, 1.0237, 1.4569, 1.5873], + device='cuda:4'), covar=tensor([0.1560, 0.2111, 0.1958, 0.1208, 0.1974, 0.1874, 0.1436, 0.2013], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0092, 0.0119, 0.0094, 0.0098, 0.0090], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 13:38:48,431 INFO [finetune.py:976] (4/7) Epoch 11, batch 4350, loss[loss=0.1476, simple_loss=0.2285, pruned_loss=0.03329, over 4733.00 frames. ], tot_loss[loss=0.189, simple_loss=0.2547, pruned_loss=0.06165, over 955955.45 frames. ], batch size: 23, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:38:54,823 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.000e+02 1.580e+02 1.801e+02 2.212e+02 3.446e+02, threshold=3.603e+02, percent-clipped=0.0 +2023-03-26 13:39:05,656 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2701, 1.0381, 1.0399, 1.0273, 1.5393, 1.4349, 1.3766, 1.0694], + device='cuda:4'), covar=tensor([0.0294, 0.0312, 0.0732, 0.0370, 0.0233, 0.0384, 0.0245, 0.0397], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0108, 0.0140, 0.0114, 0.0102, 0.0103, 0.0092, 0.0107], + device='cuda:4'), out_proj_covar=tensor([7.0732e-05, 8.3842e-05, 1.1134e-04, 8.8970e-05, 7.9377e-05, 7.6146e-05, + 6.9638e-05, 8.2237e-05], device='cuda:4') +2023-03-26 13:39:06,886 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9957, 1.9635, 1.5362, 1.8041, 1.9801, 1.6623, 2.2194, 2.0194], + device='cuda:4'), covar=tensor([0.1513, 0.2119, 0.3448, 0.3123, 0.2846, 0.1994, 0.4400, 0.2033], + device='cuda:4'), in_proj_covar=tensor([0.0175, 0.0186, 0.0232, 0.0252, 0.0238, 0.0196, 0.0212, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:39:21,859 INFO [finetune.py:976] (4/7) Epoch 11, batch 4400, loss[loss=0.1762, simple_loss=0.225, pruned_loss=0.06367, over 4809.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2563, pruned_loss=0.06217, over 956648.56 frames. ], batch size: 25, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:39:35,893 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4378, 1.4594, 1.2789, 1.4852, 1.7943, 1.6229, 1.5561, 1.2398], + device='cuda:4'), covar=tensor([0.0291, 0.0269, 0.0548, 0.0255, 0.0201, 0.0426, 0.0221, 0.0347], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0108, 0.0140, 0.0114, 0.0102, 0.0103, 0.0092, 0.0107], + device='cuda:4'), out_proj_covar=tensor([7.0914e-05, 8.3967e-05, 1.1139e-04, 8.9114e-05, 7.9526e-05, 7.6283e-05, + 6.9807e-05, 8.2407e-05], device='cuda:4') +2023-03-26 13:39:37,520 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=2.01 vs. limit=2.0 +2023-03-26 13:39:53,784 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=61717.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:40:04,333 INFO [finetune.py:976] (4/7) Epoch 11, batch 4450, loss[loss=0.1432, simple_loss=0.2139, pruned_loss=0.03627, over 4723.00 frames. ], tot_loss[loss=0.1921, simple_loss=0.2587, pruned_loss=0.06271, over 957437.98 frames. ], batch size: 23, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:40:07,485 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61732.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:40:14,298 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.223e+02 1.628e+02 1.977e+02 2.534e+02 3.640e+02, threshold=3.954e+02, percent-clipped=2.0 +2023-03-26 13:40:16,779 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 13:40:23,499 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61743.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:40:49,788 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=61765.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:40:56,858 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.45 vs. limit=5.0 +2023-03-26 13:40:57,018 INFO [finetune.py:976] (4/7) Epoch 11, batch 4500, loss[loss=0.1927, simple_loss=0.2558, pruned_loss=0.06474, over 4768.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2596, pruned_loss=0.06276, over 956311.79 frames. ], batch size: 26, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:40:59,044 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9793, 1.7358, 1.4451, 1.6570, 1.6770, 1.6967, 1.7088, 2.4001], + device='cuda:4'), covar=tensor([0.4312, 0.5321, 0.3659, 0.4549, 0.4405, 0.2650, 0.4422, 0.1935], + device='cuda:4'), in_proj_covar=tensor([0.0284, 0.0260, 0.0223, 0.0277, 0.0243, 0.0209, 0.0246, 0.0216], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:41:07,869 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61793.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:41:16,102 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61804.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:41:33,080 INFO [finetune.py:976] (4/7) Epoch 11, batch 4550, loss[loss=0.1686, simple_loss=0.2247, pruned_loss=0.05624, over 4014.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2618, pruned_loss=0.06397, over 955597.55 frames. ], batch size: 17, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:41:43,497 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.535e+01 1.607e+02 1.951e+02 2.245e+02 3.846e+02, threshold=3.902e+02, percent-clipped=0.0 +2023-03-26 13:42:13,549 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5604, 2.3842, 1.9434, 2.4723, 2.4754, 2.1485, 2.9325, 2.5779], + device='cuda:4'), covar=tensor([0.1309, 0.2636, 0.3249, 0.3173, 0.2698, 0.1613, 0.3183, 0.1793], + device='cuda:4'), in_proj_covar=tensor([0.0175, 0.0186, 0.0233, 0.0253, 0.0239, 0.0197, 0.0211, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:42:15,220 INFO [finetune.py:976] (4/7) Epoch 11, batch 4600, loss[loss=0.1921, simple_loss=0.2429, pruned_loss=0.07066, over 4699.00 frames. ], tot_loss[loss=0.195, simple_loss=0.2617, pruned_loss=0.06413, over 955767.68 frames. ], batch size: 23, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:42:40,453 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61914.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:42:48,596 INFO [finetune.py:976] (4/7) Epoch 11, batch 4650, loss[loss=0.184, simple_loss=0.2513, pruned_loss=0.05838, over 4934.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.259, pruned_loss=0.0631, over 955201.68 frames. ], batch size: 38, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:42:56,047 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.606e+02 1.934e+02 2.317e+02 5.626e+02, threshold=3.867e+02, percent-clipped=3.0 +2023-03-26 13:42:57,782 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.66 vs. limit=5.0 +2023-03-26 13:43:05,790 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9296, 4.0113, 3.8308, 1.8717, 4.2019, 3.1701, 0.7823, 2.9748], + device='cuda:4'), covar=tensor([0.2130, 0.1715, 0.1582, 0.3271, 0.0912, 0.0953, 0.4700, 0.1468], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0173, 0.0158, 0.0127, 0.0154, 0.0121, 0.0145, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 13:43:06,004 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 13:43:31,723 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61975.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:43:32,822 INFO [finetune.py:976] (4/7) Epoch 11, batch 4700, loss[loss=0.1451, simple_loss=0.2053, pruned_loss=0.04243, over 4939.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2552, pruned_loss=0.06172, over 956085.25 frames. ], batch size: 33, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:43:32,932 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2078, 1.7458, 0.6916, 2.1481, 2.5960, 1.8158, 2.1702, 1.9906], + device='cuda:4'), covar=tensor([0.1349, 0.1923, 0.2372, 0.1179, 0.1771, 0.1853, 0.1311, 0.2014], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0093, 0.0119, 0.0094, 0.0099, 0.0090], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 13:44:19,773 INFO [finetune.py:976] (4/7) Epoch 11, batch 4750, loss[loss=0.2171, simple_loss=0.277, pruned_loss=0.07862, over 4873.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.254, pruned_loss=0.06192, over 954901.31 frames. ], batch size: 31, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:44:25,602 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.474e+02 1.769e+02 2.148e+02 4.944e+02, threshold=3.539e+02, percent-clipped=1.0 +2023-03-26 13:44:53,404 INFO [finetune.py:976] (4/7) Epoch 11, batch 4800, loss[loss=0.1722, simple_loss=0.2353, pruned_loss=0.05453, over 4897.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2568, pruned_loss=0.06292, over 952731.57 frames. ], batch size: 35, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:45:06,472 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62088.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:45:13,195 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62099.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:45:13,251 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2174, 1.9387, 2.1690, 1.0023, 2.3906, 2.5540, 2.1296, 1.8458], + device='cuda:4'), covar=tensor([0.0831, 0.0739, 0.0461, 0.0687, 0.0470, 0.0590, 0.0435, 0.0691], + device='cuda:4'), in_proj_covar=tensor([0.0127, 0.0153, 0.0121, 0.0132, 0.0130, 0.0126, 0.0143, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.4033e-05, 1.1199e-04, 8.7084e-05, 9.5236e-05, 9.2986e-05, 9.1277e-05, + 1.0456e-04, 1.0647e-04], device='cuda:4') +2023-03-26 13:45:47,627 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62122.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:45:50,589 INFO [finetune.py:976] (4/7) Epoch 11, batch 4850, loss[loss=0.2198, simple_loss=0.2913, pruned_loss=0.07412, over 4926.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.2589, pruned_loss=0.06322, over 954224.72 frames. ], batch size: 42, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:46:01,539 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.225e+02 1.730e+02 2.037e+02 2.587e+02 8.043e+02, threshold=4.075e+02, percent-clipped=4.0 +2023-03-26 13:46:45,229 INFO [finetune.py:976] (4/7) Epoch 11, batch 4900, loss[loss=0.1882, simple_loss=0.255, pruned_loss=0.0607, over 4818.00 frames. ], tot_loss[loss=0.1941, simple_loss=0.2606, pruned_loss=0.06375, over 953320.56 frames. ], batch size: 30, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:46:54,064 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62183.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:47:49,178 INFO [finetune.py:976] (4/7) Epoch 11, batch 4950, loss[loss=0.2083, simple_loss=0.2743, pruned_loss=0.07116, over 4904.00 frames. ], tot_loss[loss=0.1956, simple_loss=0.2628, pruned_loss=0.06419, over 955435.85 frames. ], batch size: 46, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:47:56,655 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.284e+02 1.728e+02 2.029e+02 2.471e+02 5.736e+02, threshold=4.057e+02, percent-clipped=2.0 +2023-03-26 13:48:18,913 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62270.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:48:20,967 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 13:48:21,235 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.2805, 2.9225, 3.0589, 3.1972, 3.0545, 2.8880, 3.3078, 0.9625], + device='cuda:4'), covar=tensor([0.1063, 0.0950, 0.1010, 0.1050, 0.1592, 0.1615, 0.1095, 0.5276], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0245, 0.0278, 0.0292, 0.0333, 0.0284, 0.0304, 0.0297], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:48:21,270 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62273.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:48:24,032 INFO [finetune.py:976] (4/7) Epoch 11, batch 5000, loss[loss=0.2029, simple_loss=0.2696, pruned_loss=0.06814, over 4895.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.2595, pruned_loss=0.06279, over 954269.00 frames. ], batch size: 35, lr: 3.68e-03, grad_scale: 32.0 +2023-03-26 13:48:26,092 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-26 13:48:34,216 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2414, 2.0924, 1.6984, 1.8757, 2.1155, 1.8611, 2.3440, 2.2271], + device='cuda:4'), covar=tensor([0.1324, 0.2285, 0.3327, 0.2802, 0.2608, 0.1614, 0.4246, 0.1709], + device='cuda:4'), in_proj_covar=tensor([0.0175, 0.0186, 0.0232, 0.0252, 0.0239, 0.0196, 0.0210, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:48:41,444 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.4624, 3.8954, 4.0483, 4.2684, 4.1726, 3.9289, 4.5143, 1.3204], + device='cuda:4'), covar=tensor([0.0835, 0.0824, 0.0875, 0.0996, 0.1296, 0.1599, 0.0677, 0.5693], + device='cuda:4'), in_proj_covar=tensor([0.0348, 0.0244, 0.0277, 0.0290, 0.0331, 0.0283, 0.0302, 0.0294], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:48:57,127 INFO [finetune.py:976] (4/7) Epoch 11, batch 5050, loss[loss=0.1798, simple_loss=0.2372, pruned_loss=0.06117, over 4789.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2586, pruned_loss=0.06306, over 955867.15 frames. ], batch size: 25, lr: 3.68e-03, grad_scale: 16.0 +2023-03-26 13:49:02,475 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62334.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:49:04,173 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.504e+02 1.759e+02 2.068e+02 4.473e+02, threshold=3.518e+02, percent-clipped=1.0 +2023-03-26 13:49:32,190 INFO [finetune.py:976] (4/7) Epoch 11, batch 5100, loss[loss=0.178, simple_loss=0.2461, pruned_loss=0.05495, over 4832.00 frames. ], tot_loss[loss=0.1914, simple_loss=0.2568, pruned_loss=0.063, over 956820.41 frames. ], batch size: 30, lr: 3.68e-03, grad_scale: 16.0 +2023-03-26 13:49:40,040 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62388.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:49:42,306 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62391.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:49:47,650 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62399.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:50:05,233 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3710, 1.7962, 2.1671, 2.1733, 1.9059, 1.9440, 2.1113, 2.0709], + device='cuda:4'), covar=tensor([0.4514, 0.5163, 0.4265, 0.4860, 0.6251, 0.4746, 0.6663, 0.4057], + device='cuda:4'), in_proj_covar=tensor([0.0238, 0.0239, 0.0255, 0.0260, 0.0256, 0.0230, 0.0275, 0.0232], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:50:05,684 INFO [finetune.py:976] (4/7) Epoch 11, batch 5150, loss[loss=0.1551, simple_loss=0.2266, pruned_loss=0.04181, over 4743.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.256, pruned_loss=0.06263, over 956817.35 frames. ], batch size: 27, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:50:12,137 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62436.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:50:12,675 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.578e+02 2.001e+02 2.432e+02 3.455e+02, threshold=4.003e+02, percent-clipped=0.0 +2023-03-26 13:50:25,614 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1842, 1.9120, 1.5168, 0.6298, 1.6987, 1.8022, 1.6020, 1.8331], + device='cuda:4'), covar=tensor([0.0899, 0.0803, 0.1391, 0.1862, 0.1300, 0.2395, 0.2299, 0.0839], + device='cuda:4'), in_proj_covar=tensor([0.0166, 0.0198, 0.0201, 0.0185, 0.0215, 0.0207, 0.0222, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:50:26,783 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62447.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:50:30,454 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62452.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:50:41,649 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9150, 1.9283, 1.3752, 2.0685, 1.8987, 1.6871, 2.6223, 1.9636], + device='cuda:4'), covar=tensor([0.1394, 0.2218, 0.3622, 0.3115, 0.2776, 0.1729, 0.2825, 0.2032], + device='cuda:4'), in_proj_covar=tensor([0.0175, 0.0187, 0.0232, 0.0254, 0.0240, 0.0197, 0.0212, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:50:53,287 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 13:50:55,269 INFO [finetune.py:976] (4/7) Epoch 11, batch 5200, loss[loss=0.2072, simple_loss=0.2767, pruned_loss=0.06886, over 4937.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2603, pruned_loss=0.0642, over 955510.24 frames. ], batch size: 33, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:50:56,961 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62478.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:51:36,854 INFO [finetune.py:976] (4/7) Epoch 11, batch 5250, loss[loss=0.1809, simple_loss=0.2524, pruned_loss=0.05472, over 4170.00 frames. ], tot_loss[loss=0.1967, simple_loss=0.2628, pruned_loss=0.06525, over 955828.53 frames. ], batch size: 65, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:51:46,762 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7476, 1.3679, 0.7766, 1.5719, 2.1898, 1.4393, 1.6164, 1.6931], + device='cuda:4'), covar=tensor([0.1490, 0.2138, 0.2183, 0.1319, 0.1888, 0.2022, 0.1473, 0.1926], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0093, 0.0119, 0.0094, 0.0099, 0.0090], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 13:51:54,382 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.247e+02 1.618e+02 1.949e+02 2.406e+02 7.235e+02, threshold=3.897e+02, percent-clipped=3.0 +2023-03-26 13:52:03,525 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62545.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:52:05,827 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.4947, 1.5798, 1.5655, 0.9652, 1.5974, 1.8422, 1.8245, 1.3162], + device='cuda:4'), covar=tensor([0.0922, 0.0590, 0.0521, 0.0557, 0.0465, 0.0588, 0.0316, 0.0761], + device='cuda:4'), in_proj_covar=tensor([0.0127, 0.0153, 0.0122, 0.0132, 0.0130, 0.0125, 0.0143, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.3980e-05, 1.1177e-04, 8.7680e-05, 9.5145e-05, 9.2611e-05, 9.1145e-05, + 1.0453e-04, 1.0687e-04], device='cuda:4') +2023-03-26 13:52:19,344 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 13:52:19,512 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62570.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:52:23,690 INFO [finetune.py:976] (4/7) Epoch 11, batch 5300, loss[loss=0.2028, simple_loss=0.2604, pruned_loss=0.07267, over 4885.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2644, pruned_loss=0.06616, over 955437.47 frames. ], batch size: 32, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:52:29,601 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62585.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:52:44,283 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62606.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:52:52,191 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62618.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:52:57,602 INFO [finetune.py:976] (4/7) Epoch 11, batch 5350, loss[loss=0.1864, simple_loss=0.2573, pruned_loss=0.05779, over 4842.00 frames. ], tot_loss[loss=0.1976, simple_loss=0.2638, pruned_loss=0.06565, over 952913.55 frames. ], batch size: 47, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:52:58,903 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62629.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:53:04,200 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.504e+02 1.845e+02 2.238e+02 3.589e+02, threshold=3.690e+02, percent-clipped=0.0 +2023-03-26 13:53:10,789 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62646.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:53:24,280 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62666.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:53:30,289 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6843, 1.6288, 1.5589, 0.9381, 1.6812, 1.8570, 1.8420, 1.4556], + device='cuda:4'), covar=tensor([0.0965, 0.0646, 0.0549, 0.0570, 0.0476, 0.0544, 0.0363, 0.0658], + device='cuda:4'), in_proj_covar=tensor([0.0128, 0.0154, 0.0123, 0.0133, 0.0131, 0.0127, 0.0144, 0.0148], + device='cuda:4'), out_proj_covar=tensor([9.4700e-05, 1.1257e-04, 8.8517e-05, 9.6059e-05, 9.3316e-05, 9.2095e-05, + 1.0519e-04, 1.0774e-04], device='cuda:4') +2023-03-26 13:53:30,763 INFO [finetune.py:976] (4/7) Epoch 11, batch 5400, loss[loss=0.1573, simple_loss=0.2351, pruned_loss=0.03971, over 4905.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2606, pruned_loss=0.06461, over 953629.63 frames. ], batch size: 46, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:53:43,828 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9324, 1.4085, 1.9793, 1.8518, 1.6634, 1.6276, 1.8710, 1.8198], + device='cuda:4'), covar=tensor([0.4163, 0.4529, 0.3501, 0.4179, 0.5167, 0.4203, 0.5001, 0.3475], + device='cuda:4'), in_proj_covar=tensor([0.0238, 0.0240, 0.0256, 0.0261, 0.0257, 0.0231, 0.0276, 0.0234], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:54:04,661 INFO [finetune.py:976] (4/7) Epoch 11, batch 5450, loss[loss=0.2029, simple_loss=0.2693, pruned_loss=0.06826, over 4817.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2569, pruned_loss=0.06289, over 955121.19 frames. ], batch size: 38, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:54:04,793 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62727.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:54:10,765 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.513e+01 1.463e+02 1.876e+02 2.335e+02 4.427e+02, threshold=3.751e+02, percent-clipped=2.0 +2023-03-26 13:54:17,807 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62747.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:54:18,451 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5133, 1.3634, 1.9489, 3.2108, 2.1407, 2.1907, 0.9343, 2.5603], + device='cuda:4'), covar=tensor([0.1900, 0.1644, 0.1475, 0.0715, 0.0906, 0.1511, 0.2098, 0.0593], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0115, 0.0133, 0.0163, 0.0100, 0.0138, 0.0125, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 13:54:36,191 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62773.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:54:36,571 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 13:54:38,554 INFO [finetune.py:976] (4/7) Epoch 11, batch 5500, loss[loss=0.1336, simple_loss=0.2067, pruned_loss=0.03026, over 4930.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.254, pruned_loss=0.06156, over 956466.26 frames. ], batch size: 33, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:54:39,230 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62778.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:54:48,512 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.20 vs. limit=5.0 +2023-03-26 13:54:59,365 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.67 vs. limit=5.0 +2023-03-26 13:55:12,351 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62826.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:55:12,914 INFO [finetune.py:976] (4/7) Epoch 11, batch 5550, loss[loss=0.1958, simple_loss=0.278, pruned_loss=0.05678, over 4934.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.255, pruned_loss=0.06182, over 954822.63 frames. ], batch size: 38, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:55:17,988 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62834.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:55:19,877 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.162e+02 1.580e+02 1.841e+02 2.336e+02 5.980e+02, threshold=3.683e+02, percent-clipped=6.0 +2023-03-26 13:56:07,689 INFO [finetune.py:976] (4/7) Epoch 11, batch 5600, loss[loss=0.1946, simple_loss=0.2691, pruned_loss=0.0601, over 4806.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2601, pruned_loss=0.06357, over 954730.45 frames. ], batch size: 45, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:56:22,193 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62901.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:56:33,610 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-03-26 13:56:34,776 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 13:56:37,252 INFO [finetune.py:976] (4/7) Epoch 11, batch 5650, loss[loss=0.1894, simple_loss=0.2672, pruned_loss=0.05578, over 4827.00 frames. ], tot_loss[loss=0.1958, simple_loss=0.2634, pruned_loss=0.0641, over 955439.50 frames. ], batch size: 47, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:56:38,489 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62929.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:56:47,838 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 13:56:48,739 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.326e+01 1.606e+02 1.910e+02 2.279e+02 4.497e+02, threshold=3.820e+02, percent-clipped=2.0 +2023-03-26 13:56:51,137 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62941.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:57:20,928 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.96 vs. limit=5.0 +2023-03-26 13:57:23,399 INFO [finetune.py:976] (4/7) Epoch 11, batch 5700, loss[loss=0.1372, simple_loss=0.1967, pruned_loss=0.03879, over 3971.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2595, pruned_loss=0.06381, over 936173.42 frames. ], batch size: 17, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:57:23,433 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=62977.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:57:54,983 INFO [finetune.py:976] (4/7) Epoch 12, batch 0, loss[loss=0.183, simple_loss=0.2451, pruned_loss=0.06047, over 4357.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2451, pruned_loss=0.06047, over 4357.00 frames. ], batch size: 19, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:57:54,983 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 13:58:04,060 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1790, 1.8864, 1.7693, 1.7746, 1.8987, 1.8864, 1.8478, 2.5521], + device='cuda:4'), covar=tensor([0.4589, 0.5334, 0.3855, 0.4505, 0.4468, 0.2659, 0.4516, 0.1930], + device='cuda:4'), in_proj_covar=tensor([0.0284, 0.0260, 0.0222, 0.0275, 0.0242, 0.0209, 0.0245, 0.0216], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:58:11,587 INFO [finetune.py:1010] (4/7) Epoch 12, validation: loss=0.16, simple_loss=0.2305, pruned_loss=0.04472, over 2265189.00 frames. +2023-03-26 13:58:11,588 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 13:58:19,011 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9511, 1.9365, 2.0225, 1.4587, 1.8948, 2.0780, 2.0545, 1.6285], + device='cuda:4'), covar=tensor([0.0547, 0.0536, 0.0647, 0.0853, 0.0990, 0.0566, 0.0531, 0.1055], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0134, 0.0141, 0.0125, 0.0121, 0.0143, 0.0143, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:58:22,060 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63022.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:58:37,035 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.590e+02 1.966e+02 2.351e+02 4.424e+02, threshold=3.931e+02, percent-clipped=2.0 +2023-03-26 13:58:37,202 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2450, 2.1402, 1.8035, 2.1897, 2.0851, 2.0643, 2.0252, 3.0387], + device='cuda:4'), covar=tensor([0.4124, 0.5663, 0.3744, 0.5304, 0.5089, 0.2632, 0.4939, 0.1667], + device='cuda:4'), in_proj_covar=tensor([0.0284, 0.0259, 0.0222, 0.0275, 0.0242, 0.0208, 0.0245, 0.0215], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:58:48,209 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 13:58:49,958 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63047.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:59:00,708 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 13:59:00,859 INFO [finetune.py:976] (4/7) Epoch 12, batch 50, loss[loss=0.206, simple_loss=0.2615, pruned_loss=0.07528, over 4859.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2643, pruned_loss=0.06279, over 217578.18 frames. ], batch size: 34, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 13:59:10,741 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1730, 1.8799, 2.1637, 1.5086, 2.1961, 2.2507, 2.2368, 1.4482], + device='cuda:4'), covar=tensor([0.0619, 0.0819, 0.0716, 0.0998, 0.0602, 0.0713, 0.0704, 0.1679], + device='cuda:4'), in_proj_covar=tensor([0.0135, 0.0134, 0.0141, 0.0125, 0.0121, 0.0144, 0.0143, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 13:59:42,648 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63095.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 13:59:54,682 INFO [finetune.py:976] (4/7) Epoch 12, batch 100, loss[loss=0.1571, simple_loss=0.2219, pruned_loss=0.04617, over 4920.00 frames. ], tot_loss[loss=0.1935, simple_loss=0.2588, pruned_loss=0.06408, over 381690.44 frames. ], batch size: 33, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 14:00:15,391 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63129.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:00:21,134 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.723e+02 1.978e+02 2.544e+02 5.107e+02, threshold=3.957e+02, percent-clipped=1.0 +2023-03-26 14:00:50,145 INFO [finetune.py:976] (4/7) Epoch 12, batch 150, loss[loss=0.1802, simple_loss=0.2371, pruned_loss=0.06169, over 4760.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.2523, pruned_loss=0.06138, over 511089.98 frames. ], batch size: 27, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 14:01:04,893 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6657, 2.3065, 2.9044, 1.8157, 2.6270, 2.7702, 2.0471, 3.0048], + device='cuda:4'), covar=tensor([0.1283, 0.1867, 0.1492, 0.2129, 0.0861, 0.1452, 0.2656, 0.0691], + device='cuda:4'), in_proj_covar=tensor([0.0198, 0.0207, 0.0194, 0.0192, 0.0179, 0.0215, 0.0219, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:01:44,260 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-26 14:01:47,576 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63201.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:01:56,084 INFO [finetune.py:976] (4/7) Epoch 12, batch 200, loss[loss=0.1826, simple_loss=0.2477, pruned_loss=0.05877, over 4788.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2519, pruned_loss=0.06171, over 607646.60 frames. ], batch size: 29, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 14:02:17,484 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63221.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:02:32,834 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.551e+02 1.870e+02 2.223e+02 3.918e+02, threshold=3.740e+02, percent-clipped=0.0 +2023-03-26 14:02:41,482 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63241.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:02:46,789 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63249.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:02:50,811 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5116, 1.4364, 1.9476, 1.9450, 1.5837, 3.5185, 1.3539, 1.5280], + device='cuda:4'), covar=tensor([0.0944, 0.1904, 0.1128, 0.0911, 0.1678, 0.0211, 0.1575, 0.1872], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0075, 0.0078, 0.0092, 0.0081, 0.0085, 0.0078], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 14:02:51,342 INFO [finetune.py:976] (4/7) Epoch 12, batch 250, loss[loss=0.2128, simple_loss=0.2859, pruned_loss=0.06979, over 4900.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.258, pruned_loss=0.06376, over 685594.92 frames. ], batch size: 35, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 14:02:55,498 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8428, 3.9191, 3.7711, 1.9521, 4.0859, 3.0001, 0.9389, 2.7957], + device='cuda:4'), covar=tensor([0.2301, 0.1821, 0.1442, 0.3195, 0.0891, 0.0919, 0.4176, 0.1348], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0175, 0.0160, 0.0129, 0.0156, 0.0122, 0.0147, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 14:03:02,677 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5584, 1.4128, 1.9640, 3.3423, 2.2618, 2.2955, 0.7008, 2.6716], + device='cuda:4'), covar=tensor([0.1741, 0.1574, 0.1509, 0.0619, 0.0820, 0.1572, 0.2135, 0.0516], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0115, 0.0133, 0.0163, 0.0100, 0.0137, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 14:03:08,678 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63282.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:03:13,377 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63289.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:03:23,973 INFO [finetune.py:976] (4/7) Epoch 12, batch 300, loss[loss=0.1926, simple_loss=0.2697, pruned_loss=0.05781, over 4759.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2621, pruned_loss=0.0647, over 745307.41 frames. ], batch size: 54, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 14:03:40,236 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63322.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:03:51,186 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.663e+02 2.076e+02 2.406e+02 5.777e+02, threshold=4.151e+02, percent-clipped=4.0 +2023-03-26 14:03:55,053 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 14:04:08,572 INFO [finetune.py:976] (4/7) Epoch 12, batch 350, loss[loss=0.2265, simple_loss=0.2927, pruned_loss=0.08019, over 4791.00 frames. ], tot_loss[loss=0.195, simple_loss=0.262, pruned_loss=0.06397, over 793857.65 frames. ], batch size: 45, lr: 3.67e-03, grad_scale: 16.0 +2023-03-26 14:04:27,605 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63370.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:04:59,619 INFO [finetune.py:976] (4/7) Epoch 12, batch 400, loss[loss=0.2094, simple_loss=0.2756, pruned_loss=0.07163, over 4865.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2638, pruned_loss=0.0647, over 829807.52 frames. ], batch size: 34, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:05:02,021 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6462, 1.5003, 2.0606, 1.9448, 1.7915, 4.2021, 1.4759, 1.7048], + device='cuda:4'), covar=tensor([0.0961, 0.1857, 0.1280, 0.0996, 0.1621, 0.0178, 0.1584, 0.1805], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0082, 0.0075, 0.0078, 0.0092, 0.0082, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 14:05:08,490 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8176, 1.2811, 1.8976, 1.7170, 1.5651, 1.5244, 1.6754, 1.7216], + device='cuda:4'), covar=tensor([0.3837, 0.4507, 0.3427, 0.4139, 0.4987, 0.3955, 0.4767, 0.3399], + device='cuda:4'), in_proj_covar=tensor([0.0239, 0.0239, 0.0256, 0.0261, 0.0258, 0.0232, 0.0276, 0.0234], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:05:10,720 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63420.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:05:11,920 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63422.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:05:16,645 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63429.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:05:21,324 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.052e+02 1.591e+02 1.854e+02 2.332e+02 4.296e+02, threshold=3.709e+02, percent-clipped=1.0 +2023-03-26 14:05:38,141 INFO [finetune.py:976] (4/7) Epoch 12, batch 450, loss[loss=0.1989, simple_loss=0.265, pruned_loss=0.06638, over 4809.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2615, pruned_loss=0.06402, over 855993.13 frames. ], batch size: 25, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:05:57,264 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63477.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:05:59,774 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63481.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:06:00,971 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63483.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:06:08,821 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7237, 1.6502, 2.1345, 2.1021, 2.0553, 4.3244, 1.5389, 1.9313], + device='cuda:4'), covar=tensor([0.1085, 0.2001, 0.1225, 0.1111, 0.1621, 0.0223, 0.1779, 0.1901], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0082, 0.0075, 0.0078, 0.0092, 0.0081, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 14:06:15,170 INFO [finetune.py:976] (4/7) Epoch 12, batch 500, loss[loss=0.1484, simple_loss=0.2119, pruned_loss=0.04245, over 4851.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2596, pruned_loss=0.06329, over 878333.74 frames. ], batch size: 44, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:06:37,050 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.336e+01 1.553e+02 1.855e+02 2.331e+02 4.193e+02, threshold=3.711e+02, percent-clipped=1.0 +2023-03-26 14:06:39,609 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8163, 1.0530, 1.8491, 1.7187, 1.5597, 1.5087, 1.6096, 1.6400], + device='cuda:4'), covar=tensor([0.3625, 0.3865, 0.3257, 0.3455, 0.4432, 0.3386, 0.4161, 0.3211], + device='cuda:4'), in_proj_covar=tensor([0.0239, 0.0240, 0.0256, 0.0261, 0.0259, 0.0232, 0.0276, 0.0235], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:06:48,874 INFO [finetune.py:976] (4/7) Epoch 12, batch 550, loss[loss=0.1979, simple_loss=0.2613, pruned_loss=0.06719, over 4938.00 frames. ], tot_loss[loss=0.1922, simple_loss=0.2576, pruned_loss=0.0634, over 894993.80 frames. ], batch size: 33, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:06:58,442 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63569.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:07:03,808 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63577.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:07:10,268 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63586.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:07:22,328 INFO [finetune.py:976] (4/7) Epoch 12, batch 600, loss[loss=0.1761, simple_loss=0.2449, pruned_loss=0.05361, over 4781.00 frames. ], tot_loss[loss=0.1922, simple_loss=0.2578, pruned_loss=0.06328, over 909806.55 frames. ], batch size: 26, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:07:40,201 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63630.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:07:44,851 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.064e+02 1.685e+02 2.017e+02 2.531e+02 3.696e+02, threshold=4.034e+02, percent-clipped=0.0 +2023-03-26 14:07:51,106 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63647.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:07:56,393 INFO [finetune.py:976] (4/7) Epoch 12, batch 650, loss[loss=0.235, simple_loss=0.3007, pruned_loss=0.08464, over 4799.00 frames. ], tot_loss[loss=0.1939, simple_loss=0.2603, pruned_loss=0.06373, over 919891.98 frames. ], batch size: 41, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:08:08,335 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0779, 0.9184, 1.0076, 0.3756, 0.8279, 1.1208, 1.1903, 0.9783], + device='cuda:4'), covar=tensor([0.0991, 0.0647, 0.0559, 0.0666, 0.0592, 0.0684, 0.0411, 0.0756], + device='cuda:4'), in_proj_covar=tensor([0.0127, 0.0153, 0.0122, 0.0132, 0.0130, 0.0125, 0.0143, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.3589e-05, 1.1186e-04, 8.7735e-05, 9.5516e-05, 9.2192e-05, 9.1126e-05, + 1.0417e-04, 1.0665e-04], device='cuda:4') +2023-03-26 14:08:29,865 INFO [finetune.py:976] (4/7) Epoch 12, batch 700, loss[loss=0.2372, simple_loss=0.2821, pruned_loss=0.09617, over 4831.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2614, pruned_loss=0.06363, over 926503.97 frames. ], batch size: 49, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:08:59,839 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.754e+02 2.049e+02 2.499e+02 4.974e+02, threshold=4.098e+02, percent-clipped=3.0 +2023-03-26 14:09:05,236 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.5592, 3.8818, 4.1352, 4.3256, 4.2920, 4.0333, 4.6769, 1.6177], + device='cuda:4'), covar=tensor([0.0804, 0.0901, 0.0721, 0.0947, 0.1279, 0.1516, 0.0643, 0.5164], + device='cuda:4'), in_proj_covar=tensor([0.0348, 0.0244, 0.0276, 0.0290, 0.0330, 0.0283, 0.0300, 0.0294], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:09:11,206 INFO [finetune.py:976] (4/7) Epoch 12, batch 750, loss[loss=0.2272, simple_loss=0.292, pruned_loss=0.0812, over 4798.00 frames. ], tot_loss[loss=0.1963, simple_loss=0.2636, pruned_loss=0.06448, over 933584.88 frames. ], batch size: 51, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:09:25,545 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63776.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:09:26,759 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63778.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:09:54,801 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.58 vs. limit=5.0 +2023-03-26 14:09:56,448 INFO [finetune.py:976] (4/7) Epoch 12, batch 800, loss[loss=0.196, simple_loss=0.2607, pruned_loss=0.06572, over 4900.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.2624, pruned_loss=0.06387, over 937306.63 frames. ], batch size: 37, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:10:04,890 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63810.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:10:25,522 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5997, 1.5485, 1.4671, 1.5933, 1.2231, 3.4754, 1.3936, 1.9655], + device='cuda:4'), covar=tensor([0.3222, 0.2390, 0.2081, 0.2286, 0.1768, 0.0174, 0.2569, 0.1198], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0115, 0.0119, 0.0123, 0.0115, 0.0098, 0.0098, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 14:10:26,005 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.587e+02 1.868e+02 2.134e+02 3.136e+02, threshold=3.736e+02, percent-clipped=1.0 +2023-03-26 14:10:32,479 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63845.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:10:34,913 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7137, 1.5316, 2.2708, 3.6073, 2.5418, 2.4996, 1.0531, 2.8314], + device='cuda:4'), covar=tensor([0.1840, 0.1484, 0.1406, 0.0554, 0.0791, 0.1423, 0.2043, 0.0512], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0116, 0.0134, 0.0164, 0.0101, 0.0138, 0.0126, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 14:10:38,490 INFO [finetune.py:976] (4/7) Epoch 12, batch 850, loss[loss=0.2267, simple_loss=0.2847, pruned_loss=0.08432, over 4806.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2601, pruned_loss=0.06353, over 939610.57 frames. ], batch size: 45, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:10:41,027 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2321, 2.1449, 1.6911, 2.0230, 2.2433, 1.9099, 2.5018, 2.1819], + device='cuda:4'), covar=tensor([0.1382, 0.2397, 0.3290, 0.2977, 0.2577, 0.1690, 0.3388, 0.2142], + device='cuda:4'), in_proj_covar=tensor([0.0177, 0.0188, 0.0234, 0.0255, 0.0241, 0.0198, 0.0212, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:10:51,325 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63871.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:10:59,629 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63877.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:11:03,726 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2397, 1.4222, 1.4130, 1.5204, 1.5173, 2.9125, 1.3095, 1.6283], + device='cuda:4'), covar=tensor([0.1028, 0.1716, 0.1127, 0.0964, 0.1528, 0.0307, 0.1422, 0.1563], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0082, 0.0075, 0.0078, 0.0092, 0.0082, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 14:11:22,733 INFO [finetune.py:976] (4/7) Epoch 12, batch 900, loss[loss=0.1661, simple_loss=0.2416, pruned_loss=0.04533, over 4786.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2573, pruned_loss=0.06269, over 944263.41 frames. ], batch size: 29, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:11:23,448 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63906.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:11:29,447 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63916.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:11:35,894 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=63925.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:11:35,903 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63925.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:11:44,045 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.021e+02 1.611e+02 1.873e+02 2.372e+02 4.297e+02, threshold=3.747e+02, percent-clipped=2.0 +2023-03-26 14:11:47,185 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63942.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:11:56,455 INFO [finetune.py:976] (4/7) Epoch 12, batch 950, loss[loss=0.1863, simple_loss=0.2546, pruned_loss=0.05896, over 4932.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2559, pruned_loss=0.06264, over 947992.41 frames. ], batch size: 33, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:12:03,844 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2772, 3.7001, 3.8966, 4.0917, 4.0284, 3.7411, 4.3511, 1.3834], + device='cuda:4'), covar=tensor([0.0783, 0.0777, 0.0838, 0.1017, 0.1143, 0.1681, 0.0686, 0.5306], + device='cuda:4'), in_proj_covar=tensor([0.0346, 0.0242, 0.0274, 0.0288, 0.0329, 0.0280, 0.0299, 0.0292], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:12:09,948 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 14:12:10,885 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63977.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:12:31,104 INFO [finetune.py:976] (4/7) Epoch 12, batch 1000, loss[loss=0.1499, simple_loss=0.2198, pruned_loss=0.03997, over 4743.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2587, pruned_loss=0.06371, over 948280.90 frames. ], batch size: 23, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:12:43,091 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64023.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:12:51,954 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.200e+02 1.649e+02 1.875e+02 2.259e+02 3.443e+02, threshold=3.751e+02, percent-clipped=0.0 +2023-03-26 14:13:04,255 INFO [finetune.py:976] (4/7) Epoch 12, batch 1050, loss[loss=0.1744, simple_loss=0.2367, pruned_loss=0.05605, over 4783.00 frames. ], tot_loss[loss=0.1947, simple_loss=0.2608, pruned_loss=0.06426, over 949321.60 frames. ], batch size: 26, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:13:18,022 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64076.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:13:19,225 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64078.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:13:22,946 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64084.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:13:37,900 INFO [finetune.py:976] (4/7) Epoch 12, batch 1100, loss[loss=0.1944, simple_loss=0.2631, pruned_loss=0.06284, over 4862.00 frames. ], tot_loss[loss=0.1965, simple_loss=0.2624, pruned_loss=0.06531, over 950003.78 frames. ], batch size: 31, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:13:53,905 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64124.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:13:55,107 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64126.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:14:05,897 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.584e+02 1.925e+02 2.329e+02 4.054e+02, threshold=3.850e+02, percent-clipped=2.0 +2023-03-26 14:14:07,257 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7651, 1.6955, 1.5807, 1.6625, 1.2360, 4.2807, 1.7224, 1.9165], + device='cuda:4'), covar=tensor([0.3395, 0.2499, 0.2077, 0.2303, 0.1792, 0.0124, 0.2421, 0.1335], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0115, 0.0119, 0.0123, 0.0115, 0.0098, 0.0098, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 14:14:17,902 INFO [finetune.py:976] (4/7) Epoch 12, batch 1150, loss[loss=0.1663, simple_loss=0.2439, pruned_loss=0.0443, over 4927.00 frames. ], tot_loss[loss=0.1962, simple_loss=0.2627, pruned_loss=0.06481, over 950937.19 frames. ], batch size: 42, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:14:25,686 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64166.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:14:30,415 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6769, 1.5956, 1.5363, 1.5697, 1.0728, 3.0489, 1.1895, 1.5740], + device='cuda:4'), covar=tensor([0.3400, 0.2436, 0.2095, 0.2489, 0.2016, 0.0270, 0.2659, 0.1301], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0115, 0.0120, 0.0123, 0.0115, 0.0098, 0.0098, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 14:14:40,411 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.0900, 4.3938, 4.7044, 4.8722, 4.7898, 4.5619, 5.1869, 1.5455], + device='cuda:4'), covar=tensor([0.0760, 0.0847, 0.0771, 0.0880, 0.1192, 0.1428, 0.0565, 0.5837], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0242, 0.0275, 0.0290, 0.0330, 0.0282, 0.0300, 0.0294], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:14:48,872 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64201.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:14:56,320 INFO [finetune.py:976] (4/7) Epoch 12, batch 1200, loss[loss=0.2049, simple_loss=0.2684, pruned_loss=0.07069, over 4900.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2608, pruned_loss=0.0638, over 952940.37 frames. ], batch size: 46, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:15:14,938 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64225.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:15:24,854 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64232.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:15:31,342 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.004e+02 1.575e+02 1.833e+02 2.193e+02 5.344e+02, threshold=3.667e+02, percent-clipped=2.0 +2023-03-26 14:15:34,438 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64242.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:15:43,193 INFO [finetune.py:976] (4/7) Epoch 12, batch 1250, loss[loss=0.1797, simple_loss=0.2448, pruned_loss=0.05729, over 4830.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2584, pruned_loss=0.06334, over 953489.54 frames. ], batch size: 33, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:15:55,100 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64272.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:15:55,715 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64273.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:16:09,179 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64290.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:16:11,100 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64293.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:16:27,223 INFO [finetune.py:976] (4/7) Epoch 12, batch 1300, loss[loss=0.1763, simple_loss=0.2454, pruned_loss=0.05365, over 4917.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2555, pruned_loss=0.06234, over 955341.62 frames. ], batch size: 37, lr: 3.66e-03, grad_scale: 16.0 +2023-03-26 14:16:42,307 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 14:16:48,479 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.610e+02 1.842e+02 2.244e+02 4.381e+02, threshold=3.684e+02, percent-clipped=1.0 +2023-03-26 14:16:53,422 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64345.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:16:55,230 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2425, 1.9856, 2.4427, 4.2465, 3.0895, 2.9469, 1.1036, 3.5245], + device='cuda:4'), covar=tensor([0.1640, 0.1413, 0.1544, 0.0553, 0.0671, 0.1305, 0.2049, 0.0455], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0116, 0.0135, 0.0165, 0.0101, 0.0138, 0.0127, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 14:16:59,919 INFO [finetune.py:976] (4/7) Epoch 12, batch 1350, loss[loss=0.1832, simple_loss=0.2455, pruned_loss=0.06048, over 4734.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2558, pruned_loss=0.06285, over 956611.30 frames. ], batch size: 23, lr: 3.66e-03, grad_scale: 32.0 +2023-03-26 14:17:16,046 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64379.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:17:33,430 INFO [finetune.py:976] (4/7) Epoch 12, batch 1400, loss[loss=0.2289, simple_loss=0.2918, pruned_loss=0.08299, over 4820.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2593, pruned_loss=0.06397, over 956278.98 frames. ], batch size: 40, lr: 3.66e-03, grad_scale: 32.0 +2023-03-26 14:17:34,195 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64406.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:17:46,544 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7294, 1.5289, 1.9958, 1.4048, 1.9507, 2.0267, 1.4247, 2.1902], + device='cuda:4'), covar=tensor([0.1337, 0.2172, 0.1400, 0.1905, 0.0781, 0.1417, 0.3019, 0.0765], + device='cuda:4'), in_proj_covar=tensor([0.0196, 0.0205, 0.0195, 0.0191, 0.0179, 0.0214, 0.0218, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:17:54,259 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.198e+02 1.617e+02 1.936e+02 2.295e+02 3.610e+02, threshold=3.872e+02, percent-clipped=0.0 +2023-03-26 14:18:06,658 INFO [finetune.py:976] (4/7) Epoch 12, batch 1450, loss[loss=0.2274, simple_loss=0.2878, pruned_loss=0.08355, over 4837.00 frames. ], tot_loss[loss=0.1956, simple_loss=0.2618, pruned_loss=0.06467, over 954859.77 frames. ], batch size: 30, lr: 3.66e-03, grad_scale: 32.0 +2023-03-26 14:18:13,319 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64465.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:18:13,904 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64466.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:18:37,420 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64501.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:18:38,022 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4490, 1.3480, 1.3879, 1.3463, 0.8256, 2.3525, 0.7614, 1.2207], + device='cuda:4'), covar=tensor([0.3591, 0.2639, 0.2281, 0.2561, 0.2180, 0.0390, 0.2725, 0.1447], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0115, 0.0119, 0.0123, 0.0115, 0.0098, 0.0098, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 14:18:39,730 INFO [finetune.py:976] (4/7) Epoch 12, batch 1500, loss[loss=0.1951, simple_loss=0.2588, pruned_loss=0.06567, over 4923.00 frames. ], tot_loss[loss=0.1974, simple_loss=0.2637, pruned_loss=0.06557, over 954167.91 frames. ], batch size: 33, lr: 3.66e-03, grad_scale: 32.0 +2023-03-26 14:18:46,091 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64514.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:18:54,920 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64526.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:19:01,495 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.738e+02 2.083e+02 2.672e+02 4.064e+02, threshold=4.165e+02, percent-clipped=1.0 +2023-03-26 14:19:15,862 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64549.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:19:19,450 INFO [finetune.py:976] (4/7) Epoch 12, batch 1550, loss[loss=0.1954, simple_loss=0.2534, pruned_loss=0.06865, over 4177.00 frames. ], tot_loss[loss=0.1981, simple_loss=0.2645, pruned_loss=0.06584, over 954623.30 frames. ], batch size: 65, lr: 3.66e-03, grad_scale: 32.0 +2023-03-26 14:19:33,897 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64572.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:19:45,159 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64588.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:19:45,837 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64589.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:19:56,452 INFO [finetune.py:976] (4/7) Epoch 12, batch 1600, loss[loss=0.2026, simple_loss=0.2552, pruned_loss=0.07498, over 4866.00 frames. ], tot_loss[loss=0.1956, simple_loss=0.2617, pruned_loss=0.06481, over 954378.02 frames. ], batch size: 31, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:20:12,897 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64620.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:20:30,238 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 1.633e+02 1.922e+02 2.431e+02 4.177e+02, threshold=3.845e+02, percent-clipped=1.0 +2023-03-26 14:20:41,559 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64648.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:20:43,313 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64650.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:20:49,612 INFO [finetune.py:976] (4/7) Epoch 12, batch 1650, loss[loss=0.1406, simple_loss=0.212, pruned_loss=0.03464, over 4845.00 frames. ], tot_loss[loss=0.1929, simple_loss=0.2585, pruned_loss=0.06365, over 953213.26 frames. ], batch size: 47, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:20:51,525 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7650, 3.8856, 3.6576, 2.0184, 4.0284, 2.9971, 0.8361, 2.7436], + device='cuda:4'), covar=tensor([0.2046, 0.1619, 0.1372, 0.3357, 0.0957, 0.0921, 0.4515, 0.1571], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0173, 0.0158, 0.0128, 0.0155, 0.0121, 0.0146, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 14:21:05,207 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64679.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:21:19,529 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64701.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:21:21,787 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64703.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:21:22,888 INFO [finetune.py:976] (4/7) Epoch 12, batch 1700, loss[loss=0.2242, simple_loss=0.2805, pruned_loss=0.08393, over 4814.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2577, pruned_loss=0.06348, over 953519.37 frames. ], batch size: 41, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:21:27,405 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64709.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:21:36,624 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.45 vs. limit=5.0 +2023-03-26 14:21:46,816 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64727.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:21:53,448 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.591e+02 1.930e+02 2.225e+02 5.420e+02, threshold=3.861e+02, percent-clipped=2.0 +2023-03-26 14:21:58,922 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64745.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:22:03,092 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64752.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:22:05,310 INFO [finetune.py:976] (4/7) Epoch 12, batch 1750, loss[loss=0.1768, simple_loss=0.2416, pruned_loss=0.05602, over 4899.00 frames. ], tot_loss[loss=0.1935, simple_loss=0.2591, pruned_loss=0.06391, over 955192.63 frames. ], batch size: 35, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:22:10,827 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64764.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:22:38,078 INFO [finetune.py:976] (4/7) Epoch 12, batch 1800, loss[loss=0.2128, simple_loss=0.2772, pruned_loss=0.07417, over 4887.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.2628, pruned_loss=0.06546, over 954470.26 frames. ], batch size: 32, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:22:38,791 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64806.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:22:43,527 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64813.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:22:44,778 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0380, 1.7187, 2.0388, 1.9502, 1.7194, 1.7446, 1.9165, 1.8937], + device='cuda:4'), covar=tensor([0.4467, 0.4728, 0.3794, 0.4846, 0.5916, 0.4297, 0.5618, 0.3726], + device='cuda:4'), in_proj_covar=tensor([0.0238, 0.0238, 0.0256, 0.0259, 0.0256, 0.0231, 0.0274, 0.0232], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:22:48,369 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64821.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:22:58,987 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.145e+02 1.622e+02 1.968e+02 2.271e+02 4.247e+02, threshold=3.936e+02, percent-clipped=1.0 +2023-03-26 14:23:11,389 INFO [finetune.py:976] (4/7) Epoch 12, batch 1850, loss[loss=0.1909, simple_loss=0.2367, pruned_loss=0.07254, over 4387.00 frames. ], tot_loss[loss=0.197, simple_loss=0.2632, pruned_loss=0.06541, over 954877.37 frames. ], batch size: 19, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:23:33,507 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64888.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:23:45,141 INFO [finetune.py:976] (4/7) Epoch 12, batch 1900, loss[loss=0.1622, simple_loss=0.233, pruned_loss=0.04572, over 4185.00 frames. ], tot_loss[loss=0.1958, simple_loss=0.2621, pruned_loss=0.06475, over 952916.26 frames. ], batch size: 66, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:23:54,109 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0328, 1.8489, 1.5661, 1.5393, 1.7459, 1.7155, 1.7442, 2.5102], + device='cuda:4'), covar=tensor([0.3947, 0.4160, 0.3417, 0.3858, 0.3709, 0.2475, 0.3867, 0.1703], + device='cuda:4'), in_proj_covar=tensor([0.0284, 0.0258, 0.0223, 0.0275, 0.0242, 0.0210, 0.0246, 0.0216], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:23:56,842 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 14:23:59,627 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4787, 1.4432, 1.5011, 0.8087, 1.5549, 1.5347, 1.4651, 1.3579], + device='cuda:4'), covar=tensor([0.0635, 0.0769, 0.0746, 0.1036, 0.0802, 0.0748, 0.0629, 0.1184], + device='cuda:4'), in_proj_covar=tensor([0.0137, 0.0135, 0.0143, 0.0126, 0.0123, 0.0144, 0.0145, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:24:06,072 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=64936.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:24:06,596 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.024e+02 1.653e+02 1.911e+02 2.364e+02 4.358e+02, threshold=3.822e+02, percent-clipped=3.0 +2023-03-26 14:24:12,016 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64945.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:24:18,921 INFO [finetune.py:976] (4/7) Epoch 12, batch 1950, loss[loss=0.1583, simple_loss=0.2302, pruned_loss=0.04319, over 4820.00 frames. ], tot_loss[loss=0.1938, simple_loss=0.2602, pruned_loss=0.06367, over 955782.32 frames. ], batch size: 41, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:24:40,799 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64976.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:24:58,109 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65001.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:24:59,919 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65004.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:25:00,453 INFO [finetune.py:976] (4/7) Epoch 12, batch 2000, loss[loss=0.1982, simple_loss=0.2606, pruned_loss=0.06788, over 4877.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2572, pruned_loss=0.06271, over 954954.78 frames. ], batch size: 31, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:25:21,642 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.609e+02 1.880e+02 2.233e+02 7.388e+02, threshold=3.760e+02, percent-clipped=1.0 +2023-03-26 14:25:21,787 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65037.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:25:34,388 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65049.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:25:42,380 INFO [finetune.py:976] (4/7) Epoch 12, batch 2050, loss[loss=0.2058, simple_loss=0.2633, pruned_loss=0.07412, over 4900.00 frames. ], tot_loss[loss=0.1879, simple_loss=0.2534, pruned_loss=0.06116, over 957053.18 frames. ], batch size: 35, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:25:42,484 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65055.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:25:45,414 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65059.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:26:19,438 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.5524, 3.9832, 4.1647, 4.3616, 4.2985, 4.0392, 4.6651, 1.3254], + device='cuda:4'), covar=tensor([0.0812, 0.0853, 0.0828, 0.1003, 0.1226, 0.1447, 0.0625, 0.5715], + device='cuda:4'), in_proj_covar=tensor([0.0345, 0.0241, 0.0274, 0.0288, 0.0328, 0.0279, 0.0298, 0.0293], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:26:21,877 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65101.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:26:24,690 INFO [finetune.py:976] (4/7) Epoch 12, batch 2100, loss[loss=0.1824, simple_loss=0.2541, pruned_loss=0.05538, over 4795.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.2525, pruned_loss=0.06091, over 955483.19 frames. ], batch size: 51, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:26:27,106 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65108.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:26:32,485 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65116.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:26:35,491 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65121.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:26:52,594 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.054e+02 1.674e+02 1.985e+02 2.398e+02 5.597e+02, threshold=3.971e+02, percent-clipped=1.0 +2023-03-26 14:27:08,171 INFO [finetune.py:976] (4/7) Epoch 12, batch 2150, loss[loss=0.2099, simple_loss=0.2755, pruned_loss=0.0721, over 4857.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2563, pruned_loss=0.06218, over 956309.74 frames. ], batch size: 49, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:27:13,554 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6949, 2.3610, 2.6621, 2.5514, 2.2890, 2.3324, 2.5681, 2.4474], + device='cuda:4'), covar=tensor([0.3776, 0.4122, 0.3483, 0.4143, 0.4922, 0.3702, 0.5004, 0.3430], + device='cuda:4'), in_proj_covar=tensor([0.0240, 0.0238, 0.0257, 0.0260, 0.0258, 0.0232, 0.0276, 0.0234], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:27:15,642 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 14:27:17,729 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65169.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:27:40,969 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5938, 1.1954, 0.8620, 1.5105, 1.8984, 1.3411, 1.3111, 1.5296], + device='cuda:4'), covar=tensor([0.2009, 0.2873, 0.2557, 0.1643, 0.2648, 0.2916, 0.2047, 0.2723], + device='cuda:4'), in_proj_covar=tensor([0.0088, 0.0095, 0.0112, 0.0091, 0.0119, 0.0094, 0.0098, 0.0090], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 14:27:41,483 INFO [finetune.py:976] (4/7) Epoch 12, batch 2200, loss[loss=0.2129, simple_loss=0.2748, pruned_loss=0.07546, over 4925.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2595, pruned_loss=0.06341, over 955856.14 frames. ], batch size: 42, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:27:52,743 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3450, 1.4481, 1.5311, 0.7627, 1.4388, 1.7013, 1.6818, 1.2921], + device='cuda:4'), covar=tensor([0.1098, 0.0731, 0.0525, 0.0658, 0.0469, 0.0611, 0.0372, 0.0803], + device='cuda:4'), in_proj_covar=tensor([0.0127, 0.0153, 0.0122, 0.0132, 0.0130, 0.0126, 0.0143, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.3464e-05, 1.1191e-04, 8.7526e-05, 9.5086e-05, 9.2678e-05, 9.1602e-05, + 1.0457e-04, 1.0656e-04], device='cuda:4') +2023-03-26 14:28:03,266 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.670e+02 2.055e+02 2.491e+02 4.530e+02, threshold=4.111e+02, percent-clipped=2.0 +2023-03-26 14:28:08,186 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7258, 2.9259, 2.5989, 1.9953, 2.6527, 2.9211, 2.7469, 2.4697], + device='cuda:4'), covar=tensor([0.0456, 0.0455, 0.0602, 0.0810, 0.0554, 0.0565, 0.0574, 0.0856], + device='cuda:4'), in_proj_covar=tensor([0.0135, 0.0134, 0.0141, 0.0124, 0.0122, 0.0142, 0.0143, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:28:08,787 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65245.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:28:12,942 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1933, 2.0798, 1.7499, 2.1476, 2.1543, 1.9120, 2.5100, 2.1590], + device='cuda:4'), covar=tensor([0.1455, 0.2316, 0.3588, 0.2898, 0.2969, 0.1922, 0.3198, 0.2110], + device='cuda:4'), in_proj_covar=tensor([0.0177, 0.0187, 0.0233, 0.0254, 0.0242, 0.0198, 0.0214, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:28:15,259 INFO [finetune.py:976] (4/7) Epoch 12, batch 2250, loss[loss=0.1707, simple_loss=0.2558, pruned_loss=0.0428, over 4897.00 frames. ], tot_loss[loss=0.1941, simple_loss=0.2612, pruned_loss=0.06355, over 958115.87 frames. ], batch size: 43, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:28:16,006 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0422, 1.9611, 1.6644, 1.9832, 2.0567, 1.7702, 2.3608, 2.0342], + device='cuda:4'), covar=tensor([0.1419, 0.2170, 0.3049, 0.2629, 0.2539, 0.1694, 0.3211, 0.1899], + device='cuda:4'), in_proj_covar=tensor([0.0177, 0.0187, 0.0233, 0.0254, 0.0242, 0.0198, 0.0214, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:28:35,381 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3303, 1.2957, 1.2816, 1.2916, 0.9206, 2.0160, 0.8379, 1.2597], + device='cuda:4'), covar=tensor([0.2935, 0.2136, 0.1900, 0.2183, 0.1707, 0.0437, 0.2846, 0.1188], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0116, 0.0120, 0.0124, 0.0116, 0.0098, 0.0098, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 14:28:35,987 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65285.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:28:39,086 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.81 vs. limit=5.0 +2023-03-26 14:28:41,278 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65293.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:28:48,517 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65304.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:28:49,046 INFO [finetune.py:976] (4/7) Epoch 12, batch 2300, loss[loss=0.201, simple_loss=0.2659, pruned_loss=0.06806, over 4742.00 frames. ], tot_loss[loss=0.1941, simple_loss=0.2612, pruned_loss=0.06346, over 955294.77 frames. ], batch size: 54, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:28:52,237 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6786, 1.5987, 1.9355, 1.1580, 1.7239, 1.9153, 1.5215, 2.0623], + device='cuda:4'), covar=tensor([0.1240, 0.2141, 0.1270, 0.1838, 0.0925, 0.1292, 0.2749, 0.0850], + device='cuda:4'), in_proj_covar=tensor([0.0198, 0.0208, 0.0197, 0.0194, 0.0180, 0.0217, 0.0220, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:29:07,489 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65332.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:29:10,445 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.717e+01 1.673e+02 1.949e+02 2.267e+02 6.743e+02, threshold=3.897e+02, percent-clipped=1.0 +2023-03-26 14:29:16,550 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65346.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:29:20,083 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65352.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:29:22,342 INFO [finetune.py:976] (4/7) Epoch 12, batch 2350, loss[loss=0.2202, simple_loss=0.2707, pruned_loss=0.0849, over 4112.00 frames. ], tot_loss[loss=0.1921, simple_loss=0.2587, pruned_loss=0.06272, over 954411.37 frames. ], batch size: 65, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:29:24,813 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65359.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:30:02,654 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65401.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:30:04,998 INFO [finetune.py:976] (4/7) Epoch 12, batch 2400, loss[loss=0.1415, simple_loss=0.21, pruned_loss=0.03656, over 4839.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2549, pruned_loss=0.06117, over 954860.75 frames. ], batch size: 25, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:30:06,755 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65407.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:30:06,816 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0847, 0.8584, 0.9708, 0.1970, 0.8861, 1.1613, 1.1896, 0.9766], + device='cuda:4'), covar=tensor([0.0847, 0.0732, 0.0526, 0.0633, 0.0597, 0.0558, 0.0403, 0.0602], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0152, 0.0121, 0.0131, 0.0130, 0.0126, 0.0142, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.2970e-05, 1.1143e-04, 8.6971e-05, 9.4606e-05, 9.2613e-05, 9.1432e-05, + 1.0393e-04, 1.0595e-04], device='cuda:4') +2023-03-26 14:30:07,399 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65408.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:30:09,187 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65411.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:30:26,324 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.174e+02 1.534e+02 1.885e+02 2.327e+02 5.518e+02, threshold=3.771e+02, percent-clipped=1.0 +2023-03-26 14:30:34,674 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65449.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:30:38,808 INFO [finetune.py:976] (4/7) Epoch 12, batch 2450, loss[loss=0.1836, simple_loss=0.248, pruned_loss=0.0596, over 4905.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2514, pruned_loss=0.05958, over 954765.85 frames. ], batch size: 36, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:30:39,469 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65456.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:31:31,338 INFO [finetune.py:976] (4/7) Epoch 12, batch 2500, loss[loss=0.2364, simple_loss=0.2988, pruned_loss=0.08702, over 4798.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2542, pruned_loss=0.06074, over 952911.56 frames. ], batch size: 51, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:31:49,235 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65532.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:31:52,647 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.473e+01 1.650e+02 2.020e+02 2.341e+02 4.049e+02, threshold=4.040e+02, percent-clipped=1.0 +2023-03-26 14:31:55,332 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-26 14:32:06,906 INFO [finetune.py:976] (4/7) Epoch 12, batch 2550, loss[loss=0.2107, simple_loss=0.2697, pruned_loss=0.07591, over 4769.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2585, pruned_loss=0.06188, over 952917.99 frames. ], batch size: 54, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:32:27,336 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7846, 1.8214, 1.5540, 1.9408, 2.4576, 2.0230, 1.6028, 1.4934], + device='cuda:4'), covar=tensor([0.2319, 0.2005, 0.1939, 0.1658, 0.1730, 0.1097, 0.2373, 0.1999], + device='cuda:4'), in_proj_covar=tensor([0.0240, 0.0208, 0.0210, 0.0189, 0.0242, 0.0183, 0.0213, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:32:40,910 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65593.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:32:43,306 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5565, 2.3005, 2.9801, 1.7203, 2.6233, 2.6464, 2.2156, 2.8884], + device='cuda:4'), covar=tensor([0.1328, 0.1867, 0.1250, 0.2207, 0.0849, 0.1606, 0.2375, 0.0879], + device='cuda:4'), in_proj_covar=tensor([0.0197, 0.0206, 0.0194, 0.0191, 0.0179, 0.0216, 0.0217, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:32:48,714 INFO [finetune.py:976] (4/7) Epoch 12, batch 2600, loss[loss=0.2451, simple_loss=0.3088, pruned_loss=0.09073, over 4812.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2613, pruned_loss=0.06308, over 951408.70 frames. ], batch size: 40, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:32:56,513 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.4676, 3.8577, 4.0637, 4.2882, 4.2326, 3.9258, 4.5288, 1.4244], + device='cuda:4'), covar=tensor([0.0590, 0.0726, 0.0743, 0.0852, 0.0940, 0.1394, 0.0572, 0.5174], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0242, 0.0276, 0.0291, 0.0330, 0.0282, 0.0300, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:33:06,628 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65632.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:33:08,927 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9013, 1.8352, 1.9436, 1.1932, 1.9736, 1.9940, 1.8433, 1.6469], + device='cuda:4'), covar=tensor([0.0574, 0.0781, 0.0728, 0.0973, 0.0710, 0.0758, 0.0767, 0.1173], + device='cuda:4'), in_proj_covar=tensor([0.0136, 0.0135, 0.0142, 0.0125, 0.0123, 0.0143, 0.0145, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:33:10,011 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.524e+01 1.645e+02 2.049e+02 2.494e+02 4.393e+02, threshold=4.097e+02, percent-clipped=1.0 +2023-03-26 14:33:10,740 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6120, 1.5956, 1.6012, 0.8442, 1.6496, 1.9229, 1.9276, 1.4764], + device='cuda:4'), covar=tensor([0.0945, 0.0696, 0.0483, 0.0675, 0.0478, 0.0694, 0.0307, 0.0679], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0153, 0.0122, 0.0132, 0.0130, 0.0126, 0.0143, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.3271e-05, 1.1166e-04, 8.7761e-05, 9.4853e-05, 9.2776e-05, 9.1318e-05, + 1.0414e-04, 1.0645e-04], device='cuda:4') +2023-03-26 14:33:12,533 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65641.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:33:22,368 INFO [finetune.py:976] (4/7) Epoch 12, batch 2650, loss[loss=0.2147, simple_loss=0.2885, pruned_loss=0.07041, over 4872.00 frames. ], tot_loss[loss=0.194, simple_loss=0.2615, pruned_loss=0.06328, over 951053.64 frames. ], batch size: 34, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:33:38,618 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65680.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:33:47,464 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4952, 3.3509, 3.1985, 1.6289, 3.5043, 2.5746, 0.6941, 2.3082], + device='cuda:4'), covar=tensor([0.2409, 0.1874, 0.1739, 0.3157, 0.1190, 0.1110, 0.4360, 0.1494], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0174, 0.0159, 0.0127, 0.0155, 0.0122, 0.0146, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 14:33:55,610 INFO [finetune.py:976] (4/7) Epoch 12, batch 2700, loss[loss=0.191, simple_loss=0.2423, pruned_loss=0.0699, over 4022.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2603, pruned_loss=0.06293, over 948581.32 frames. ], batch size: 17, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:33:59,758 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65711.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:34:17,010 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.115e+02 1.547e+02 1.884e+02 2.200e+02 3.210e+02, threshold=3.769e+02, percent-clipped=0.0 +2023-03-26 14:34:30,267 INFO [finetune.py:976] (4/7) Epoch 12, batch 2750, loss[loss=0.1665, simple_loss=0.2239, pruned_loss=0.05459, over 4796.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.258, pruned_loss=0.06263, over 949022.54 frames. ], batch size: 25, lr: 3.65e-03, grad_scale: 32.0 +2023-03-26 14:34:38,469 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65759.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:35:30,778 INFO [finetune.py:976] (4/7) Epoch 12, batch 2800, loss[loss=0.1714, simple_loss=0.2382, pruned_loss=0.05227, over 4912.00 frames. ], tot_loss[loss=0.1902, simple_loss=0.2558, pruned_loss=0.06234, over 951391.76 frames. ], batch size: 36, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:35:52,231 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.914e+01 1.578e+02 1.887e+02 2.176e+02 5.167e+02, threshold=3.774e+02, percent-clipped=1.0 +2023-03-26 14:35:52,956 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65838.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:35:57,679 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8232, 3.9583, 3.8100, 1.9407, 4.0992, 3.0108, 0.8373, 2.8259], + device='cuda:4'), covar=tensor([0.2393, 0.2145, 0.1561, 0.3456, 0.0885, 0.1070, 0.4837, 0.1579], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0174, 0.0159, 0.0127, 0.0156, 0.0121, 0.0146, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 14:36:04,206 INFO [finetune.py:976] (4/7) Epoch 12, batch 2850, loss[loss=0.2387, simple_loss=0.2992, pruned_loss=0.08913, over 4894.00 frames. ], tot_loss[loss=0.1879, simple_loss=0.2539, pruned_loss=0.06098, over 954112.16 frames. ], batch size: 35, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:36:34,054 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 14:36:36,965 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65888.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 14:36:45,243 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65899.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:36:48,786 INFO [finetune.py:976] (4/7) Epoch 12, batch 2900, loss[loss=0.1486, simple_loss=0.22, pruned_loss=0.03861, over 4758.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.257, pruned_loss=0.06206, over 952706.26 frames. ], batch size: 27, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:37:10,264 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.527e+02 1.842e+02 2.377e+02 4.547e+02, threshold=3.684e+02, percent-clipped=3.0 +2023-03-26 14:37:12,782 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65941.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:37:20,122 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-26 14:37:27,696 INFO [finetune.py:976] (4/7) Epoch 12, batch 2950, loss[loss=0.1952, simple_loss=0.2672, pruned_loss=0.06165, over 4799.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.2604, pruned_loss=0.06326, over 950976.02 frames. ], batch size: 51, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:37:43,488 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.7913, 3.3030, 3.4839, 3.6275, 3.5492, 3.4066, 3.8668, 1.1313], + device='cuda:4'), covar=tensor([0.0914, 0.0932, 0.0897, 0.1193, 0.1414, 0.1550, 0.0892, 0.5744], + device='cuda:4'), in_proj_covar=tensor([0.0348, 0.0242, 0.0275, 0.0292, 0.0329, 0.0282, 0.0300, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:38:02,423 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=65989.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:38:20,434 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-26 14:38:26,404 INFO [finetune.py:976] (4/7) Epoch 12, batch 3000, loss[loss=0.1952, simple_loss=0.2728, pruned_loss=0.05884, over 4811.00 frames. ], tot_loss[loss=0.1946, simple_loss=0.2615, pruned_loss=0.06387, over 951391.50 frames. ], batch size: 40, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:38:26,404 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 14:38:28,779 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8759, 1.7654, 1.7540, 1.7335, 1.1947, 3.0163, 1.2619, 1.7495], + device='cuda:4'), covar=tensor([0.2959, 0.2058, 0.1802, 0.2150, 0.1618, 0.0269, 0.2256, 0.1096], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0115, 0.0119, 0.0123, 0.0115, 0.0098, 0.0098, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 14:38:32,603 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8516, 1.3824, 1.0164, 1.6209, 2.0479, 1.3172, 1.5793, 1.6908], + device='cuda:4'), covar=tensor([0.1211, 0.1712, 0.1733, 0.1055, 0.1820, 0.1906, 0.1228, 0.1722], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0096, 0.0113, 0.0092, 0.0120, 0.0094, 0.0100, 0.0091], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 14:38:37,073 INFO [finetune.py:1010] (4/7) Epoch 12, validation: loss=0.1571, simple_loss=0.2281, pruned_loss=0.04309, over 2265189.00 frames. +2023-03-26 14:38:37,074 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 14:38:49,188 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 14:38:58,512 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.619e+02 1.943e+02 2.343e+02 4.325e+02, threshold=3.886e+02, percent-clipped=3.0 +2023-03-26 14:39:21,465 INFO [finetune.py:976] (4/7) Epoch 12, batch 3050, loss[loss=0.2779, simple_loss=0.3254, pruned_loss=0.1152, over 4849.00 frames. ], tot_loss[loss=0.1958, simple_loss=0.263, pruned_loss=0.06434, over 951809.52 frames. ], batch size: 44, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:40:06,364 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0983, 1.9009, 1.6032, 1.7482, 1.8582, 1.7975, 1.8592, 2.5684], + device='cuda:4'), covar=tensor([0.3994, 0.4694, 0.3408, 0.4153, 0.4154, 0.2560, 0.4050, 0.1720], + device='cuda:4'), in_proj_covar=tensor([0.0282, 0.0258, 0.0222, 0.0274, 0.0242, 0.0209, 0.0245, 0.0216], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:40:36,353 INFO [finetune.py:976] (4/7) Epoch 12, batch 3100, loss[loss=0.1677, simple_loss=0.245, pruned_loss=0.04524, over 4820.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2612, pruned_loss=0.06374, over 949869.01 frames. ], batch size: 41, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:41:19,790 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.628e+02 1.972e+02 2.398e+02 4.316e+02, threshold=3.945e+02, percent-clipped=3.0 +2023-03-26 14:41:22,271 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6479, 1.5020, 1.4837, 0.8339, 1.5613, 1.7478, 1.7099, 1.4575], + device='cuda:4'), covar=tensor([0.0704, 0.0535, 0.0523, 0.0501, 0.0452, 0.0495, 0.0293, 0.0587], + device='cuda:4'), in_proj_covar=tensor([0.0127, 0.0154, 0.0123, 0.0132, 0.0131, 0.0127, 0.0145, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.3960e-05, 1.1254e-04, 8.8647e-05, 9.5535e-05, 9.3301e-05, 9.2046e-05, + 1.0549e-04, 1.0728e-04], device='cuda:4') +2023-03-26 14:41:26,917 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2360, 1.9339, 2.8101, 1.6738, 2.4603, 2.5343, 1.8521, 2.6745], + device='cuda:4'), covar=tensor([0.1238, 0.1714, 0.1315, 0.2079, 0.0631, 0.1392, 0.2378, 0.0748], + device='cuda:4'), in_proj_covar=tensor([0.0196, 0.0205, 0.0193, 0.0190, 0.0178, 0.0214, 0.0217, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:41:27,552 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66144.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:41:40,147 INFO [finetune.py:976] (4/7) Epoch 12, batch 3150, loss[loss=0.2086, simple_loss=0.2491, pruned_loss=0.08406, over 4021.00 frames. ], tot_loss[loss=0.1922, simple_loss=0.2586, pruned_loss=0.0629, over 952180.41 frames. ], batch size: 17, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:41:43,358 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66160.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:42:04,561 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6011, 2.3600, 2.0980, 2.6146, 2.5144, 2.1986, 3.0504, 2.5402], + device='cuda:4'), covar=tensor([0.1475, 0.2834, 0.3314, 0.3165, 0.2791, 0.1902, 0.3246, 0.2253], + device='cuda:4'), in_proj_covar=tensor([0.0177, 0.0187, 0.0232, 0.0254, 0.0241, 0.0197, 0.0213, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:42:24,559 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66188.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 14:42:33,442 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66194.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 14:42:35,555 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 14:42:40,455 INFO [finetune.py:976] (4/7) Epoch 12, batch 3200, loss[loss=0.1877, simple_loss=0.2503, pruned_loss=0.06257, over 4783.00 frames. ], tot_loss[loss=0.189, simple_loss=0.2551, pruned_loss=0.06141, over 954544.36 frames. ], batch size: 26, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:42:40,569 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66205.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:42:51,199 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66221.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 14:43:01,313 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66236.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:43:01,824 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.059e+02 1.602e+02 1.878e+02 2.419e+02 4.134e+02, threshold=3.755e+02, percent-clipped=1.0 +2023-03-26 14:43:02,317 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 14:43:14,214 INFO [finetune.py:976] (4/7) Epoch 12, batch 3250, loss[loss=0.2019, simple_loss=0.2703, pruned_loss=0.06677, over 4838.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2546, pruned_loss=0.06107, over 955140.35 frames. ], batch size: 49, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:43:16,210 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66258.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:43:48,272 INFO [finetune.py:976] (4/7) Epoch 12, batch 3300, loss[loss=0.2082, simple_loss=0.2729, pruned_loss=0.07178, over 4930.00 frames. ], tot_loss[loss=0.1918, simple_loss=0.2587, pruned_loss=0.06245, over 955600.03 frames. ], batch size: 33, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:43:52,507 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-26 14:43:56,990 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66319.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:44:14,654 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.247e+02 1.642e+02 1.904e+02 2.370e+02 4.024e+02, threshold=3.808e+02, percent-clipped=2.0 +2023-03-26 14:44:20,883 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5755, 1.4139, 1.1112, 1.2261, 1.7564, 1.7739, 1.5459, 1.2906], + device='cuda:4'), covar=tensor([0.0265, 0.0345, 0.0869, 0.0394, 0.0304, 0.0417, 0.0350, 0.0434], + device='cuda:4'), in_proj_covar=tensor([0.0092, 0.0107, 0.0139, 0.0113, 0.0102, 0.0103, 0.0093, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.1604e-05, 8.3504e-05, 1.1019e-04, 8.7958e-05, 7.9406e-05, 7.6640e-05, + 7.0267e-05, 8.3275e-05], device='cuda:4') +2023-03-26 14:44:29,723 INFO [finetune.py:976] (4/7) Epoch 12, batch 3350, loss[loss=0.1814, simple_loss=0.2511, pruned_loss=0.05582, over 4889.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.2606, pruned_loss=0.06305, over 956309.95 frames. ], batch size: 32, lr: 3.64e-03, grad_scale: 64.0 +2023-03-26 14:44:38,932 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.51 vs. limit=5.0 +2023-03-26 14:45:02,613 INFO [finetune.py:976] (4/7) Epoch 12, batch 3400, loss[loss=0.1759, simple_loss=0.2485, pruned_loss=0.05167, over 4824.00 frames. ], tot_loss[loss=0.1947, simple_loss=0.2617, pruned_loss=0.06381, over 956316.51 frames. ], batch size: 30, lr: 3.64e-03, grad_scale: 64.0 +2023-03-26 14:45:11,010 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3753, 2.2819, 2.3583, 1.6079, 2.3848, 2.5740, 2.4689, 2.0002], + device='cuda:4'), covar=tensor([0.0673, 0.0651, 0.0699, 0.0935, 0.0573, 0.0649, 0.0623, 0.1046], + device='cuda:4'), in_proj_covar=tensor([0.0135, 0.0133, 0.0141, 0.0123, 0.0122, 0.0142, 0.0142, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:45:24,440 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.132e+02 1.693e+02 1.994e+02 2.432e+02 3.824e+02, threshold=3.988e+02, percent-clipped=2.0 +2023-03-26 14:45:36,093 INFO [finetune.py:976] (4/7) Epoch 12, batch 3450, loss[loss=0.169, simple_loss=0.2359, pruned_loss=0.05102, over 4918.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.2607, pruned_loss=0.06292, over 956824.57 frames. ], batch size: 33, lr: 3.64e-03, grad_scale: 64.0 +2023-03-26 14:45:40,271 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1969, 1.3093, 1.5340, 1.0371, 1.1826, 1.3892, 1.2967, 1.5292], + device='cuda:4'), covar=tensor([0.1290, 0.1985, 0.1264, 0.1428, 0.0874, 0.1196, 0.2619, 0.0840], + device='cuda:4'), in_proj_covar=tensor([0.0194, 0.0203, 0.0192, 0.0189, 0.0177, 0.0212, 0.0215, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:46:02,769 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66494.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 14:46:06,887 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66500.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:46:09,852 INFO [finetune.py:976] (4/7) Epoch 12, batch 3500, loss[loss=0.1866, simple_loss=0.2569, pruned_loss=0.05816, over 4868.00 frames. ], tot_loss[loss=0.1917, simple_loss=0.2585, pruned_loss=0.06242, over 957270.52 frames. ], batch size: 31, lr: 3.64e-03, grad_scale: 64.0 +2023-03-26 14:46:20,800 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66516.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 14:46:23,850 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66521.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:46:36,432 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.655e+02 1.937e+02 2.486e+02 6.010e+02, threshold=3.875e+02, percent-clipped=2.0 +2023-03-26 14:46:40,007 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66542.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:46:57,618 INFO [finetune.py:976] (4/7) Epoch 12, batch 3550, loss[loss=0.2586, simple_loss=0.3009, pruned_loss=0.1082, over 4868.00 frames. ], tot_loss[loss=0.1892, simple_loss=0.2552, pruned_loss=0.06158, over 957056.31 frames. ], batch size: 34, lr: 3.64e-03, grad_scale: 64.0 +2023-03-26 14:47:18,499 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 14:47:23,089 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66582.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:47:43,652 INFO [finetune.py:976] (4/7) Epoch 12, batch 3600, loss[loss=0.1502, simple_loss=0.2236, pruned_loss=0.03841, over 4827.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2528, pruned_loss=0.06089, over 956088.87 frames. ], batch size: 33, lr: 3.64e-03, grad_scale: 64.0 +2023-03-26 14:47:53,890 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66614.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:48:08,738 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.581e+02 1.999e+02 2.430e+02 3.919e+02, threshold=3.999e+02, percent-clipped=1.0 +2023-03-26 14:48:21,117 INFO [finetune.py:976] (4/7) Epoch 12, batch 3650, loss[loss=0.1843, simple_loss=0.2485, pruned_loss=0.06007, over 4806.00 frames. ], tot_loss[loss=0.1895, simple_loss=0.2555, pruned_loss=0.06174, over 955570.26 frames. ], batch size: 25, lr: 3.64e-03, grad_scale: 64.0 +2023-03-26 14:48:26,025 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2091, 2.1060, 2.2951, 1.4769, 2.2744, 2.3611, 2.3018, 1.8917], + device='cuda:4'), covar=tensor([0.0685, 0.0679, 0.0697, 0.0986, 0.0616, 0.0683, 0.0597, 0.1101], + device='cuda:4'), in_proj_covar=tensor([0.0135, 0.0133, 0.0141, 0.0123, 0.0122, 0.0142, 0.0142, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:48:54,888 INFO [finetune.py:976] (4/7) Epoch 12, batch 3700, loss[loss=0.2129, simple_loss=0.2822, pruned_loss=0.07177, over 4852.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.258, pruned_loss=0.06184, over 955748.64 frames. ], batch size: 44, lr: 3.64e-03, grad_scale: 64.0 +2023-03-26 14:49:14,933 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66731.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:49:18,460 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.675e+02 1.999e+02 2.466e+02 3.717e+02, threshold=3.997e+02, percent-clipped=0.0 +2023-03-26 14:49:38,724 INFO [finetune.py:976] (4/7) Epoch 12, batch 3750, loss[loss=0.1849, simple_loss=0.2575, pruned_loss=0.05617, over 4922.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2589, pruned_loss=0.0621, over 956242.09 frames. ], batch size: 38, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:50:03,078 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66792.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:50:09,001 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66800.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 14:50:12,433 INFO [finetune.py:976] (4/7) Epoch 12, batch 3800, loss[loss=0.2592, simple_loss=0.3055, pruned_loss=0.1065, over 4101.00 frames. ], tot_loss[loss=0.195, simple_loss=0.262, pruned_loss=0.06399, over 955335.89 frames. ], batch size: 65, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:50:19,262 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66816.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 14:50:33,186 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6009, 1.5882, 1.3575, 1.5161, 1.9118, 1.6722, 1.5502, 1.3955], + device='cuda:4'), covar=tensor([0.0288, 0.0256, 0.0618, 0.0275, 0.0190, 0.0521, 0.0332, 0.0343], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0110, 0.0141, 0.0115, 0.0103, 0.0105, 0.0095, 0.0110], + device='cuda:4'), out_proj_covar=tensor([7.2604e-05, 8.5300e-05, 1.1204e-04, 8.9374e-05, 8.0706e-05, 7.8019e-05, + 7.1969e-05, 8.4380e-05], device='cuda:4') +2023-03-26 14:50:34,020 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.56 vs. limit=5.0 +2023-03-26 14:50:34,226 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.175e+02 1.668e+02 2.101e+02 2.666e+02 4.038e+02, threshold=4.202e+02, percent-clipped=1.0 +2023-03-26 14:50:40,361 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66848.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:50:45,466 INFO [finetune.py:976] (4/7) Epoch 12, batch 3850, loss[loss=0.1902, simple_loss=0.2515, pruned_loss=0.06442, over 4769.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2598, pruned_loss=0.0629, over 955799.15 frames. ], batch size: 28, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:50:51,375 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66864.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:51:00,117 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66877.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:51:04,705 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8976, 1.7859, 1.6204, 1.9655, 2.4271, 1.9899, 1.6272, 1.5264], + device='cuda:4'), covar=tensor([0.2015, 0.1953, 0.1823, 0.1493, 0.1685, 0.1136, 0.2335, 0.1935], + device='cuda:4'), in_proj_covar=tensor([0.0241, 0.0207, 0.0211, 0.0190, 0.0241, 0.0183, 0.0213, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:51:17,301 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6191, 1.4930, 1.9321, 1.2951, 1.5564, 1.9133, 1.4651, 2.0050], + device='cuda:4'), covar=tensor([0.1196, 0.2101, 0.1269, 0.1686, 0.0938, 0.1235, 0.2704, 0.0798], + device='cuda:4'), in_proj_covar=tensor([0.0197, 0.0207, 0.0195, 0.0193, 0.0180, 0.0216, 0.0218, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:51:18,888 INFO [finetune.py:976] (4/7) Epoch 12, batch 3900, loss[loss=0.1851, simple_loss=0.2528, pruned_loss=0.05869, over 4910.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2568, pruned_loss=0.06206, over 955131.56 frames. ], batch size: 37, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:51:24,972 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66914.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:51:40,886 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.325e+01 1.578e+02 1.785e+02 2.294e+02 5.103e+02, threshold=3.570e+02, percent-clipped=1.0 +2023-03-26 14:51:51,226 INFO [finetune.py:976] (4/7) Epoch 12, batch 3950, loss[loss=0.209, simple_loss=0.2489, pruned_loss=0.08454, over 4063.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2533, pruned_loss=0.06068, over 954899.82 frames. ], batch size: 17, lr: 3.64e-03, grad_scale: 32.0 +2023-03-26 14:51:53,038 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66957.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:51:58,484 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=66962.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:52:09,634 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.69 vs. limit=5.0 +2023-03-26 14:52:47,108 INFO [finetune.py:976] (4/7) Epoch 12, batch 4000, loss[loss=0.1901, simple_loss=0.2637, pruned_loss=0.05822, over 4822.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2522, pruned_loss=0.06007, over 955357.80 frames. ], batch size: 38, lr: 3.63e-03, grad_scale: 32.0 +2023-03-26 14:53:04,490 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67018.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 14:53:18,612 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.442e+01 1.595e+02 2.014e+02 2.521e+02 4.335e+02, threshold=4.027e+02, percent-clipped=3.0 +2023-03-26 14:53:28,877 INFO [finetune.py:976] (4/7) Epoch 12, batch 4050, loss[loss=0.2328, simple_loss=0.3062, pruned_loss=0.07971, over 4906.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2555, pruned_loss=0.0612, over 955382.68 frames. ], batch size: 37, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:53:47,904 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67083.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:53:50,807 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67087.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:54:02,029 INFO [finetune.py:976] (4/7) Epoch 12, batch 4100, loss[loss=0.2323, simple_loss=0.2919, pruned_loss=0.0863, over 4816.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2582, pruned_loss=0.0621, over 955350.41 frames. ], batch size: 30, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:54:17,653 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 14:54:29,985 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.152e+02 1.725e+02 1.998e+02 2.409e+02 3.172e+02, threshold=3.997e+02, percent-clipped=0.0 +2023-03-26 14:54:34,094 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67144.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:54:44,209 INFO [finetune.py:976] (4/7) Epoch 12, batch 4150, loss[loss=0.2107, simple_loss=0.2823, pruned_loss=0.06954, over 4930.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2621, pruned_loss=0.0642, over 956896.79 frames. ], batch size: 38, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:54:51,448 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4284, 1.3446, 1.8554, 2.8630, 1.9353, 2.1984, 1.0560, 2.3089], + device='cuda:4'), covar=tensor([0.1738, 0.1458, 0.1202, 0.0676, 0.0819, 0.1556, 0.1622, 0.0632], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0116, 0.0135, 0.0166, 0.0101, 0.0138, 0.0126, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 14:54:59,055 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67177.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:55:02,658 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9479, 1.7890, 2.3788, 1.5835, 2.1706, 2.3872, 1.7867, 2.4554], + device='cuda:4'), covar=tensor([0.1476, 0.2032, 0.1509, 0.2187, 0.0930, 0.1336, 0.2488, 0.0878], + device='cuda:4'), in_proj_covar=tensor([0.0198, 0.0208, 0.0197, 0.0195, 0.0180, 0.0217, 0.0219, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:55:17,546 INFO [finetune.py:976] (4/7) Epoch 12, batch 4200, loss[loss=0.1753, simple_loss=0.2394, pruned_loss=0.05559, over 4890.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2615, pruned_loss=0.06356, over 956395.61 frames. ], batch size: 35, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:55:30,582 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=67225.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:55:39,456 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.549e+02 1.852e+02 2.427e+02 4.145e+02, threshold=3.704e+02, percent-clipped=1.0 +2023-03-26 14:55:42,942 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5590, 1.4514, 1.9752, 2.8547, 2.0464, 2.2448, 1.2376, 2.3716], + device='cuda:4'), covar=tensor([0.1549, 0.1351, 0.1028, 0.0640, 0.0720, 0.1566, 0.1438, 0.0543], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0115, 0.0134, 0.0164, 0.0100, 0.0137, 0.0125, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 14:55:50,532 INFO [finetune.py:976] (4/7) Epoch 12, batch 4250, loss[loss=0.1431, simple_loss=0.2205, pruned_loss=0.03291, over 4820.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2588, pruned_loss=0.06283, over 954368.38 frames. ], batch size: 40, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:56:21,795 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7789, 1.2406, 0.6285, 1.7005, 2.1205, 1.3304, 1.4654, 1.6920], + device='cuda:4'), covar=tensor([0.1509, 0.2248, 0.2413, 0.1273, 0.1987, 0.2075, 0.1534, 0.2162], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0092, 0.0120, 0.0094, 0.0100, 0.0091], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 14:56:32,185 INFO [finetune.py:976] (4/7) Epoch 12, batch 4300, loss[loss=0.1605, simple_loss=0.2303, pruned_loss=0.04532, over 4820.00 frames. ], tot_loss[loss=0.1898, simple_loss=0.2558, pruned_loss=0.06191, over 955215.46 frames. ], batch size: 38, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:56:37,182 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67313.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 14:56:39,076 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7917, 1.5939, 1.4475, 1.7195, 2.0494, 1.8111, 1.2700, 1.4743], + device='cuda:4'), covar=tensor([0.2058, 0.2053, 0.1894, 0.1575, 0.1625, 0.1147, 0.2600, 0.1864], + device='cuda:4'), in_proj_covar=tensor([0.0239, 0.0206, 0.0210, 0.0190, 0.0240, 0.0182, 0.0213, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:56:54,399 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.954e+01 1.550e+02 1.913e+02 2.348e+02 5.397e+02, threshold=3.825e+02, percent-clipped=3.0 +2023-03-26 14:56:56,995 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0231, 2.2515, 1.8607, 1.7472, 2.4255, 2.5971, 2.2906, 2.0457], + device='cuda:4'), covar=tensor([0.0345, 0.0303, 0.0486, 0.0339, 0.0263, 0.0408, 0.0290, 0.0360], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0109, 0.0140, 0.0114, 0.0102, 0.0104, 0.0094, 0.0109], + device='cuda:4'), out_proj_covar=tensor([7.2350e-05, 8.4868e-05, 1.1102e-04, 8.8725e-05, 7.9956e-05, 7.7459e-05, + 7.1415e-05, 8.3984e-05], device='cuda:4') +2023-03-26 14:57:05,082 INFO [finetune.py:976] (4/7) Epoch 12, batch 4350, loss[loss=0.2187, simple_loss=0.2724, pruned_loss=0.08248, over 4869.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2527, pruned_loss=0.06086, over 955515.67 frames. ], batch size: 34, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:57:22,139 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7817, 1.7256, 1.5664, 1.7773, 1.1178, 3.7282, 1.3826, 1.9772], + device='cuda:4'), covar=tensor([0.3224, 0.2372, 0.2185, 0.2232, 0.1808, 0.0178, 0.2557, 0.1272], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0116, 0.0121, 0.0124, 0.0116, 0.0098, 0.0098, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 14:57:28,540 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67387.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:57:39,912 INFO [finetune.py:976] (4/7) Epoch 12, batch 4400, loss[loss=0.2058, simple_loss=0.2728, pruned_loss=0.06946, over 4812.00 frames. ], tot_loss[loss=0.1897, simple_loss=0.2546, pruned_loss=0.06238, over 955361.42 frames. ], batch size: 45, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:58:11,840 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3692, 1.3075, 1.2041, 1.3227, 1.5670, 1.3792, 1.2880, 1.2005], + device='cuda:4'), covar=tensor([0.0277, 0.0285, 0.0549, 0.0291, 0.0210, 0.0557, 0.0299, 0.0409], + device='cuda:4'), in_proj_covar=tensor([0.0092, 0.0109, 0.0139, 0.0113, 0.0102, 0.0104, 0.0094, 0.0109], + device='cuda:4'), out_proj_covar=tensor([7.1953e-05, 8.4359e-05, 1.1043e-04, 8.8285e-05, 7.9630e-05, 7.7071e-05, + 7.1083e-05, 8.3491e-05], device='cuda:4') +2023-03-26 14:58:14,141 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=67435.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 14:58:16,970 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.656e+02 1.972e+02 2.339e+02 4.406e+02, threshold=3.944e+02, percent-clipped=2.0 +2023-03-26 14:58:17,071 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67439.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:58:30,811 INFO [finetune.py:976] (4/7) Epoch 12, batch 4450, loss[loss=0.1574, simple_loss=0.228, pruned_loss=0.04346, over 4735.00 frames. ], tot_loss[loss=0.1914, simple_loss=0.2574, pruned_loss=0.06266, over 954159.65 frames. ], batch size: 23, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:58:59,274 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1483, 1.2762, 1.1396, 1.3487, 1.4274, 2.4067, 1.2789, 1.4141], + device='cuda:4'), covar=tensor([0.1047, 0.1983, 0.1182, 0.1003, 0.1662, 0.0399, 0.1519, 0.1789], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0081, 0.0085, 0.0078], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 14:59:03,971 INFO [finetune.py:976] (4/7) Epoch 12, batch 4500, loss[loss=0.2184, simple_loss=0.2807, pruned_loss=0.07804, over 4226.00 frames. ], tot_loss[loss=0.1941, simple_loss=0.26, pruned_loss=0.06406, over 955425.31 frames. ], batch size: 65, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:59:09,863 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67513.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 14:59:22,555 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7687, 1.5922, 1.4119, 1.3137, 1.5450, 1.4687, 1.5150, 2.0912], + device='cuda:4'), covar=tensor([0.3729, 0.3900, 0.3039, 0.3507, 0.3732, 0.2239, 0.3598, 0.1640], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0260, 0.0224, 0.0277, 0.0245, 0.0211, 0.0247, 0.0219], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 14:59:26,032 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.163e+02 1.686e+02 1.980e+02 2.352e+02 4.001e+02, threshold=3.961e+02, percent-clipped=1.0 +2023-03-26 14:59:37,240 INFO [finetune.py:976] (4/7) Epoch 12, batch 4550, loss[loss=0.1621, simple_loss=0.239, pruned_loss=0.0426, over 4910.00 frames. ], tot_loss[loss=0.1952, simple_loss=0.2619, pruned_loss=0.06423, over 953827.47 frames. ], batch size: 38, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 14:59:56,259 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67574.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 15:00:19,990 INFO [finetune.py:976] (4/7) Epoch 12, batch 4600, loss[loss=0.1773, simple_loss=0.2358, pruned_loss=0.05941, over 4317.00 frames. ], tot_loss[loss=0.1958, simple_loss=0.2624, pruned_loss=0.0646, over 953226.74 frames. ], batch size: 18, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:00:20,116 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3869, 1.3046, 1.2621, 1.3081, 1.5936, 1.5254, 1.3582, 1.2044], + device='cuda:4'), covar=tensor([0.0265, 0.0335, 0.0603, 0.0305, 0.0225, 0.0439, 0.0315, 0.0450], + device='cuda:4'), in_proj_covar=tensor([0.0092, 0.0108, 0.0139, 0.0113, 0.0102, 0.0104, 0.0094, 0.0109], + device='cuda:4'), out_proj_covar=tensor([7.2028e-05, 8.4287e-05, 1.1012e-04, 8.8147e-05, 7.9581e-05, 7.6915e-05, + 7.0973e-05, 8.3454e-05], device='cuda:4') +2023-03-26 15:00:24,951 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67613.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:00:24,997 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3642, 2.1748, 1.8619, 2.4630, 2.1839, 2.0611, 2.1004, 3.0957], + device='cuda:4'), covar=tensor([0.4273, 0.5559, 0.3901, 0.4806, 0.4580, 0.2742, 0.4983, 0.1833], + device='cuda:4'), in_proj_covar=tensor([0.0284, 0.0259, 0.0223, 0.0276, 0.0243, 0.0210, 0.0246, 0.0218], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:00:42,114 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.172e+02 1.477e+02 1.878e+02 2.272e+02 4.960e+02, threshold=3.756e+02, percent-clipped=1.0 +2023-03-26 15:00:53,235 INFO [finetune.py:976] (4/7) Epoch 12, batch 4650, loss[loss=0.1481, simple_loss=0.2099, pruned_loss=0.04313, over 4907.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2595, pruned_loss=0.06393, over 954184.79 frames. ], batch size: 46, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:00:56,982 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=67661.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:01:00,581 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.7119, 1.7108, 1.7936, 1.0224, 1.8010, 1.9157, 1.9782, 1.5507], + device='cuda:4'), covar=tensor([0.0770, 0.0602, 0.0425, 0.0517, 0.0391, 0.0535, 0.0287, 0.0595], + device='cuda:4'), in_proj_covar=tensor([0.0127, 0.0155, 0.0123, 0.0133, 0.0132, 0.0127, 0.0145, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.3864e-05, 1.1302e-04, 8.8597e-05, 9.5849e-05, 9.3961e-05, 9.2557e-05, + 1.0579e-04, 1.0703e-04], device='cuda:4') +2023-03-26 15:01:10,469 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67680.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:01:15,446 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-03-26 15:01:17,695 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67692.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:01:31,298 INFO [finetune.py:976] (4/7) Epoch 12, batch 4700, loss[loss=0.1732, simple_loss=0.2371, pruned_loss=0.05462, over 4906.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2567, pruned_loss=0.06298, over 955423.34 frames. ], batch size: 32, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:01:38,880 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.97 vs. limit=5.0 +2023-03-26 15:01:56,970 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.553e+02 1.823e+02 2.116e+02 3.808e+02, threshold=3.646e+02, percent-clipped=1.0 +2023-03-26 15:01:57,075 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67739.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:01:58,324 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67741.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:02:05,965 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67753.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 15:02:07,550 INFO [finetune.py:976] (4/7) Epoch 12, batch 4750, loss[loss=0.1819, simple_loss=0.2518, pruned_loss=0.05599, over 4749.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2545, pruned_loss=0.06212, over 954708.55 frames. ], batch size: 54, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:02:23,064 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4114, 0.9285, 0.8444, 1.2830, 1.8252, 0.7916, 1.1163, 1.3730], + device='cuda:4'), covar=tensor([0.1512, 0.2415, 0.1856, 0.1332, 0.2114, 0.2127, 0.1718, 0.1919], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0092, 0.0121, 0.0094, 0.0100, 0.0090], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 15:02:28,909 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=67787.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:02:40,333 INFO [finetune.py:976] (4/7) Epoch 12, batch 4800, loss[loss=0.1897, simple_loss=0.25, pruned_loss=0.06472, over 4792.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.258, pruned_loss=0.06329, over 954806.41 frames. ], batch size: 29, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:03:07,513 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.086e+02 1.756e+02 1.975e+02 2.556e+02 4.813e+02, threshold=3.950e+02, percent-clipped=3.0 +2023-03-26 15:03:25,933 INFO [finetune.py:976] (4/7) Epoch 12, batch 4850, loss[loss=0.1736, simple_loss=0.2235, pruned_loss=0.06187, over 4219.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2608, pruned_loss=0.06379, over 955354.30 frames. ], batch size: 17, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:03:33,963 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5946, 3.4437, 3.2705, 1.3671, 3.4913, 2.6368, 0.6874, 2.3030], + device='cuda:4'), covar=tensor([0.2353, 0.2160, 0.1703, 0.3691, 0.1107, 0.1059, 0.4505, 0.1571], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0173, 0.0160, 0.0129, 0.0156, 0.0121, 0.0146, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 15:03:39,909 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67869.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 15:04:03,160 INFO [finetune.py:976] (4/7) Epoch 12, batch 4900, loss[loss=0.1913, simple_loss=0.2644, pruned_loss=0.05909, over 4814.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.2622, pruned_loss=0.06422, over 953996.63 frames. ], batch size: 38, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:04:26,940 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.121e+02 1.717e+02 1.971e+02 2.418e+02 4.222e+02, threshold=3.942e+02, percent-clipped=1.0 +2023-03-26 15:04:36,659 INFO [finetune.py:976] (4/7) Epoch 12, batch 4950, loss[loss=0.174, simple_loss=0.2367, pruned_loss=0.05568, over 4701.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2623, pruned_loss=0.0638, over 954027.56 frames. ], batch size: 23, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:04:53,972 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67981.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:05:20,899 INFO [finetune.py:976] (4/7) Epoch 12, batch 5000, loss[loss=0.1695, simple_loss=0.2416, pruned_loss=0.04867, over 4822.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2605, pruned_loss=0.06299, over 955892.40 frames. ], batch size: 38, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:05:41,231 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68036.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:05:43,413 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.964e+01 1.543e+02 1.867e+02 2.301e+02 3.447e+02, threshold=3.734e+02, percent-clipped=0.0 +2023-03-26 15:05:46,841 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68042.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:05:50,385 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68048.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 15:05:54,512 INFO [finetune.py:976] (4/7) Epoch 12, batch 5050, loss[loss=0.1868, simple_loss=0.262, pruned_loss=0.05585, over 4901.00 frames. ], tot_loss[loss=0.1914, simple_loss=0.2579, pruned_loss=0.06243, over 955436.95 frames. ], batch size: 32, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:06:15,503 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68087.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 15:06:19,541 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.7114, 4.0617, 4.3115, 4.5543, 4.4526, 4.1566, 4.7684, 1.5681], + device='cuda:4'), covar=tensor([0.0648, 0.0768, 0.0758, 0.0768, 0.1043, 0.1410, 0.0577, 0.4990], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0242, 0.0276, 0.0291, 0.0329, 0.0282, 0.0300, 0.0293], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:06:22,384 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4476, 1.3665, 1.5035, 1.6482, 1.5286, 3.0663, 1.1317, 1.4617], + device='cuda:4'), covar=tensor([0.1137, 0.2383, 0.1252, 0.1112, 0.1977, 0.0324, 0.2196, 0.2356], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0084, 0.0078], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 15:06:27,695 INFO [finetune.py:976] (4/7) Epoch 12, batch 5100, loss[loss=0.1943, simple_loss=0.2557, pruned_loss=0.06647, over 4907.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2538, pruned_loss=0.06073, over 956345.51 frames. ], batch size: 35, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:06:36,274 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9263, 1.7475, 1.6110, 1.9603, 2.1504, 2.0534, 1.3351, 1.5849], + device='cuda:4'), covar=tensor([0.2139, 0.2018, 0.1943, 0.1609, 0.1735, 0.1100, 0.2605, 0.1883], + device='cuda:4'), in_proj_covar=tensor([0.0237, 0.0205, 0.0209, 0.0189, 0.0238, 0.0180, 0.0212, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:06:59,404 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.056e+02 1.565e+02 1.837e+02 2.198e+02 4.078e+02, threshold=3.675e+02, percent-clipped=2.0 +2023-03-26 15:07:05,462 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68148.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 15:07:10,950 INFO [finetune.py:976] (4/7) Epoch 12, batch 5150, loss[loss=0.2272, simple_loss=0.2861, pruned_loss=0.08415, over 4908.00 frames. ], tot_loss[loss=0.1879, simple_loss=0.2539, pruned_loss=0.06102, over 956821.23 frames. ], batch size: 37, lr: 3.63e-03, grad_scale: 16.0 +2023-03-26 15:07:16,509 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0273, 1.5174, 0.8150, 2.0117, 2.4455, 1.6115, 1.9659, 1.9112], + device='cuda:4'), covar=tensor([0.1392, 0.2033, 0.2244, 0.1108, 0.1857, 0.2007, 0.1362, 0.1898], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0096, 0.0113, 0.0092, 0.0121, 0.0094, 0.0100, 0.0091], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 15:07:19,522 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68169.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:07:43,721 INFO [finetune.py:976] (4/7) Epoch 12, batch 5200, loss[loss=0.2029, simple_loss=0.2634, pruned_loss=0.07119, over 4783.00 frames. ], tot_loss[loss=0.192, simple_loss=0.2583, pruned_loss=0.06282, over 958052.59 frames. ], batch size: 28, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:07:50,564 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9890, 1.9019, 1.4455, 1.8819, 1.9566, 1.6475, 2.5224, 1.9039], + device='cuda:4'), covar=tensor([0.1561, 0.2304, 0.3532, 0.3196, 0.2859, 0.1855, 0.2714, 0.2125], + device='cuda:4'), in_proj_covar=tensor([0.0179, 0.0186, 0.0233, 0.0256, 0.0242, 0.0198, 0.0213, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:07:51,093 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68217.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:08:05,789 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.664e+02 1.889e+02 2.252e+02 3.665e+02, threshold=3.778e+02, percent-clipped=0.0 +2023-03-26 15:08:12,438 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.7401, 4.1106, 4.2949, 4.4765, 4.4475, 4.2507, 4.8429, 1.4674], + device='cuda:4'), covar=tensor([0.0760, 0.0813, 0.0768, 0.1004, 0.1320, 0.1546, 0.0654, 0.6015], + device='cuda:4'), in_proj_covar=tensor([0.0351, 0.0242, 0.0276, 0.0291, 0.0329, 0.0282, 0.0300, 0.0294], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:08:16,523 INFO [finetune.py:976] (4/7) Epoch 12, batch 5250, loss[loss=0.2329, simple_loss=0.2984, pruned_loss=0.08371, over 4865.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2607, pruned_loss=0.06336, over 957606.87 frames. ], batch size: 34, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:08:24,217 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6271, 1.2655, 0.8842, 1.5748, 1.9746, 1.3825, 1.4911, 1.6460], + device='cuda:4'), covar=tensor([0.1498, 0.2017, 0.1998, 0.1166, 0.2063, 0.2093, 0.1418, 0.1903], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0096, 0.0113, 0.0092, 0.0121, 0.0094, 0.0100, 0.0091], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 15:09:03,121 INFO [finetune.py:976] (4/7) Epoch 12, batch 5300, loss[loss=0.1842, simple_loss=0.2632, pruned_loss=0.05259, over 4810.00 frames. ], tot_loss[loss=0.1939, simple_loss=0.2609, pruned_loss=0.06349, over 956556.96 frames. ], batch size: 38, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:09:24,996 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68336.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:09:25,576 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68337.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:09:26,707 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.294e+02 1.838e+02 2.123e+02 2.651e+02 4.524e+02, threshold=4.245e+02, percent-clipped=5.0 +2023-03-26 15:09:32,252 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68348.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:09:36,486 INFO [finetune.py:976] (4/7) Epoch 12, batch 5350, loss[loss=0.2067, simple_loss=0.2755, pruned_loss=0.0689, over 4735.00 frames. ], tot_loss[loss=0.1935, simple_loss=0.2611, pruned_loss=0.06296, over 955529.43 frames. ], batch size: 54, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:09:55,985 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68384.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:10:04,679 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68396.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:10:06,566 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68399.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:10:10,287 INFO [finetune.py:976] (4/7) Epoch 12, batch 5400, loss[loss=0.1524, simple_loss=0.2266, pruned_loss=0.03913, over 4689.00 frames. ], tot_loss[loss=0.1914, simple_loss=0.2584, pruned_loss=0.06217, over 954801.99 frames. ], batch size: 23, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:10:14,574 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8559, 1.8189, 1.8006, 1.8802, 1.5333, 3.4325, 1.6139, 2.0123], + device='cuda:4'), covar=tensor([0.3226, 0.2350, 0.2068, 0.2357, 0.1737, 0.0247, 0.2525, 0.1219], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0116, 0.0120, 0.0124, 0.0116, 0.0098, 0.0098, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 15:10:19,309 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.69 vs. limit=5.0 +2023-03-26 15:10:28,985 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.88 vs. limit=5.0 +2023-03-26 15:10:40,847 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.541e+02 1.801e+02 2.082e+02 4.267e+02, threshold=3.602e+02, percent-clipped=1.0 +2023-03-26 15:10:44,357 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68443.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 15:10:47,429 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4504, 1.3744, 1.3208, 1.3436, 0.8434, 2.4143, 0.7968, 1.3399], + device='cuda:4'), covar=tensor([0.4074, 0.3042, 0.2607, 0.3060, 0.2139, 0.0464, 0.2630, 0.1353], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0116, 0.0120, 0.0124, 0.0115, 0.0098, 0.0098, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 15:10:51,600 INFO [finetune.py:976] (4/7) Epoch 12, batch 5450, loss[loss=0.1941, simple_loss=0.2634, pruned_loss=0.06241, over 4829.00 frames. ], tot_loss[loss=0.1898, simple_loss=0.2559, pruned_loss=0.0618, over 953750.95 frames. ], batch size: 33, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:10:54,772 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68460.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:11:15,495 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-26 15:11:23,398 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3743, 1.4265, 1.5970, 1.5749, 1.6856, 3.0063, 1.2844, 1.5935], + device='cuda:4'), covar=tensor([0.1034, 0.1907, 0.1103, 0.1027, 0.1611, 0.0321, 0.1616, 0.1759], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0084, 0.0078], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 15:11:24,505 INFO [finetune.py:976] (4/7) Epoch 12, batch 5500, loss[loss=0.1886, simple_loss=0.2511, pruned_loss=0.06306, over 4750.00 frames. ], tot_loss[loss=0.1879, simple_loss=0.2535, pruned_loss=0.06114, over 953572.59 frames. ], batch size: 54, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:11:47,045 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.509e+02 1.942e+02 2.407e+02 6.603e+02, threshold=3.884e+02, percent-clipped=3.0 +2023-03-26 15:11:59,913 INFO [finetune.py:976] (4/7) Epoch 12, batch 5550, loss[loss=0.195, simple_loss=0.2544, pruned_loss=0.06779, over 4901.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2549, pruned_loss=0.06195, over 951614.29 frames. ], batch size: 37, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:12:23,730 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68578.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:12:39,652 INFO [finetune.py:976] (4/7) Epoch 12, batch 5600, loss[loss=0.1885, simple_loss=0.27, pruned_loss=0.05343, over 4904.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2603, pruned_loss=0.06405, over 951815.10 frames. ], batch size: 43, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:12:58,321 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68637.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:12:59,421 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.238e+02 1.664e+02 1.965e+02 2.319e+02 3.885e+02, threshold=3.931e+02, percent-clipped=1.0 +2023-03-26 15:12:59,535 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68639.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:13:09,174 INFO [finetune.py:976] (4/7) Epoch 12, batch 5650, loss[loss=0.1972, simple_loss=0.2754, pruned_loss=0.0595, over 4910.00 frames. ], tot_loss[loss=0.1956, simple_loss=0.2621, pruned_loss=0.06452, over 952142.73 frames. ], batch size: 37, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:13:15,024 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.75 vs. limit=5.0 +2023-03-26 15:13:27,862 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68685.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:13:27,904 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68685.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:13:41,827 INFO [finetune.py:976] (4/7) Epoch 12, batch 5700, loss[loss=0.1888, simple_loss=0.2328, pruned_loss=0.07237, over 3583.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2593, pruned_loss=0.06403, over 934488.51 frames. ], batch size: 15, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:13:42,560 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8255, 1.6199, 1.4840, 1.1912, 1.6319, 1.6417, 1.6082, 2.1094], + device='cuda:4'), covar=tensor([0.4155, 0.4266, 0.3625, 0.3959, 0.3762, 0.2554, 0.3666, 0.2022], + device='cuda:4'), in_proj_covar=tensor([0.0286, 0.0260, 0.0225, 0.0277, 0.0245, 0.0211, 0.0248, 0.0221], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:14:27,874 INFO [finetune.py:976] (4/7) Epoch 13, batch 0, loss[loss=0.1795, simple_loss=0.2403, pruned_loss=0.0593, over 4849.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2403, pruned_loss=0.0593, over 4849.00 frames. ], batch size: 31, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:14:27,874 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 15:14:42,136 INFO [finetune.py:1010] (4/7) Epoch 13, validation: loss=0.1598, simple_loss=0.23, pruned_loss=0.04482, over 2265189.00 frames. +2023-03-26 15:14:42,136 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 15:14:47,273 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.040e+02 1.546e+02 1.915e+02 2.253e+02 4.332e+02, threshold=3.830e+02, percent-clipped=1.0 +2023-03-26 15:14:49,807 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68743.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 15:14:52,124 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68746.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:14:58,519 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68755.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:15:15,980 INFO [finetune.py:976] (4/7) Epoch 13, batch 50, loss[loss=0.1733, simple_loss=0.234, pruned_loss=0.05633, over 4696.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.2658, pruned_loss=0.06699, over 217045.64 frames. ], batch size: 23, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:15:21,851 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=68791.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 15:15:57,665 INFO [finetune.py:976] (4/7) Epoch 13, batch 100, loss[loss=0.1484, simple_loss=0.2139, pruned_loss=0.04148, over 4838.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2556, pruned_loss=0.06175, over 378849.73 frames. ], batch size: 49, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:16:02,754 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.134e+02 1.681e+02 1.901e+02 2.429e+02 4.753e+02, threshold=3.802e+02, percent-clipped=2.0 +2023-03-26 15:16:19,043 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 15:16:31,426 INFO [finetune.py:976] (4/7) Epoch 13, batch 150, loss[loss=0.1727, simple_loss=0.2379, pruned_loss=0.05374, over 4745.00 frames. ], tot_loss[loss=0.1868, simple_loss=0.252, pruned_loss=0.06079, over 506501.16 frames. ], batch size: 54, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:17:05,108 INFO [finetune.py:976] (4/7) Epoch 13, batch 200, loss[loss=0.1982, simple_loss=0.2668, pruned_loss=0.06476, over 4919.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2512, pruned_loss=0.06144, over 605788.12 frames. ], batch size: 38, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:17:05,753 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68934.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:17:09,208 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.222e+02 1.603e+02 1.930e+02 2.189e+02 8.191e+02, threshold=3.861e+02, percent-clipped=2.0 +2023-03-26 15:17:46,316 INFO [finetune.py:976] (4/7) Epoch 13, batch 250, loss[loss=0.1747, simple_loss=0.2498, pruned_loss=0.04983, over 4843.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2539, pruned_loss=0.06218, over 683987.09 frames. ], batch size: 30, lr: 3.62e-03, grad_scale: 16.0 +2023-03-26 15:18:05,434 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8878, 1.3168, 0.8177, 1.7640, 2.2041, 1.3516, 1.5625, 1.5877], + device='cuda:4'), covar=tensor([0.1871, 0.2771, 0.2618, 0.1608, 0.2227, 0.2541, 0.2037, 0.3049], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0111, 0.0091, 0.0119, 0.0093, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 15:18:19,709 INFO [finetune.py:976] (4/7) Epoch 13, batch 300, loss[loss=0.1809, simple_loss=0.2494, pruned_loss=0.05621, over 4828.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2575, pruned_loss=0.06281, over 744022.30 frames. ], batch size: 51, lr: 3.62e-03, grad_scale: 32.0 +2023-03-26 15:18:23,316 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.215e+02 1.585e+02 1.877e+02 2.328e+02 4.201e+02, threshold=3.755e+02, percent-clipped=2.0 +2023-03-26 15:18:24,576 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69041.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:18:25,870 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69043.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:18:30,878 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5308, 1.6438, 1.2328, 1.5784, 1.9668, 1.7755, 1.6162, 1.3517], + device='cuda:4'), covar=tensor([0.0336, 0.0307, 0.0654, 0.0318, 0.0203, 0.0569, 0.0300, 0.0437], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0108, 0.0139, 0.0113, 0.0101, 0.0104, 0.0094, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.2202e-05, 8.3653e-05, 1.1029e-04, 8.7830e-05, 7.8874e-05, 7.7218e-05, + 7.1315e-05, 8.3284e-05], device='cuda:4') +2023-03-26 15:18:34,496 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69055.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:18:55,351 INFO [finetune.py:976] (4/7) Epoch 13, batch 350, loss[loss=0.2245, simple_loss=0.294, pruned_loss=0.07745, over 4747.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2607, pruned_loss=0.06383, over 792594.54 frames. ], batch size: 54, lr: 3.62e-03, grad_scale: 32.0 +2023-03-26 15:19:18,985 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69103.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:19:19,632 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69104.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:19:41,449 INFO [finetune.py:976] (4/7) Epoch 13, batch 400, loss[loss=0.1885, simple_loss=0.2649, pruned_loss=0.05607, over 4774.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2616, pruned_loss=0.06338, over 829495.68 frames. ], batch size: 26, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:19:50,066 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.689e+02 1.999e+02 2.345e+02 4.076e+02, threshold=3.998e+02, percent-clipped=3.0 +2023-03-26 15:20:09,355 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 15:20:09,757 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69162.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:20:13,410 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69168.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:20:23,370 INFO [finetune.py:976] (4/7) Epoch 13, batch 450, loss[loss=0.1956, simple_loss=0.2586, pruned_loss=0.0663, over 4741.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2599, pruned_loss=0.0626, over 857504.04 frames. ], batch size: 27, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:20:29,472 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5648, 1.4046, 1.2544, 1.5281, 1.6583, 1.5726, 1.0439, 1.3238], + device='cuda:4'), covar=tensor([0.2140, 0.2081, 0.2003, 0.1719, 0.1511, 0.1222, 0.2485, 0.1868], + device='cuda:4'), in_proj_covar=tensor([0.0238, 0.0206, 0.0209, 0.0190, 0.0239, 0.0181, 0.0212, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:20:30,223 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-26 15:21:04,198 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69223.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:21:05,420 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69225.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:21:07,851 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69229.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:21:10,197 INFO [finetune.py:976] (4/7) Epoch 13, batch 500, loss[loss=0.138, simple_loss=0.2066, pruned_loss=0.03468, over 4827.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2575, pruned_loss=0.06171, over 880612.38 frames. ], batch size: 41, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:21:10,908 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69234.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:21:14,293 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.659e+02 1.928e+02 2.205e+02 4.798e+02, threshold=3.855e+02, percent-clipped=1.0 +2023-03-26 15:21:18,079 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69245.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 15:21:37,033 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69273.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:21:43,322 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69282.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:21:43,877 INFO [finetune.py:976] (4/7) Epoch 13, batch 550, loss[loss=0.1408, simple_loss=0.2257, pruned_loss=0.028, over 4756.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2537, pruned_loss=0.06044, over 895947.26 frames. ], batch size: 27, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:21:45,819 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69286.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:21:49,014 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-26 15:21:56,582 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6626, 1.7430, 2.1965, 2.0328, 1.9077, 3.5977, 1.5292, 1.8770], + device='cuda:4'), covar=tensor([0.0848, 0.1432, 0.1348, 0.0835, 0.1286, 0.0229, 0.1269, 0.1419], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0092, 0.0081, 0.0084, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 15:21:59,414 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69306.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 15:22:08,840 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69320.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:22:17,493 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=2.06 vs. limit=2.0 +2023-03-26 15:22:17,555 INFO [finetune.py:976] (4/7) Epoch 13, batch 600, loss[loss=0.2959, simple_loss=0.3409, pruned_loss=0.1254, over 3948.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2542, pruned_loss=0.06131, over 904182.38 frames. ], batch size: 65, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:22:18,297 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69334.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 15:22:21,204 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.059e+02 1.536e+02 1.861e+02 2.296e+02 3.946e+02, threshold=3.721e+02, percent-clipped=1.0 +2023-03-26 15:22:22,512 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69341.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:22:50,000 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69368.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:22:58,839 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69381.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:22:59,983 INFO [finetune.py:976] (4/7) Epoch 13, batch 650, loss[loss=0.167, simple_loss=0.2479, pruned_loss=0.0431, over 4929.00 frames. ], tot_loss[loss=0.1914, simple_loss=0.2576, pruned_loss=0.06257, over 915901.90 frames. ], batch size: 33, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:23:03,692 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69389.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:23:06,749 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69394.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:23:10,353 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69399.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:23:18,896 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0116, 2.0032, 1.8573, 2.0955, 2.4894, 2.0423, 1.8290, 1.5608], + device='cuda:4'), covar=tensor([0.2692, 0.2368, 0.2191, 0.1901, 0.2154, 0.1316, 0.2654, 0.2253], + device='cuda:4'), in_proj_covar=tensor([0.0239, 0.0207, 0.0210, 0.0190, 0.0241, 0.0182, 0.0213, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:23:30,655 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69429.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:23:33,419 INFO [finetune.py:976] (4/7) Epoch 13, batch 700, loss[loss=0.1857, simple_loss=0.2647, pruned_loss=0.05331, over 4821.00 frames. ], tot_loss[loss=0.1922, simple_loss=0.2591, pruned_loss=0.06261, over 925061.51 frames. ], batch size: 38, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:23:37,537 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.702e+02 1.957e+02 2.425e+02 4.096e+02, threshold=3.913e+02, percent-clipped=2.0 +2023-03-26 15:23:47,861 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69455.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:23:54,779 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5737, 1.4528, 1.4365, 1.5322, 1.2186, 3.4628, 1.3358, 1.7162], + device='cuda:4'), covar=tensor([0.3493, 0.2613, 0.2216, 0.2497, 0.1877, 0.0186, 0.2809, 0.1421], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0116, 0.0120, 0.0123, 0.0115, 0.0098, 0.0098, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 15:23:58,907 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.7502, 4.0183, 4.3021, 4.4969, 4.4696, 4.2101, 4.8567, 1.4485], + device='cuda:4'), covar=tensor([0.0680, 0.0891, 0.0793, 0.0921, 0.1058, 0.1540, 0.0563, 0.5540], + device='cuda:4'), in_proj_covar=tensor([0.0352, 0.0243, 0.0277, 0.0292, 0.0329, 0.0283, 0.0302, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:24:06,514 INFO [finetune.py:976] (4/7) Epoch 13, batch 750, loss[loss=0.2329, simple_loss=0.2936, pruned_loss=0.08607, over 4898.00 frames. ], tot_loss[loss=0.1943, simple_loss=0.2616, pruned_loss=0.06352, over 933335.04 frames. ], batch size: 37, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:24:40,545 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69518.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:24:43,994 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6309, 1.0848, 0.6913, 1.5874, 2.0980, 1.2716, 1.2908, 1.5223], + device='cuda:4'), covar=tensor([0.2055, 0.3051, 0.3043, 0.1651, 0.2290, 0.3002, 0.2203, 0.2817], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0112, 0.0091, 0.0120, 0.0094, 0.0099, 0.0090], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 15:24:44,562 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69524.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:24:45,057 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.62 vs. limit=2.0 +2023-03-26 15:24:50,491 INFO [finetune.py:976] (4/7) Epoch 13, batch 800, loss[loss=0.2032, simple_loss=0.2675, pruned_loss=0.0694, over 4765.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2606, pruned_loss=0.06272, over 937867.94 frames. ], batch size: 28, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:24:57,843 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.694e+02 1.982e+02 2.355e+02 4.334e+02, threshold=3.964e+02, percent-clipped=1.0 +2023-03-26 15:25:08,475 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1218, 2.2565, 2.0307, 2.4268, 2.6490, 2.2309, 2.0886, 1.7045], + device='cuda:4'), covar=tensor([0.2496, 0.2093, 0.1982, 0.1698, 0.2043, 0.1176, 0.2284, 0.2170], + device='cuda:4'), in_proj_covar=tensor([0.0238, 0.0207, 0.0209, 0.0190, 0.0240, 0.0182, 0.0212, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:25:47,577 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69581.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:25:48,750 INFO [finetune.py:976] (4/7) Epoch 13, batch 850, loss[loss=0.1517, simple_loss=0.2225, pruned_loss=0.04047, over 4849.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2584, pruned_loss=0.06168, over 942183.50 frames. ], batch size: 44, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:25:52,438 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2222, 2.1055, 2.1916, 1.0820, 2.5327, 2.8023, 2.2511, 2.0541], + device='cuda:4'), covar=tensor([0.0909, 0.0701, 0.0489, 0.0703, 0.0722, 0.0474, 0.0431, 0.0647], + device='cuda:4'), in_proj_covar=tensor([0.0125, 0.0153, 0.0123, 0.0130, 0.0130, 0.0126, 0.0143, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.2788e-05, 1.1173e-04, 8.8324e-05, 9.3617e-05, 9.2401e-05, 9.1099e-05, + 1.0450e-04, 1.0596e-04], device='cuda:4') +2023-03-26 15:26:03,124 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69601.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 15:26:06,197 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4724, 2.3302, 1.9526, 2.5313, 2.3553, 2.1356, 2.7396, 2.4500], + device='cuda:4'), covar=tensor([0.1465, 0.2371, 0.3371, 0.2610, 0.2822, 0.1821, 0.3145, 0.2056], + device='cuda:4'), in_proj_covar=tensor([0.0180, 0.0188, 0.0234, 0.0256, 0.0244, 0.0200, 0.0214, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:26:17,339 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6801, 1.6361, 1.4254, 1.7402, 2.0521, 1.9087, 1.6430, 1.4705], + device='cuda:4'), covar=tensor([0.0322, 0.0314, 0.0606, 0.0290, 0.0203, 0.0437, 0.0339, 0.0402], + device='cuda:4'), in_proj_covar=tensor([0.0095, 0.0110, 0.0141, 0.0114, 0.0102, 0.0106, 0.0095, 0.0110], + device='cuda:4'), out_proj_covar=tensor([7.3684e-05, 8.5231e-05, 1.1159e-04, 8.8691e-05, 7.9938e-05, 7.8309e-05, + 7.1986e-05, 8.4338e-05], device='cuda:4') +2023-03-26 15:26:21,359 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69629.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 15:26:24,229 INFO [finetune.py:976] (4/7) Epoch 13, batch 900, loss[loss=0.2154, simple_loss=0.2671, pruned_loss=0.08183, over 4714.00 frames. ], tot_loss[loss=0.1892, simple_loss=0.2558, pruned_loss=0.06132, over 944760.83 frames. ], batch size: 23, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:26:27,891 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.145e+02 1.604e+02 1.856e+02 2.224e+02 3.601e+02, threshold=3.711e+02, percent-clipped=0.0 +2023-03-26 15:26:55,509 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69674.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 15:26:56,704 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69676.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:27:06,462 INFO [finetune.py:976] (4/7) Epoch 13, batch 950, loss[loss=0.1729, simple_loss=0.2364, pruned_loss=0.05471, over 4757.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.255, pruned_loss=0.06162, over 948165.98 frames. ], batch size: 27, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:27:29,375 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69699.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:27:53,701 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-26 15:27:54,599 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.9934, 0.8332, 0.7818, 1.0519, 1.1323, 1.0225, 0.9243, 0.8425], + device='cuda:4'), covar=tensor([0.0355, 0.0316, 0.0574, 0.0281, 0.0299, 0.0394, 0.0299, 0.0413], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0109, 0.0140, 0.0113, 0.0102, 0.0105, 0.0095, 0.0109], + device='cuda:4'), out_proj_covar=tensor([7.3458e-05, 8.4809e-05, 1.1124e-04, 8.8382e-05, 7.9712e-05, 7.8105e-05, + 7.1765e-05, 8.3802e-05], device='cuda:4') +2023-03-26 15:28:02,008 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 15:28:02,977 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69724.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:28:08,358 INFO [finetune.py:976] (4/7) Epoch 13, batch 1000, loss[loss=0.2707, simple_loss=0.319, pruned_loss=0.1112, over 4830.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2568, pruned_loss=0.06206, over 951540.59 frames. ], batch size: 39, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:28:09,644 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 15:28:10,719 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69735.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 15:28:12,999 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.598e+02 1.856e+02 2.406e+02 4.029e+02, threshold=3.712e+02, percent-clipped=2.0 +2023-03-26 15:28:18,440 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69747.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:28:20,809 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69750.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:28:40,317 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.74 vs. limit=5.0 +2023-03-26 15:28:52,947 INFO [finetune.py:976] (4/7) Epoch 13, batch 1050, loss[loss=0.1929, simple_loss=0.2717, pruned_loss=0.05706, over 4862.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2589, pruned_loss=0.06189, over 952479.03 frames. ], batch size: 44, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:29:38,125 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69818.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:29:48,114 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69824.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:29:59,172 INFO [finetune.py:976] (4/7) Epoch 13, batch 1100, loss[loss=0.1892, simple_loss=0.2616, pruned_loss=0.05843, over 4909.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2604, pruned_loss=0.06214, over 953142.06 frames. ], batch size: 36, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:30:02,882 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.063e+02 1.609e+02 1.898e+02 2.282e+02 6.010e+02, threshold=3.795e+02, percent-clipped=2.0 +2023-03-26 15:30:09,472 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7988, 1.2171, 0.8203, 1.6289, 2.0693, 1.4134, 1.4982, 1.6150], + device='cuda:4'), covar=tensor([0.1410, 0.2106, 0.2092, 0.1178, 0.1926, 0.2044, 0.1396, 0.1880], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0096, 0.0113, 0.0093, 0.0121, 0.0094, 0.0099, 0.0090], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 15:30:12,373 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.28 vs. limit=5.0 +2023-03-26 15:30:35,892 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69866.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:30:43,313 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69872.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:30:51,894 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69881.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:30:53,487 INFO [finetune.py:976] (4/7) Epoch 13, batch 1150, loss[loss=0.2405, simple_loss=0.299, pruned_loss=0.09103, over 4889.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2617, pruned_loss=0.06276, over 952804.79 frames. ], batch size: 43, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:31:12,320 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69901.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 15:31:42,226 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69929.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:31:42,274 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69929.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 15:31:44,611 INFO [finetune.py:976] (4/7) Epoch 13, batch 1200, loss[loss=0.1791, simple_loss=0.2497, pruned_loss=0.05425, over 4882.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2609, pruned_loss=0.06321, over 953648.71 frames. ], batch size: 32, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:31:48,753 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.031e+02 1.603e+02 1.893e+02 2.321e+02 3.158e+02, threshold=3.786e+02, percent-clipped=0.0 +2023-03-26 15:31:55,835 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69949.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 15:32:13,210 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69976.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:32:13,759 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=69977.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:32:17,364 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6966, 1.6909, 1.4855, 1.7438, 2.0754, 1.9628, 1.6871, 1.4542], + device='cuda:4'), covar=tensor([0.0329, 0.0322, 0.0576, 0.0269, 0.0181, 0.0394, 0.0356, 0.0398], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0109, 0.0140, 0.0113, 0.0101, 0.0105, 0.0095, 0.0109], + device='cuda:4'), out_proj_covar=tensor([7.3165e-05, 8.4740e-05, 1.1092e-04, 8.8067e-05, 7.9069e-05, 7.8170e-05, + 7.1828e-05, 8.3652e-05], device='cuda:4') +2023-03-26 15:32:17,838 INFO [finetune.py:976] (4/7) Epoch 13, batch 1250, loss[loss=0.1509, simple_loss=0.2254, pruned_loss=0.03821, over 4765.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2575, pruned_loss=0.06173, over 956259.23 frames. ], batch size: 26, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:32:45,370 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0694, 1.0412, 0.9857, 0.3410, 0.9278, 1.1897, 1.2016, 0.9940], + device='cuda:4'), covar=tensor([0.0940, 0.0734, 0.0563, 0.0622, 0.0588, 0.0727, 0.0444, 0.0775], + device='cuda:4'), in_proj_covar=tensor([0.0124, 0.0151, 0.0120, 0.0128, 0.0129, 0.0125, 0.0141, 0.0143], + device='cuda:4'), out_proj_covar=tensor([9.1747e-05, 1.1010e-04, 8.6705e-05, 9.2450e-05, 9.1264e-05, 9.0504e-05, + 1.0303e-04, 1.0419e-04], device='cuda:4') +2023-03-26 15:32:46,509 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70024.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:32:46,536 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70024.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:32:47,251 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.16 vs. limit=5.0 +2023-03-26 15:32:50,168 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=70030.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 15:32:52,398 INFO [finetune.py:976] (4/7) Epoch 13, batch 1300, loss[loss=0.185, simple_loss=0.2497, pruned_loss=0.06017, over 4928.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2551, pruned_loss=0.06128, over 957549.52 frames. ], batch size: 38, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:32:52,526 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0906, 1.8918, 2.0344, 0.8374, 2.2317, 2.4355, 2.0212, 1.8374], + device='cuda:4'), covar=tensor([0.0943, 0.0815, 0.0456, 0.0766, 0.0431, 0.0610, 0.0500, 0.0698], + device='cuda:4'), in_proj_covar=tensor([0.0124, 0.0150, 0.0120, 0.0128, 0.0129, 0.0125, 0.0141, 0.0143], + device='cuda:4'), out_proj_covar=tensor([9.1692e-05, 1.1003e-04, 8.6630e-05, 9.2386e-05, 9.1150e-05, 9.0448e-05, + 1.0294e-04, 1.0406e-04], device='cuda:4') +2023-03-26 15:32:56,052 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.095e+02 1.649e+02 1.897e+02 2.309e+02 4.234e+02, threshold=3.795e+02, percent-clipped=2.0 +2023-03-26 15:33:03,827 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70050.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:33:06,697 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2170, 2.6750, 2.4165, 1.8493, 2.5667, 2.7163, 2.6012, 2.2202], + device='cuda:4'), covar=tensor([0.0689, 0.0557, 0.0726, 0.0878, 0.0728, 0.0674, 0.0601, 0.0904], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0130, 0.0140, 0.0123, 0.0122, 0.0140, 0.0140, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:33:10,858 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6271, 3.3925, 3.2525, 1.4770, 3.5359, 2.6616, 1.1369, 2.2649], + device='cuda:4'), covar=tensor([0.2325, 0.2139, 0.1519, 0.3421, 0.1197, 0.1079, 0.3727, 0.1654], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0175, 0.0160, 0.0129, 0.0156, 0.0121, 0.0145, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 15:33:16,965 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 15:33:19,135 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70072.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:33:25,839 INFO [finetune.py:976] (4/7) Epoch 13, batch 1350, loss[loss=0.1495, simple_loss=0.2219, pruned_loss=0.03853, over 4753.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2548, pruned_loss=0.06121, over 957340.79 frames. ], batch size: 28, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:33:36,434 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70098.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:33:53,159 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=70110.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:34:08,079 INFO [finetune.py:976] (4/7) Epoch 13, batch 1400, loss[loss=0.1652, simple_loss=0.2249, pruned_loss=0.05277, over 3912.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2577, pruned_loss=0.06231, over 954075.07 frames. ], batch size: 17, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:34:12,154 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.202e+02 1.588e+02 1.939e+02 2.393e+02 8.943e+02, threshold=3.877e+02, percent-clipped=1.0 +2023-03-26 15:34:34,200 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=70171.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:34:41,772 INFO [finetune.py:976] (4/7) Epoch 13, batch 1450, loss[loss=0.2061, simple_loss=0.2675, pruned_loss=0.07232, over 4929.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2597, pruned_loss=0.06316, over 952062.39 frames. ], batch size: 33, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:35:26,440 INFO [finetune.py:976] (4/7) Epoch 13, batch 1500, loss[loss=0.182, simple_loss=0.2364, pruned_loss=0.06378, over 4346.00 frames. ], tot_loss[loss=0.1928, simple_loss=0.2596, pruned_loss=0.06301, over 953116.58 frames. ], batch size: 19, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:35:30,134 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.195e+02 1.613e+02 1.899e+02 2.364e+02 4.350e+02, threshold=3.798e+02, percent-clipped=1.0 +2023-03-26 15:35:46,642 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 15:35:46,860 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=70260.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:35:53,455 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9124, 1.6765, 2.2721, 1.6548, 2.1097, 2.2779, 1.6382, 2.3606], + device='cuda:4'), covar=tensor([0.1499, 0.1902, 0.1637, 0.2010, 0.0838, 0.1413, 0.2740, 0.0790], + device='cuda:4'), in_proj_covar=tensor([0.0196, 0.0205, 0.0195, 0.0192, 0.0178, 0.0214, 0.0217, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:36:10,535 INFO [finetune.py:976] (4/7) Epoch 13, batch 1550, loss[loss=0.2164, simple_loss=0.2804, pruned_loss=0.07623, over 4864.00 frames. ], tot_loss[loss=0.1921, simple_loss=0.2597, pruned_loss=0.06225, over 954846.24 frames. ], batch size: 34, lr: 3.61e-03, grad_scale: 32.0 +2023-03-26 15:36:49,026 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0795, 1.0181, 1.0282, 0.4522, 0.8860, 1.1883, 1.2128, 1.0336], + device='cuda:4'), covar=tensor([0.0861, 0.0577, 0.0514, 0.0535, 0.0560, 0.0611, 0.0377, 0.0628], + device='cuda:4'), in_proj_covar=tensor([0.0125, 0.0151, 0.0122, 0.0129, 0.0130, 0.0126, 0.0142, 0.0144], + device='cuda:4'), out_proj_covar=tensor([9.2455e-05, 1.1064e-04, 8.7716e-05, 9.3188e-05, 9.1858e-05, 9.1104e-05, + 1.0348e-04, 1.0456e-04], device='cuda:4') +2023-03-26 15:36:49,645 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=70321.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:36:58,941 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70330.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 15:37:00,630 INFO [finetune.py:976] (4/7) Epoch 13, batch 1600, loss[loss=0.1658, simple_loss=0.2346, pruned_loss=0.04844, over 4768.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2577, pruned_loss=0.06171, over 956406.75 frames. ], batch size: 28, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:37:04,736 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.529e+02 1.873e+02 2.318e+02 5.550e+02, threshold=3.745e+02, percent-clipped=4.0 +2023-03-26 15:37:18,731 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-03-26 15:37:30,809 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70378.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 15:37:34,182 INFO [finetune.py:976] (4/7) Epoch 13, batch 1650, loss[loss=0.1529, simple_loss=0.2251, pruned_loss=0.0404, over 4722.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2562, pruned_loss=0.06121, over 957698.13 frames. ], batch size: 23, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:37:52,268 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 15:38:08,089 INFO [finetune.py:976] (4/7) Epoch 13, batch 1700, loss[loss=0.1291, simple_loss=0.1937, pruned_loss=0.0323, over 4795.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.2527, pruned_loss=0.05974, over 958843.08 frames. ], batch size: 25, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:38:10,637 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.2280, 2.7983, 2.5719, 1.4174, 2.6947, 2.2800, 2.2509, 2.4121], + device='cuda:4'), covar=tensor([0.1080, 0.0857, 0.1894, 0.2354, 0.1944, 0.2263, 0.2080, 0.1280], + device='cuda:4'), in_proj_covar=tensor([0.0166, 0.0197, 0.0200, 0.0186, 0.0215, 0.0208, 0.0224, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:38:11,734 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.265e+02 1.610e+02 1.926e+02 2.276e+02 4.227e+02, threshold=3.852e+02, percent-clipped=1.0 +2023-03-26 15:38:30,205 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=70466.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:38:41,456 INFO [finetune.py:976] (4/7) Epoch 13, batch 1750, loss[loss=0.1456, simple_loss=0.2231, pruned_loss=0.03409, over 4792.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.2541, pruned_loss=0.05965, over 959979.53 frames. ], batch size: 29, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:39:24,239 INFO [finetune.py:976] (4/7) Epoch 13, batch 1800, loss[loss=0.228, simple_loss=0.3026, pruned_loss=0.07676, over 4806.00 frames. ], tot_loss[loss=0.1887, simple_loss=0.2567, pruned_loss=0.06039, over 959397.87 frames. ], batch size: 51, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:39:28,345 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.316e+01 1.597e+02 2.051e+02 2.548e+02 3.844e+02, threshold=4.101e+02, percent-clipped=0.0 +2023-03-26 15:39:38,130 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7387, 1.5948, 1.5741, 1.7559, 1.2317, 3.8179, 1.4874, 2.0898], + device='cuda:4'), covar=tensor([0.3412, 0.2508, 0.2133, 0.2302, 0.1796, 0.0159, 0.2538, 0.1260], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0116, 0.0120, 0.0123, 0.0115, 0.0098, 0.0098, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 15:39:58,060 INFO [finetune.py:976] (4/7) Epoch 13, batch 1850, loss[loss=0.1924, simple_loss=0.2662, pruned_loss=0.05925, over 4814.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2585, pruned_loss=0.0615, over 957184.44 frames. ], batch size: 39, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:40:08,309 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6022, 1.4594, 1.4409, 1.5438, 1.4657, 3.6649, 1.5039, 2.1629], + device='cuda:4'), covar=tensor([0.4453, 0.3161, 0.2503, 0.3022, 0.1664, 0.0266, 0.2519, 0.1183], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0116, 0.0120, 0.0124, 0.0115, 0.0098, 0.0098, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 15:40:26,914 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=70616.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:40:42,679 INFO [finetune.py:976] (4/7) Epoch 13, batch 1900, loss[loss=0.182, simple_loss=0.2564, pruned_loss=0.0538, over 4919.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2608, pruned_loss=0.06218, over 957608.39 frames. ], batch size: 42, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:40:46,779 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.570e+02 1.884e+02 2.217e+02 6.026e+02, threshold=3.769e+02, percent-clipped=2.0 +2023-03-26 15:41:27,335 INFO [finetune.py:976] (4/7) Epoch 13, batch 1950, loss[loss=0.1632, simple_loss=0.2305, pruned_loss=0.04799, over 4762.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2592, pruned_loss=0.06123, over 958932.21 frames. ], batch size: 26, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:41:34,559 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1394, 2.1749, 1.7739, 1.9905, 2.6467, 2.6163, 2.1983, 1.9577], + device='cuda:4'), covar=tensor([0.0340, 0.0367, 0.0526, 0.0310, 0.0236, 0.0450, 0.0309, 0.0418], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0109, 0.0140, 0.0113, 0.0101, 0.0105, 0.0095, 0.0109], + device='cuda:4'), out_proj_covar=tensor([7.2937e-05, 8.4819e-05, 1.1104e-04, 8.8209e-05, 7.8768e-05, 7.8023e-05, + 7.2183e-05, 8.4054e-05], device='cuda:4') +2023-03-26 15:41:44,813 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6004, 1.4263, 1.2819, 1.5469, 1.5271, 1.5929, 0.8417, 1.3448], + device='cuda:4'), covar=tensor([0.2136, 0.2122, 0.1980, 0.1640, 0.1748, 0.1239, 0.2726, 0.1896], + device='cuda:4'), in_proj_covar=tensor([0.0239, 0.0207, 0.0209, 0.0190, 0.0239, 0.0182, 0.0213, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:42:06,899 INFO [finetune.py:976] (4/7) Epoch 13, batch 2000, loss[loss=0.1827, simple_loss=0.2501, pruned_loss=0.05763, over 4279.00 frames. ], tot_loss[loss=0.1898, simple_loss=0.2574, pruned_loss=0.06112, over 956312.60 frames. ], batch size: 66, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:42:15,807 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.177e+02 1.535e+02 1.807e+02 2.194e+02 3.140e+02, threshold=3.615e+02, percent-clipped=0.0 +2023-03-26 15:42:36,899 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70766.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:42:48,489 INFO [finetune.py:976] (4/7) Epoch 13, batch 2050, loss[loss=0.1652, simple_loss=0.2402, pruned_loss=0.04514, over 4751.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.2537, pruned_loss=0.05981, over 954211.08 frames. ], batch size: 26, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:43:09,368 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70814.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:43:11,804 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9430, 1.9878, 2.0790, 1.6730, 2.0243, 2.2974, 2.1171, 1.7349], + device='cuda:4'), covar=tensor([0.0565, 0.0504, 0.0627, 0.0787, 0.1015, 0.0459, 0.0485, 0.0907], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0131, 0.0140, 0.0123, 0.0122, 0.0140, 0.0140, 0.0159], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:43:22,317 INFO [finetune.py:976] (4/7) Epoch 13, batch 2100, loss[loss=0.2216, simple_loss=0.2846, pruned_loss=0.07927, over 4932.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.2539, pruned_loss=0.06055, over 952180.93 frames. ], batch size: 38, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:43:26,462 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.827e+01 1.609e+02 1.892e+02 2.240e+02 3.187e+02, threshold=3.783e+02, percent-clipped=0.0 +2023-03-26 15:43:56,100 INFO [finetune.py:976] (4/7) Epoch 13, batch 2150, loss[loss=0.1674, simple_loss=0.2489, pruned_loss=0.04293, over 4820.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2581, pruned_loss=0.06216, over 951972.09 frames. ], batch size: 25, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:44:14,243 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4537, 1.4396, 1.7990, 1.7507, 1.5397, 3.4048, 1.3529, 1.4705], + device='cuda:4'), covar=tensor([0.1255, 0.2367, 0.1380, 0.1260, 0.1940, 0.0298, 0.1983, 0.2419], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0084, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 15:44:35,060 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70916.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:44:46,790 INFO [finetune.py:976] (4/7) Epoch 13, batch 2200, loss[loss=0.1958, simple_loss=0.2633, pruned_loss=0.06419, over 4815.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.2612, pruned_loss=0.0627, over 952216.31 frames. ], batch size: 51, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:44:50,484 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.006e+02 1.701e+02 1.958e+02 2.316e+02 4.574e+02, threshold=3.916e+02, percent-clipped=1.0 +2023-03-26 15:45:07,668 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=70964.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:45:12,618 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1410, 2.0498, 1.5733, 2.0982, 2.0288, 1.7641, 2.4428, 2.1636], + device='cuda:4'), covar=tensor([0.1398, 0.2335, 0.3268, 0.2709, 0.2705, 0.1887, 0.3034, 0.1850], + device='cuda:4'), in_proj_covar=tensor([0.0179, 0.0188, 0.0234, 0.0255, 0.0245, 0.0200, 0.0213, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:45:19,186 INFO [finetune.py:976] (4/7) Epoch 13, batch 2250, loss[loss=0.2135, simple_loss=0.2781, pruned_loss=0.07441, over 4759.00 frames. ], tot_loss[loss=0.1941, simple_loss=0.262, pruned_loss=0.06314, over 952391.14 frames. ], batch size: 28, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:45:26,593 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=70992.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:45:39,207 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1002, 1.6715, 1.9186, 1.9182, 1.7265, 1.7665, 1.9304, 1.7823], + device='cuda:4'), covar=tensor([0.5327, 0.5143, 0.4693, 0.5093, 0.6612, 0.4993, 0.6445, 0.4475], + device='cuda:4'), in_proj_covar=tensor([0.0240, 0.0238, 0.0255, 0.0262, 0.0259, 0.0234, 0.0275, 0.0233], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:46:01,454 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2467, 3.6613, 3.8463, 4.0674, 4.0139, 3.7598, 4.2991, 1.4584], + device='cuda:4'), covar=tensor([0.0692, 0.0782, 0.0824, 0.0987, 0.1089, 0.1432, 0.0746, 0.5234], + device='cuda:4'), in_proj_covar=tensor([0.0349, 0.0244, 0.0276, 0.0291, 0.0330, 0.0282, 0.0301, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:46:02,073 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=71030.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:46:03,707 INFO [finetune.py:976] (4/7) Epoch 13, batch 2300, loss[loss=0.1891, simple_loss=0.2576, pruned_loss=0.0603, over 4779.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2624, pruned_loss=0.06298, over 953233.89 frames. ], batch size: 29, lr: 3.60e-03, grad_scale: 64.0 +2023-03-26 15:46:08,249 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.237e+01 1.685e+02 2.000e+02 2.324e+02 3.629e+02, threshold=3.999e+02, percent-clipped=0.0 +2023-03-26 15:46:23,772 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71053.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:46:46,193 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 15:46:59,590 INFO [finetune.py:976] (4/7) Epoch 13, batch 2350, loss[loss=0.1355, simple_loss=0.2021, pruned_loss=0.03442, over 4690.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.2611, pruned_loss=0.06277, over 953098.66 frames. ], batch size: 23, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:47:10,257 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71091.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:47:46,930 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.15 vs. limit=5.0 +2023-03-26 15:48:00,797 INFO [finetune.py:976] (4/7) Epoch 13, batch 2400, loss[loss=0.2016, simple_loss=0.2541, pruned_loss=0.0746, over 4802.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2576, pruned_loss=0.062, over 954488.42 frames. ], batch size: 51, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:48:09,285 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.076e+01 1.502e+02 1.791e+02 2.104e+02 3.987e+02, threshold=3.583e+02, percent-clipped=0.0 +2023-03-26 15:48:27,196 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.1776, 4.8281, 4.5453, 2.8335, 4.9309, 3.6749, 1.3453, 3.3589], + device='cuda:4'), covar=tensor([0.2088, 0.1440, 0.1183, 0.2947, 0.0667, 0.0848, 0.4367, 0.1523], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0172, 0.0158, 0.0127, 0.0155, 0.0120, 0.0144, 0.0121], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 15:49:05,621 INFO [finetune.py:976] (4/7) Epoch 13, batch 2450, loss[loss=0.1573, simple_loss=0.2236, pruned_loss=0.04543, over 4823.00 frames. ], tot_loss[loss=0.1882, simple_loss=0.2545, pruned_loss=0.06089, over 954856.85 frames. ], batch size: 51, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:50:04,538 INFO [finetune.py:976] (4/7) Epoch 13, batch 2500, loss[loss=0.189, simple_loss=0.2759, pruned_loss=0.05103, over 4835.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2572, pruned_loss=0.06195, over 955658.78 frames. ], batch size: 47, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:50:08,818 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.134e+02 1.629e+02 1.890e+02 2.415e+02 4.682e+02, threshold=3.780e+02, percent-clipped=4.0 +2023-03-26 15:50:40,376 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-26 15:50:41,417 INFO [finetune.py:976] (4/7) Epoch 13, batch 2550, loss[loss=0.2189, simple_loss=0.2874, pruned_loss=0.07521, over 4159.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2605, pruned_loss=0.06289, over 954332.42 frames. ], batch size: 65, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:51:01,090 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 15:51:22,575 INFO [finetune.py:976] (4/7) Epoch 13, batch 2600, loss[loss=0.1988, simple_loss=0.2734, pruned_loss=0.06216, over 4903.00 frames. ], tot_loss[loss=0.1958, simple_loss=0.263, pruned_loss=0.06431, over 955752.99 frames. ], batch size: 43, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:51:26,872 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.678e+02 1.922e+02 2.428e+02 5.321e+02, threshold=3.843e+02, percent-clipped=3.0 +2023-03-26 15:51:31,781 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=71348.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:51:44,507 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6455, 1.7817, 1.7917, 1.0234, 1.8416, 2.0819, 2.0255, 1.5915], + device='cuda:4'), covar=tensor([0.1007, 0.0568, 0.0473, 0.0610, 0.0404, 0.0468, 0.0356, 0.0641], + device='cuda:4'), in_proj_covar=tensor([0.0127, 0.0153, 0.0123, 0.0131, 0.0131, 0.0127, 0.0144, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.3827e-05, 1.1188e-04, 8.8530e-05, 9.4017e-05, 9.2975e-05, 9.2045e-05, + 1.0461e-04, 1.0624e-04], device='cuda:4') +2023-03-26 15:51:47,606 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.67 vs. limit=2.0 +2023-03-26 15:51:55,370 INFO [finetune.py:976] (4/7) Epoch 13, batch 2650, loss[loss=0.2312, simple_loss=0.2939, pruned_loss=0.0843, over 4803.00 frames. ], tot_loss[loss=0.1952, simple_loss=0.2629, pruned_loss=0.06375, over 955909.26 frames. ], batch size: 41, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:51:58,312 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=71386.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:52:29,325 INFO [finetune.py:976] (4/7) Epoch 13, batch 2700, loss[loss=0.1428, simple_loss=0.2091, pruned_loss=0.03826, over 4737.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.2602, pruned_loss=0.06236, over 955314.14 frames. ], batch size: 54, lr: 3.60e-03, grad_scale: 32.0 +2023-03-26 15:52:32,885 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0798, 2.0037, 1.6152, 1.9956, 2.0962, 1.7904, 2.5035, 2.1240], + device='cuda:4'), covar=tensor([0.1450, 0.2484, 0.3396, 0.2863, 0.2759, 0.1887, 0.2940, 0.1870], + device='cuda:4'), in_proj_covar=tensor([0.0179, 0.0187, 0.0234, 0.0254, 0.0244, 0.0199, 0.0213, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:52:34,537 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.578e+02 1.884e+02 2.307e+02 4.300e+02, threshold=3.769e+02, percent-clipped=2.0 +2023-03-26 15:53:02,925 INFO [finetune.py:976] (4/7) Epoch 13, batch 2750, loss[loss=0.2108, simple_loss=0.2747, pruned_loss=0.07346, over 4930.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2574, pruned_loss=0.06181, over 955744.24 frames. ], batch size: 33, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:53:05,701 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 15:53:34,931 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6396, 1.6476, 1.8138, 0.9614, 1.8048, 1.9593, 1.9352, 1.5441], + device='cuda:4'), covar=tensor([0.0911, 0.0640, 0.0429, 0.0549, 0.0393, 0.0477, 0.0326, 0.0632], + device='cuda:4'), in_proj_covar=tensor([0.0127, 0.0153, 0.0122, 0.0130, 0.0131, 0.0127, 0.0143, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.3632e-05, 1.1167e-04, 8.7933e-05, 9.3505e-05, 9.2790e-05, 9.1819e-05, + 1.0419e-04, 1.0561e-04], device='cuda:4') +2023-03-26 15:53:36,644 INFO [finetune.py:976] (4/7) Epoch 13, batch 2800, loss[loss=0.201, simple_loss=0.2588, pruned_loss=0.07165, over 4930.00 frames. ], tot_loss[loss=0.1868, simple_loss=0.2533, pruned_loss=0.06017, over 954441.93 frames. ], batch size: 33, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:53:40,886 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.138e+02 1.564e+02 1.863e+02 2.304e+02 3.302e+02, threshold=3.726e+02, percent-clipped=0.0 +2023-03-26 15:54:07,212 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 15:54:08,469 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 15:54:23,060 INFO [finetune.py:976] (4/7) Epoch 13, batch 2850, loss[loss=0.216, simple_loss=0.2756, pruned_loss=0.07816, over 4828.00 frames. ], tot_loss[loss=0.1858, simple_loss=0.2522, pruned_loss=0.05971, over 954566.35 frames. ], batch size: 30, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:54:43,630 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.18 vs. limit=5.0 +2023-03-26 15:54:52,346 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=71616.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:54:56,241 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1159, 2.1442, 2.2019, 1.5076, 2.0871, 2.3379, 2.2167, 1.8304], + device='cuda:4'), covar=tensor([0.0620, 0.0574, 0.0679, 0.0977, 0.0708, 0.0696, 0.0645, 0.1036], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0133, 0.0143, 0.0125, 0.0124, 0.0143, 0.0142, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:55:06,324 INFO [finetune.py:976] (4/7) Epoch 13, batch 2900, loss[loss=0.172, simple_loss=0.2412, pruned_loss=0.05139, over 4761.00 frames. ], tot_loss[loss=0.1897, simple_loss=0.2564, pruned_loss=0.06153, over 952351.23 frames. ], batch size: 27, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:55:15,493 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.635e+01 1.661e+02 1.944e+02 2.530e+02 6.475e+02, threshold=3.888e+02, percent-clipped=5.0 +2023-03-26 15:55:24,624 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=71648.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:55:27,638 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.5385, 1.6930, 1.8219, 0.9368, 1.7827, 2.0219, 1.9498, 1.5581], + device='cuda:4'), covar=tensor([0.0791, 0.0576, 0.0412, 0.0529, 0.0392, 0.0515, 0.0300, 0.0621], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0153, 0.0123, 0.0130, 0.0131, 0.0127, 0.0143, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.3606e-05, 1.1203e-04, 8.8265e-05, 9.3692e-05, 9.3188e-05, 9.2534e-05, + 1.0437e-04, 1.0629e-04], device='cuda:4') +2023-03-26 15:55:37,949 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5892, 1.5822, 1.9356, 2.0262, 1.6669, 3.6440, 1.3846, 1.5636], + device='cuda:4'), covar=tensor([0.0961, 0.1757, 0.1136, 0.0927, 0.1573, 0.0188, 0.1510, 0.1790], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0092, 0.0081, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 15:55:49,610 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71677.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:55:58,583 INFO [finetune.py:976] (4/7) Epoch 13, batch 2950, loss[loss=0.2097, simple_loss=0.2663, pruned_loss=0.07658, over 4215.00 frames. ], tot_loss[loss=0.1922, simple_loss=0.2594, pruned_loss=0.06253, over 951097.74 frames. ], batch size: 18, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:56:00,473 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=71686.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:56:11,125 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=71696.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:56:39,793 INFO [finetune.py:976] (4/7) Epoch 13, batch 3000, loss[loss=0.2374, simple_loss=0.3053, pruned_loss=0.08482, over 4892.00 frames. ], tot_loss[loss=0.1938, simple_loss=0.2612, pruned_loss=0.06323, over 953356.32 frames. ], batch size: 43, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:56:39,793 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 15:56:50,410 INFO [finetune.py:1010] (4/7) Epoch 13, validation: loss=0.1572, simple_loss=0.2278, pruned_loss=0.04333, over 2265189.00 frames. +2023-03-26 15:56:50,410 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 15:56:51,090 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=71734.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:56:55,669 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.204e+02 1.624e+02 1.953e+02 2.376e+02 4.887e+02, threshold=3.907e+02, percent-clipped=1.0 +2023-03-26 15:57:11,443 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9280, 1.8279, 1.6605, 1.9983, 2.3813, 2.0715, 1.7268, 1.5427], + device='cuda:4'), covar=tensor([0.2297, 0.1979, 0.2039, 0.1743, 0.1688, 0.1083, 0.2373, 0.2149], + device='cuda:4'), in_proj_covar=tensor([0.0240, 0.0208, 0.0211, 0.0191, 0.0241, 0.0184, 0.0215, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:57:11,996 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0473, 1.7802, 2.2200, 1.5110, 2.1657, 2.3014, 1.7594, 2.4798], + device='cuda:4'), covar=tensor([0.1355, 0.1842, 0.1592, 0.1919, 0.0878, 0.1295, 0.2597, 0.0918], + device='cuda:4'), in_proj_covar=tensor([0.0197, 0.0206, 0.0195, 0.0193, 0.0180, 0.0216, 0.0218, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 15:57:22,728 INFO [finetune.py:976] (4/7) Epoch 13, batch 3050, loss[loss=0.2147, simple_loss=0.2807, pruned_loss=0.07438, over 4774.00 frames. ], tot_loss[loss=0.1939, simple_loss=0.2615, pruned_loss=0.06316, over 953805.44 frames. ], batch size: 51, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:57:55,481 INFO [finetune.py:976] (4/7) Epoch 13, batch 3100, loss[loss=0.2026, simple_loss=0.2524, pruned_loss=0.07639, over 4808.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2582, pruned_loss=0.06159, over 953655.81 frames. ], batch size: 51, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:58:01,083 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.937e+01 1.560e+02 1.843e+02 2.215e+02 5.565e+02, threshold=3.687e+02, percent-clipped=1.0 +2023-03-26 15:58:08,161 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2301, 2.0178, 2.0301, 0.9587, 2.3154, 2.4190, 2.1397, 1.9947], + device='cuda:4'), covar=tensor([0.0952, 0.0743, 0.0546, 0.0704, 0.0463, 0.0789, 0.0480, 0.0705], + device='cuda:4'), in_proj_covar=tensor([0.0125, 0.0152, 0.0121, 0.0129, 0.0130, 0.0126, 0.0142, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.2834e-05, 1.1089e-04, 8.7332e-05, 9.2646e-05, 9.2054e-05, 9.1662e-05, + 1.0324e-04, 1.0496e-04], device='cuda:4') +2023-03-26 15:58:26,838 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6943, 1.5660, 1.5041, 1.6026, 1.9177, 1.8180, 1.5954, 1.3814], + device='cuda:4'), covar=tensor([0.0273, 0.0322, 0.0553, 0.0280, 0.0242, 0.0372, 0.0348, 0.0439], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0109, 0.0140, 0.0112, 0.0101, 0.0105, 0.0095, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.2538e-05, 8.4274e-05, 1.1052e-04, 8.7043e-05, 7.8868e-05, 7.7587e-05, + 7.1683e-05, 8.3093e-05], device='cuda:4') +2023-03-26 15:58:29,155 INFO [finetune.py:976] (4/7) Epoch 13, batch 3150, loss[loss=0.1989, simple_loss=0.2623, pruned_loss=0.06774, over 4912.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2553, pruned_loss=0.0609, over 954023.28 frames. ], batch size: 36, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:58:53,751 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=71919.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:59:03,054 INFO [finetune.py:976] (4/7) Epoch 13, batch 3200, loss[loss=0.2127, simple_loss=0.278, pruned_loss=0.07369, over 4909.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2523, pruned_loss=0.06002, over 955798.28 frames. ], batch size: 36, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:59:07,315 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.138e+02 1.561e+02 1.912e+02 2.265e+02 3.518e+02, threshold=3.824e+02, percent-clipped=0.0 +2023-03-26 15:59:40,774 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=71972.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:59:54,371 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71980.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 15:59:56,075 INFO [finetune.py:976] (4/7) Epoch 13, batch 3250, loss[loss=0.1415, simple_loss=0.2126, pruned_loss=0.03525, over 4830.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2533, pruned_loss=0.06068, over 953418.79 frames. ], batch size: 30, lr: 3.59e-03, grad_scale: 32.0 +2023-03-26 15:59:57,392 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2003, 3.6546, 3.8409, 4.0288, 3.9350, 3.7478, 4.2924, 1.3813], + device='cuda:4'), covar=tensor([0.0812, 0.0897, 0.0940, 0.1029, 0.1291, 0.1564, 0.0719, 0.5602], + device='cuda:4'), in_proj_covar=tensor([0.0348, 0.0243, 0.0276, 0.0290, 0.0330, 0.0282, 0.0302, 0.0297], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:00:04,178 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8003, 1.6716, 1.5126, 1.9512, 2.1021, 1.8869, 1.4070, 1.4012], + device='cuda:4'), covar=tensor([0.2302, 0.2163, 0.2050, 0.1675, 0.1900, 0.1277, 0.2655, 0.2071], + device='cuda:4'), in_proj_covar=tensor([0.0238, 0.0208, 0.0210, 0.0190, 0.0241, 0.0184, 0.0214, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:00:39,651 INFO [finetune.py:976] (4/7) Epoch 13, batch 3300, loss[loss=0.1902, simple_loss=0.2731, pruned_loss=0.05363, over 4901.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2571, pruned_loss=0.06185, over 951444.53 frames. ], batch size: 35, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:00:44,479 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.095e+02 1.593e+02 1.995e+02 2.341e+02 5.205e+02, threshold=3.991e+02, percent-clipped=4.0 +2023-03-26 16:01:29,182 INFO [finetune.py:976] (4/7) Epoch 13, batch 3350, loss[loss=0.1954, simple_loss=0.2638, pruned_loss=0.06352, over 4920.00 frames. ], tot_loss[loss=0.1918, simple_loss=0.2593, pruned_loss=0.06211, over 952627.07 frames. ], batch size: 38, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:01:45,261 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0752, 2.0354, 1.5699, 2.0408, 2.0099, 1.7418, 2.3601, 2.0211], + device='cuda:4'), covar=tensor([0.1368, 0.2061, 0.3017, 0.2325, 0.2665, 0.1644, 0.2971, 0.1776], + device='cuda:4'), in_proj_covar=tensor([0.0180, 0.0187, 0.0234, 0.0255, 0.0245, 0.0199, 0.0214, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:01:58,843 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.71 vs. limit=2.0 +2023-03-26 16:01:59,488 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-03-26 16:02:07,936 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72129.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:02:11,042 INFO [finetune.py:976] (4/7) Epoch 13, batch 3400, loss[loss=0.1755, simple_loss=0.2452, pruned_loss=0.05289, over 4724.00 frames. ], tot_loss[loss=0.1919, simple_loss=0.26, pruned_loss=0.06197, over 951913.54 frames. ], batch size: 54, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:02:16,764 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.109e+02 1.708e+02 2.008e+02 2.371e+02 4.954e+02, threshold=4.015e+02, percent-clipped=4.0 +2023-03-26 16:02:49,868 INFO [finetune.py:976] (4/7) Epoch 13, batch 3450, loss[loss=0.166, simple_loss=0.2398, pruned_loss=0.04607, over 4813.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2599, pruned_loss=0.06158, over 954487.19 frames. ], batch size: 39, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:02:53,946 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6867, 1.7921, 1.9043, 1.1155, 1.9121, 1.8718, 1.7823, 1.6041], + device='cuda:4'), covar=tensor([0.0633, 0.0640, 0.0631, 0.0860, 0.0634, 0.0725, 0.0606, 0.1088], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0133, 0.0141, 0.0123, 0.0124, 0.0142, 0.0141, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:02:55,188 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=72190.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:03:06,763 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-26 16:03:23,344 INFO [finetune.py:976] (4/7) Epoch 13, batch 3500, loss[loss=0.2204, simple_loss=0.2725, pruned_loss=0.08422, over 4816.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2576, pruned_loss=0.06143, over 955610.25 frames. ], batch size: 25, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:03:29,066 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.641e+02 1.993e+02 2.438e+02 4.377e+02, threshold=3.986e+02, percent-clipped=2.0 +2023-03-26 16:03:45,133 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4566, 1.3691, 1.3546, 1.3289, 0.9564, 2.3382, 0.7441, 1.3234], + device='cuda:4'), covar=tensor([0.4220, 0.3038, 0.2437, 0.3164, 0.1841, 0.0442, 0.2781, 0.1261], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0115, 0.0119, 0.0123, 0.0115, 0.0097, 0.0097, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 16:03:49,882 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=72272.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:03:51,651 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=72275.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:03:56,423 INFO [finetune.py:976] (4/7) Epoch 13, batch 3550, loss[loss=0.1857, simple_loss=0.2523, pruned_loss=0.0596, over 4859.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2547, pruned_loss=0.06096, over 957429.76 frames. ], batch size: 44, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:04:04,829 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 16:04:14,523 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.56 vs. limit=5.0 +2023-03-26 16:04:37,462 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=72320.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:04:47,562 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72326.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:04:51,696 INFO [finetune.py:976] (4/7) Epoch 13, batch 3600, loss[loss=0.1795, simple_loss=0.2532, pruned_loss=0.05291, over 4747.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2541, pruned_loss=0.06105, over 956298.43 frames. ], batch size: 23, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:04:58,283 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.845e+01 1.525e+02 1.754e+02 2.048e+02 3.586e+02, threshold=3.507e+02, percent-clipped=0.0 +2023-03-26 16:05:42,447 INFO [finetune.py:976] (4/7) Epoch 13, batch 3650, loss[loss=0.1807, simple_loss=0.2539, pruned_loss=0.05379, over 4911.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2554, pruned_loss=0.06143, over 956011.23 frames. ], batch size: 37, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:05:50,447 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=72387.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:06:20,306 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7499, 1.3772, 0.9282, 1.6823, 2.1742, 1.4526, 1.5289, 1.5991], + device='cuda:4'), covar=tensor([0.1507, 0.2033, 0.1893, 0.1183, 0.1819, 0.1921, 0.1475, 0.1943], + device='cuda:4'), in_proj_covar=tensor([0.0088, 0.0094, 0.0111, 0.0091, 0.0119, 0.0093, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 16:06:22,159 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.5011, 4.7235, 5.0321, 5.2911, 5.1936, 4.8998, 5.5978, 1.6715], + device='cuda:4'), covar=tensor([0.0601, 0.0834, 0.0689, 0.0944, 0.1120, 0.1360, 0.0428, 0.5586], + device='cuda:4'), in_proj_covar=tensor([0.0345, 0.0241, 0.0274, 0.0289, 0.0329, 0.0279, 0.0301, 0.0294], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:06:54,369 INFO [finetune.py:976] (4/7) Epoch 13, batch 3700, loss[loss=0.1951, simple_loss=0.2637, pruned_loss=0.06332, over 4932.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2587, pruned_loss=0.06222, over 952819.86 frames. ], batch size: 33, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:06:55,110 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2086, 2.0935, 1.6554, 2.1510, 2.0517, 1.8302, 2.4838, 2.1879], + device='cuda:4'), covar=tensor([0.1232, 0.1963, 0.2858, 0.2423, 0.2494, 0.1624, 0.2960, 0.1594], + device='cuda:4'), in_proj_covar=tensor([0.0180, 0.0187, 0.0234, 0.0254, 0.0245, 0.0199, 0.0213, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:07:04,383 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.108e+02 1.616e+02 1.915e+02 2.308e+02 4.437e+02, threshold=3.829e+02, percent-clipped=1.0 +2023-03-26 16:07:17,015 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5882, 1.4171, 1.9273, 3.1321, 2.1804, 2.2143, 0.9475, 2.4899], + device='cuda:4'), covar=tensor([0.1724, 0.1551, 0.1273, 0.0552, 0.0769, 0.1555, 0.1864, 0.0607], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0115, 0.0133, 0.0164, 0.0100, 0.0137, 0.0125, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 16:07:52,819 INFO [finetune.py:976] (4/7) Epoch 13, batch 3750, loss[loss=0.2173, simple_loss=0.2879, pruned_loss=0.07331, over 4768.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.2605, pruned_loss=0.06315, over 953586.37 frames. ], batch size: 28, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:07:54,153 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=72485.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:08:29,283 INFO [finetune.py:976] (4/7) Epoch 13, batch 3800, loss[loss=0.1874, simple_loss=0.2586, pruned_loss=0.05809, over 4917.00 frames. ], tot_loss[loss=0.1929, simple_loss=0.2606, pruned_loss=0.06264, over 952976.81 frames. ], batch size: 41, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:08:33,612 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-26 16:08:34,666 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.445e+01 1.578e+02 1.803e+02 2.155e+02 3.901e+02, threshold=3.607e+02, percent-clipped=1.0 +2023-03-26 16:08:56,832 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=72575.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:09:02,586 INFO [finetune.py:976] (4/7) Epoch 13, batch 3850, loss[loss=0.1717, simple_loss=0.2386, pruned_loss=0.05243, over 4828.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.258, pruned_loss=0.06129, over 954980.33 frames. ], batch size: 30, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:09:09,805 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.1647, 4.4672, 4.7205, 4.9675, 4.8998, 4.6522, 5.2333, 1.5961], + device='cuda:4'), covar=tensor([0.0665, 0.0814, 0.0738, 0.0787, 0.1200, 0.1314, 0.0536, 0.5273], + device='cuda:4'), in_proj_covar=tensor([0.0345, 0.0241, 0.0273, 0.0288, 0.0328, 0.0278, 0.0300, 0.0293], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:09:12,222 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7101, 1.3170, 0.8491, 1.5911, 2.0544, 1.2790, 1.4561, 1.5555], + device='cuda:4'), covar=tensor([0.1506, 0.1981, 0.1939, 0.1204, 0.1927, 0.1866, 0.1420, 0.1928], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0094, 0.0111, 0.0091, 0.0119, 0.0093, 0.0098, 0.0090], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 16:09:35,401 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=72623.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:09:46,306 INFO [finetune.py:976] (4/7) Epoch 13, batch 3900, loss[loss=0.1239, simple_loss=0.2006, pruned_loss=0.02367, over 4816.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2549, pruned_loss=0.06024, over 955037.29 frames. ], batch size: 51, lr: 3.59e-03, grad_scale: 16.0 +2023-03-26 16:09:51,190 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.402e+01 1.493e+02 1.751e+02 2.217e+02 3.590e+02, threshold=3.501e+02, percent-clipped=0.0 +2023-03-26 16:10:07,955 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.3331, 2.8243, 2.6942, 1.3228, 2.8588, 2.4719, 2.2804, 2.5763], + device='cuda:4'), covar=tensor([0.0771, 0.0869, 0.1590, 0.2284, 0.1477, 0.2249, 0.2057, 0.1034], + device='cuda:4'), in_proj_covar=tensor([0.0166, 0.0197, 0.0201, 0.0186, 0.0213, 0.0208, 0.0223, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:10:15,172 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8027, 4.0355, 3.8615, 2.0213, 4.2032, 3.0559, 0.7838, 2.8269], + device='cuda:4'), covar=tensor([0.2212, 0.1667, 0.1382, 0.3118, 0.1033, 0.0908, 0.4368, 0.1440], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0174, 0.0159, 0.0128, 0.0156, 0.0121, 0.0145, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 16:10:18,076 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=72682.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:10:18,646 INFO [finetune.py:976] (4/7) Epoch 13, batch 3950, loss[loss=0.1577, simple_loss=0.2317, pruned_loss=0.04184, over 4788.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2522, pruned_loss=0.05932, over 956048.53 frames. ], batch size: 29, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:10:41,449 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 16:10:51,393 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.7907, 1.8090, 2.0055, 1.1346, 2.0471, 2.2252, 2.1223, 1.7491], + device='cuda:4'), covar=tensor([0.0991, 0.0777, 0.0474, 0.0651, 0.0431, 0.0640, 0.0386, 0.0665], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0153, 0.0123, 0.0129, 0.0131, 0.0127, 0.0143, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.3420e-05, 1.1151e-04, 8.8160e-05, 9.3126e-05, 9.3126e-05, 9.1890e-05, + 1.0394e-04, 1.0563e-04], device='cuda:4') +2023-03-26 16:11:10,485 INFO [finetune.py:976] (4/7) Epoch 13, batch 4000, loss[loss=0.2173, simple_loss=0.2859, pruned_loss=0.07437, over 4902.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2518, pruned_loss=0.05888, over 954903.93 frames. ], batch size: 43, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:11:16,800 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.220e+02 1.596e+02 1.921e+02 2.181e+02 4.609e+02, threshold=3.842e+02, percent-clipped=3.0 +2023-03-26 16:11:19,239 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5840, 1.4542, 1.4468, 1.5107, 1.0473, 3.3418, 1.2864, 1.8859], + device='cuda:4'), covar=tensor([0.4034, 0.2974, 0.2366, 0.2847, 0.1895, 0.0295, 0.2486, 0.1186], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0116, 0.0120, 0.0123, 0.0115, 0.0098, 0.0097, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 16:11:44,632 INFO [finetune.py:976] (4/7) Epoch 13, batch 4050, loss[loss=0.1434, simple_loss=0.2134, pruned_loss=0.0367, over 4800.00 frames. ], tot_loss[loss=0.1887, simple_loss=0.2558, pruned_loss=0.06076, over 955622.23 frames. ], batch size: 25, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:11:46,034 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=72785.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:12:39,938 INFO [finetune.py:976] (4/7) Epoch 13, batch 4100, loss[loss=0.2068, simple_loss=0.2779, pruned_loss=0.06787, over 4894.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2579, pruned_loss=0.06134, over 954029.80 frames. ], batch size: 35, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:12:39,999 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=72833.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:12:45,293 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.089e+02 1.592e+02 1.875e+02 2.230e+02 3.624e+02, threshold=3.749e+02, percent-clipped=0.0 +2023-03-26 16:13:12,227 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8929, 1.2986, 1.8252, 1.8558, 1.6450, 1.5720, 1.8227, 1.6742], + device='cuda:4'), covar=tensor([0.3621, 0.3773, 0.3174, 0.3526, 0.4428, 0.3461, 0.4150, 0.3249], + device='cuda:4'), in_proj_covar=tensor([0.0241, 0.0237, 0.0256, 0.0264, 0.0261, 0.0236, 0.0276, 0.0234], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:13:13,291 INFO [finetune.py:976] (4/7) Epoch 13, batch 4150, loss[loss=0.2053, simple_loss=0.2667, pruned_loss=0.07195, over 4904.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2591, pruned_loss=0.06153, over 953103.01 frames. ], batch size: 36, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:13:26,385 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72894.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:13:44,276 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0854, 1.9721, 1.6685, 1.7890, 2.0331, 1.7424, 2.2644, 2.0283], + device='cuda:4'), covar=tensor([0.1423, 0.2091, 0.3238, 0.2748, 0.2742, 0.1817, 0.3356, 0.1900], + device='cuda:4'), in_proj_covar=tensor([0.0181, 0.0188, 0.0236, 0.0255, 0.0245, 0.0200, 0.0215, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:14:03,787 INFO [finetune.py:976] (4/7) Epoch 13, batch 4200, loss[loss=0.1845, simple_loss=0.2597, pruned_loss=0.05461, over 4832.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.259, pruned_loss=0.06089, over 954927.36 frames. ], batch size: 39, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:14:08,708 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.693e+01 1.496e+02 1.812e+02 2.169e+02 4.504e+02, threshold=3.624e+02, percent-clipped=2.0 +2023-03-26 16:14:18,724 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=72955.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:14:35,609 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72964.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 16:14:52,552 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=72982.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:14:53,056 INFO [finetune.py:976] (4/7) Epoch 13, batch 4250, loss[loss=0.1743, simple_loss=0.2455, pruned_loss=0.05153, over 4895.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2565, pruned_loss=0.06064, over 955576.95 frames. ], batch size: 32, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:15:15,247 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6895, 1.8081, 2.2345, 1.9692, 1.8517, 4.2231, 1.5221, 1.9589], + device='cuda:4'), covar=tensor([0.0895, 0.1637, 0.1015, 0.0958, 0.1416, 0.0184, 0.1444, 0.1628], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0073, 0.0077, 0.0091, 0.0081, 0.0084, 0.0078], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 16:15:19,028 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.57 vs. limit=5.0 +2023-03-26 16:15:21,372 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73025.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 16:15:23,184 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5742, 1.5851, 1.7009, 1.7789, 1.7252, 3.3505, 1.4237, 1.7655], + device='cuda:4'), covar=tensor([0.0942, 0.1750, 0.1029, 0.0913, 0.1481, 0.0286, 0.1493, 0.1603], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0081, 0.0084, 0.0078], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 16:15:24,380 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=73030.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:15:26,657 INFO [finetune.py:976] (4/7) Epoch 13, batch 4300, loss[loss=0.2338, simple_loss=0.2937, pruned_loss=0.08697, over 4943.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2544, pruned_loss=0.06039, over 954781.48 frames. ], batch size: 38, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:15:31,991 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.052e+02 1.498e+02 1.782e+02 2.254e+02 4.055e+02, threshold=3.563e+02, percent-clipped=2.0 +2023-03-26 16:15:32,767 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1371, 2.0885, 1.8038, 2.0960, 1.9598, 1.9818, 1.9878, 2.7205], + device='cuda:4'), covar=tensor([0.3988, 0.4506, 0.3376, 0.4163, 0.4443, 0.2521, 0.4479, 0.1680], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0259, 0.0223, 0.0276, 0.0246, 0.0212, 0.0246, 0.0221], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:15:59,435 INFO [finetune.py:976] (4/7) Epoch 13, batch 4350, loss[loss=0.1928, simple_loss=0.2513, pruned_loss=0.06714, over 4864.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2524, pruned_loss=0.06012, over 956243.31 frames. ], batch size: 31, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:16:28,096 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.23 vs. limit=5.0 +2023-03-26 16:16:29,856 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.18 vs. limit=5.0 +2023-03-26 16:16:34,970 INFO [finetune.py:976] (4/7) Epoch 13, batch 4400, loss[loss=0.1789, simple_loss=0.2516, pruned_loss=0.05309, over 4808.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2532, pruned_loss=0.06068, over 954569.20 frames. ], batch size: 45, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:16:40,306 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.328e+01 1.433e+02 1.829e+02 2.142e+02 3.915e+02, threshold=3.659e+02, percent-clipped=1.0 +2023-03-26 16:16:56,788 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0582, 1.9516, 1.6677, 1.8116, 1.8648, 1.8187, 1.8563, 2.5254], + device='cuda:4'), covar=tensor([0.3933, 0.4456, 0.3362, 0.3851, 0.4074, 0.2402, 0.4109, 0.1680], + device='cuda:4'), in_proj_covar=tensor([0.0286, 0.0260, 0.0223, 0.0276, 0.0246, 0.0212, 0.0247, 0.0221], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:17:08,725 INFO [finetune.py:976] (4/7) Epoch 13, batch 4450, loss[loss=0.2018, simple_loss=0.2558, pruned_loss=0.07391, over 4711.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2572, pruned_loss=0.06196, over 954320.23 frames. ], batch size: 23, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:17:43,141 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.0969, 3.5795, 3.7481, 3.9639, 3.8533, 3.5985, 4.1482, 1.3344], + device='cuda:4'), covar=tensor([0.0832, 0.0851, 0.0797, 0.0885, 0.1211, 0.1492, 0.0714, 0.5156], + device='cuda:4'), in_proj_covar=tensor([0.0349, 0.0243, 0.0276, 0.0290, 0.0331, 0.0282, 0.0301, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:17:53,150 INFO [finetune.py:976] (4/7) Epoch 13, batch 4500, loss[loss=0.1933, simple_loss=0.2715, pruned_loss=0.05752, over 4847.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2582, pruned_loss=0.062, over 953824.46 frames. ], batch size: 44, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:17:58,022 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.220e+02 1.732e+02 2.105e+02 2.505e+02 4.470e+02, threshold=4.210e+02, percent-clipped=3.0 +2023-03-26 16:18:04,479 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73250.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:18:05,146 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0043, 1.9912, 2.1119, 1.3939, 2.0976, 2.1526, 2.1858, 1.7803], + device='cuda:4'), covar=tensor([0.0565, 0.0575, 0.0642, 0.0882, 0.0598, 0.0715, 0.0546, 0.0963], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0131, 0.0139, 0.0121, 0.0121, 0.0139, 0.0139, 0.0159], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:18:26,893 INFO [finetune.py:976] (4/7) Epoch 13, batch 4550, loss[loss=0.1843, simple_loss=0.2568, pruned_loss=0.05593, over 4826.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.26, pruned_loss=0.06272, over 953751.08 frames. ], batch size: 39, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:18:34,965 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=73293.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:19:07,529 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73320.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 16:19:16,564 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3702, 2.3970, 1.9452, 1.9378, 2.7197, 2.7181, 2.2780, 2.0500], + device='cuda:4'), covar=tensor([0.0294, 0.0283, 0.0525, 0.0345, 0.0257, 0.0502, 0.0274, 0.0435], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0108, 0.0139, 0.0112, 0.0101, 0.0105, 0.0095, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.2555e-05, 8.4031e-05, 1.1005e-04, 8.7030e-05, 7.8630e-05, 7.7467e-05, + 7.1325e-05, 8.3049e-05], device='cuda:4') +2023-03-26 16:19:19,978 INFO [finetune.py:976] (4/7) Epoch 13, batch 4600, loss[loss=0.1837, simple_loss=0.2541, pruned_loss=0.05669, over 4908.00 frames. ], tot_loss[loss=0.1921, simple_loss=0.2593, pruned_loss=0.06244, over 954496.46 frames. ], batch size: 37, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:19:24,890 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.809e+01 1.600e+02 1.901e+02 2.194e+02 3.702e+02, threshold=3.803e+02, percent-clipped=0.0 +2023-03-26 16:19:41,999 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73354.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:20:11,745 INFO [finetune.py:976] (4/7) Epoch 13, batch 4650, loss[loss=0.1963, simple_loss=0.2602, pruned_loss=0.06622, over 4864.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2569, pruned_loss=0.06204, over 955797.69 frames. ], batch size: 34, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:20:20,931 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 16:20:45,240 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-03-26 16:20:45,673 INFO [finetune.py:976] (4/7) Epoch 13, batch 4700, loss[loss=0.1492, simple_loss=0.2244, pruned_loss=0.03696, over 4819.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2532, pruned_loss=0.06024, over 955510.77 frames. ], batch size: 41, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:20:50,438 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.031e+02 1.614e+02 1.909e+02 2.257e+02 3.771e+02, threshold=3.817e+02, percent-clipped=0.0 +2023-03-26 16:21:10,181 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0066, 1.8190, 1.5578, 1.5000, 1.7337, 1.7257, 1.7516, 2.3836], + device='cuda:4'), covar=tensor([0.3896, 0.4564, 0.3235, 0.4083, 0.3974, 0.2525, 0.3943, 0.1816], + device='cuda:4'), in_proj_covar=tensor([0.0286, 0.0260, 0.0223, 0.0276, 0.0246, 0.0212, 0.0247, 0.0222], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:21:18,783 INFO [finetune.py:976] (4/7) Epoch 13, batch 4750, loss[loss=0.1493, simple_loss=0.2209, pruned_loss=0.03889, over 4857.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2511, pruned_loss=0.0594, over 955567.44 frames. ], batch size: 49, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:21:40,706 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-26 16:21:51,903 INFO [finetune.py:976] (4/7) Epoch 13, batch 4800, loss[loss=0.1383, simple_loss=0.2104, pruned_loss=0.03308, over 4777.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.253, pruned_loss=0.06038, over 953007.53 frames. ], batch size: 28, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:21:57,196 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.195e+02 1.637e+02 2.007e+02 2.318e+02 3.852e+02, threshold=4.014e+02, percent-clipped=1.0 +2023-03-26 16:22:03,309 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=73550.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:22:24,806 INFO [finetune.py:976] (4/7) Epoch 13, batch 4850, loss[loss=0.1816, simple_loss=0.2596, pruned_loss=0.05186, over 4810.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.256, pruned_loss=0.06085, over 953209.41 frames. ], batch size: 51, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:22:30,084 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=73590.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:22:37,222 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=73598.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:22:48,551 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3455, 1.3305, 1.5948, 2.2475, 1.5454, 1.9482, 1.1368, 1.8721], + device='cuda:4'), covar=tensor([0.1471, 0.1189, 0.0978, 0.0703, 0.0838, 0.1733, 0.1149, 0.0643], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0115, 0.0133, 0.0165, 0.0101, 0.0138, 0.0126, 0.0103], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 16:23:00,189 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=73620.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 16:23:08,518 INFO [finetune.py:976] (4/7) Epoch 13, batch 4900, loss[loss=0.2157, simple_loss=0.2753, pruned_loss=0.07802, over 4895.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2586, pruned_loss=0.06221, over 950534.79 frames. ], batch size: 35, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:23:14,277 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.751e+02 2.108e+02 2.596e+02 5.059e+02, threshold=4.217e+02, percent-clipped=3.0 +2023-03-26 16:23:19,737 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73649.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:23:20,790 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.98 vs. limit=5.0 +2023-03-26 16:23:21,042 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73651.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:23:31,884 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=73668.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 16:23:33,393 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 16:23:41,382 INFO [finetune.py:976] (4/7) Epoch 13, batch 4950, loss[loss=0.2504, simple_loss=0.3183, pruned_loss=0.09128, over 4702.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2598, pruned_loss=0.06273, over 953143.46 frames. ], batch size: 59, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:24:24,420 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=73732.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:24:24,938 INFO [finetune.py:976] (4/7) Epoch 13, batch 5000, loss[loss=0.1662, simple_loss=0.24, pruned_loss=0.04622, over 4896.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2575, pruned_loss=0.06151, over 951914.47 frames. ], batch size: 36, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:24:33,742 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.586e+02 1.888e+02 2.371e+02 3.310e+02, threshold=3.776e+02, percent-clipped=1.0 +2023-03-26 16:24:48,045 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 16:24:52,642 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.1735, 2.7818, 2.6090, 1.3560, 2.6850, 2.2906, 2.1645, 2.4377], + device='cuda:4'), covar=tensor([0.0861, 0.0845, 0.1725, 0.2115, 0.1797, 0.2076, 0.1994, 0.1267], + device='cuda:4'), in_proj_covar=tensor([0.0166, 0.0196, 0.0199, 0.0185, 0.0212, 0.0207, 0.0222, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:25:13,969 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.54 vs. limit=5.0 +2023-03-26 16:25:17,331 INFO [finetune.py:976] (4/7) Epoch 13, batch 5050, loss[loss=0.1863, simple_loss=0.2492, pruned_loss=0.06169, over 4756.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2536, pruned_loss=0.06008, over 951779.43 frames. ], batch size: 23, lr: 3.58e-03, grad_scale: 16.0 +2023-03-26 16:25:23,910 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.70 vs. limit=5.0 +2023-03-26 16:25:26,640 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73793.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:25:34,588 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 16:25:53,901 INFO [finetune.py:976] (4/7) Epoch 13, batch 5100, loss[loss=0.1474, simple_loss=0.2145, pruned_loss=0.04017, over 4760.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.251, pruned_loss=0.05901, over 954046.21 frames. ], batch size: 23, lr: 3.57e-03, grad_scale: 16.0 +2023-03-26 16:25:59,155 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.804e+01 1.435e+02 1.752e+02 2.086e+02 3.868e+02, threshold=3.504e+02, percent-clipped=1.0 +2023-03-26 16:26:02,889 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5611, 3.4213, 3.1954, 1.5279, 3.4833, 2.7039, 0.8395, 2.2260], + device='cuda:4'), covar=tensor([0.2512, 0.2274, 0.1745, 0.3217, 0.1234, 0.0955, 0.4230, 0.1542], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0173, 0.0160, 0.0128, 0.0156, 0.0121, 0.0146, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 16:26:12,882 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0501, 0.9470, 0.9343, 1.1040, 1.2352, 1.1572, 1.0084, 0.9386], + device='cuda:4'), covar=tensor([0.0368, 0.0300, 0.0564, 0.0276, 0.0261, 0.0383, 0.0325, 0.0369], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0108, 0.0138, 0.0112, 0.0100, 0.0104, 0.0095, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.2444e-05, 8.3529e-05, 1.0959e-04, 8.7001e-05, 7.8451e-05, 7.6996e-05, + 7.1642e-05, 8.2765e-05], device='cuda:4') +2023-03-26 16:26:16,003 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.37 vs. limit=5.0 +2023-03-26 16:26:22,518 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3509, 1.2366, 1.7082, 2.4451, 1.6192, 2.1069, 0.8476, 2.0387], + device='cuda:4'), covar=tensor([0.1855, 0.1596, 0.1156, 0.0789, 0.1011, 0.1223, 0.1678, 0.0725], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0116, 0.0134, 0.0165, 0.0102, 0.0138, 0.0127, 0.0103], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 16:26:27,642 INFO [finetune.py:976] (4/7) Epoch 13, batch 5150, loss[loss=0.2207, simple_loss=0.2765, pruned_loss=0.08244, over 4866.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2509, pruned_loss=0.05906, over 954110.29 frames. ], batch size: 31, lr: 3.57e-03, grad_scale: 16.0 +2023-03-26 16:27:01,332 INFO [finetune.py:976] (4/7) Epoch 13, batch 5200, loss[loss=0.2161, simple_loss=0.2912, pruned_loss=0.07052, over 4854.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2545, pruned_loss=0.06121, over 951809.74 frames. ], batch size: 44, lr: 3.57e-03, grad_scale: 16.0 +2023-03-26 16:27:06,219 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.103e+02 1.675e+02 1.952e+02 2.217e+02 3.649e+02, threshold=3.904e+02, percent-clipped=2.0 +2023-03-26 16:27:09,801 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73946.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:27:11,690 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=73949.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:27:30,644 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8609, 1.7546, 2.2311, 3.1519, 2.1813, 2.4002, 1.3052, 2.4822], + device='cuda:4'), covar=tensor([0.1469, 0.1206, 0.1054, 0.0578, 0.0723, 0.1728, 0.1492, 0.0600], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0116, 0.0133, 0.0165, 0.0101, 0.0138, 0.0127, 0.0103], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 16:27:34,705 INFO [finetune.py:976] (4/7) Epoch 13, batch 5250, loss[loss=0.1048, simple_loss=0.1642, pruned_loss=0.02266, over 4134.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2573, pruned_loss=0.06195, over 949564.40 frames. ], batch size: 17, lr: 3.57e-03, grad_scale: 16.0 +2023-03-26 16:27:44,364 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=73997.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:27:47,368 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7940, 4.1568, 3.8800, 2.0199, 4.2431, 3.1173, 0.7167, 2.7662], + device='cuda:4'), covar=tensor([0.2170, 0.1425, 0.1420, 0.3084, 0.0852, 0.0878, 0.4583, 0.1411], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0173, 0.0161, 0.0128, 0.0157, 0.0122, 0.0146, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 16:27:53,799 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2811, 1.1440, 1.0845, 1.1598, 1.5299, 1.3948, 1.2832, 1.0575], + device='cuda:4'), covar=tensor([0.0361, 0.0352, 0.0625, 0.0344, 0.0223, 0.0479, 0.0307, 0.0457], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0109, 0.0140, 0.0113, 0.0101, 0.0105, 0.0096, 0.0109], + device='cuda:4'), out_proj_covar=tensor([7.3075e-05, 8.4324e-05, 1.1051e-04, 8.7701e-05, 7.9158e-05, 7.7751e-05, + 7.2079e-05, 8.3712e-05], device='cuda:4') +2023-03-26 16:28:11,316 INFO [finetune.py:976] (4/7) Epoch 13, batch 5300, loss[loss=0.2177, simple_loss=0.2758, pruned_loss=0.07981, over 4737.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2592, pruned_loss=0.0628, over 952062.97 frames. ], batch size: 54, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:28:17,126 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.638e+02 2.059e+02 2.515e+02 4.122e+02, threshold=4.117e+02, percent-clipped=3.0 +2023-03-26 16:28:33,156 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74064.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:28:45,017 INFO [finetune.py:976] (4/7) Epoch 13, batch 5350, loss[loss=0.1804, simple_loss=0.2536, pruned_loss=0.05362, over 4921.00 frames. ], tot_loss[loss=0.1921, simple_loss=0.2596, pruned_loss=0.06231, over 952670.80 frames. ], batch size: 33, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:28:48,545 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74088.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:29:12,968 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=74125.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:29:14,230 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8561, 0.9952, 1.8875, 1.7460, 1.6397, 1.5403, 1.6835, 1.6687], + device='cuda:4'), covar=tensor([0.3475, 0.4028, 0.3228, 0.3526, 0.4492, 0.3388, 0.3994, 0.3032], + device='cuda:4'), in_proj_covar=tensor([0.0239, 0.0236, 0.0254, 0.0263, 0.0260, 0.0234, 0.0274, 0.0233], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:29:18,159 INFO [finetune.py:976] (4/7) Epoch 13, batch 5400, loss[loss=0.2081, simple_loss=0.2701, pruned_loss=0.07301, over 4907.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2569, pruned_loss=0.06093, over 951616.04 frames. ], batch size: 37, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:29:20,122 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2955, 2.1986, 1.9205, 2.0996, 2.0625, 2.0515, 2.1284, 2.7959], + device='cuda:4'), covar=tensor([0.3956, 0.5125, 0.3451, 0.4378, 0.4291, 0.2723, 0.4417, 0.1774], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0259, 0.0223, 0.0275, 0.0245, 0.0212, 0.0247, 0.0222], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:29:27,406 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1655, 3.6125, 3.7648, 3.9828, 3.9315, 3.6541, 4.2721, 1.3690], + device='cuda:4'), covar=tensor([0.0754, 0.0797, 0.0787, 0.0958, 0.1091, 0.1410, 0.0634, 0.5115], + device='cuda:4'), in_proj_covar=tensor([0.0352, 0.0245, 0.0278, 0.0291, 0.0335, 0.0282, 0.0302, 0.0297], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:29:27,917 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.534e+02 1.853e+02 2.261e+02 4.254e+02, threshold=3.706e+02, percent-clipped=1.0 +2023-03-26 16:29:28,636 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6044, 1.5564, 1.3775, 1.5251, 1.8656, 1.7966, 1.6285, 1.3694], + device='cuda:4'), covar=tensor([0.0312, 0.0291, 0.0585, 0.0304, 0.0228, 0.0380, 0.0287, 0.0426], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0109, 0.0140, 0.0113, 0.0101, 0.0105, 0.0095, 0.0109], + device='cuda:4'), out_proj_covar=tensor([7.2980e-05, 8.4373e-05, 1.1076e-04, 8.8074e-05, 7.9209e-05, 7.7774e-05, + 7.1990e-05, 8.3801e-05], device='cuda:4') +2023-03-26 16:30:00,917 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.47 vs. limit=5.0 +2023-03-26 16:30:02,314 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7138, 1.3450, 0.8148, 1.5433, 2.0841, 1.2669, 1.4843, 1.5500], + device='cuda:4'), covar=tensor([0.1464, 0.2029, 0.2129, 0.1286, 0.2028, 0.1878, 0.1485, 0.2031], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0094, 0.0112, 0.0092, 0.0120, 0.0093, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 16:30:11,889 INFO [finetune.py:976] (4/7) Epoch 13, batch 5450, loss[loss=0.1478, simple_loss=0.2175, pruned_loss=0.03905, over 4903.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2532, pruned_loss=0.05973, over 951474.74 frames. ], batch size: 35, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:30:56,458 INFO [finetune.py:976] (4/7) Epoch 13, batch 5500, loss[loss=0.2084, simple_loss=0.2696, pruned_loss=0.07359, over 4099.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2512, pruned_loss=0.05884, over 952289.26 frames. ], batch size: 65, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:30:59,618 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6678, 1.5309, 1.9185, 1.3905, 1.7015, 1.9431, 1.4858, 2.1030], + device='cuda:4'), covar=tensor([0.1327, 0.2125, 0.1192, 0.1668, 0.0850, 0.1127, 0.2949, 0.0861], + device='cuda:4'), in_proj_covar=tensor([0.0194, 0.0203, 0.0192, 0.0189, 0.0177, 0.0212, 0.0214, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:31:01,344 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74240.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:31:01,848 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.610e+01 1.534e+02 1.911e+02 2.187e+02 5.924e+02, threshold=3.822e+02, percent-clipped=2.0 +2023-03-26 16:31:05,017 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74246.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:31:30,484 INFO [finetune.py:976] (4/7) Epoch 13, batch 5550, loss[loss=0.2166, simple_loss=0.29, pruned_loss=0.07159, over 4832.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2537, pruned_loss=0.06005, over 954139.21 frames. ], batch size: 47, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:31:37,728 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=74294.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:31:42,505 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=74301.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:32:02,275 INFO [finetune.py:976] (4/7) Epoch 13, batch 5600, loss[loss=0.1937, simple_loss=0.2588, pruned_loss=0.06432, over 4185.00 frames. ], tot_loss[loss=0.1892, simple_loss=0.2572, pruned_loss=0.0606, over 952224.60 frames. ], batch size: 65, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:32:06,846 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.439e+01 1.528e+02 1.833e+02 2.251e+02 4.644e+02, threshold=3.666e+02, percent-clipped=1.0 +2023-03-26 16:32:06,966 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2865, 1.2830, 1.1478, 1.3198, 1.5749, 1.4166, 1.3343, 1.1287], + device='cuda:4'), covar=tensor([0.0364, 0.0272, 0.0603, 0.0300, 0.0222, 0.0522, 0.0335, 0.0444], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0109, 0.0140, 0.0113, 0.0102, 0.0106, 0.0096, 0.0110], + device='cuda:4'), out_proj_covar=tensor([7.2949e-05, 8.4586e-05, 1.1090e-04, 8.8085e-05, 7.9217e-05, 7.8208e-05, + 7.2437e-05, 8.4349e-05], device='cuda:4') +2023-03-26 16:32:31,595 INFO [finetune.py:976] (4/7) Epoch 13, batch 5650, loss[loss=0.2465, simple_loss=0.2997, pruned_loss=0.09668, over 4884.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2588, pruned_loss=0.0606, over 953951.02 frames. ], batch size: 32, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:32:34,743 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 16:32:35,019 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74388.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:32:54,216 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74420.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:33:01,853 INFO [finetune.py:976] (4/7) Epoch 13, batch 5700, loss[loss=0.1523, simple_loss=0.2117, pruned_loss=0.04642, over 3957.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2553, pruned_loss=0.06009, over 937346.31 frames. ], batch size: 17, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:33:03,647 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=74436.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:33:06,486 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.627e+01 1.487e+02 1.916e+02 2.529e+02 4.839e+02, threshold=3.833e+02, percent-clipped=5.0 +2023-03-26 16:33:31,117 INFO [finetune.py:976] (4/7) Epoch 14, batch 0, loss[loss=0.1834, simple_loss=0.2608, pruned_loss=0.05298, over 4855.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2608, pruned_loss=0.05298, over 4855.00 frames. ], batch size: 31, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:33:31,117 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 16:33:41,691 INFO [finetune.py:1010] (4/7) Epoch 14, validation: loss=0.1582, simple_loss=0.2295, pruned_loss=0.04344, over 2265189.00 frames. +2023-03-26 16:33:41,692 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 16:33:53,665 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6547, 1.7445, 2.3298, 2.0669, 1.9128, 4.5622, 1.5574, 1.9813], + device='cuda:4'), covar=tensor([0.1001, 0.1830, 0.1104, 0.0998, 0.1607, 0.0255, 0.1574, 0.1709], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0084, 0.0078], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 16:34:14,920 INFO [finetune.py:976] (4/7) Epoch 14, batch 50, loss[loss=0.1569, simple_loss=0.2279, pruned_loss=0.04301, over 4841.00 frames. ], tot_loss[loss=0.1951, simple_loss=0.2613, pruned_loss=0.06448, over 216655.26 frames. ], batch size: 49, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:34:42,637 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.553e+01 1.578e+02 1.920e+02 2.248e+02 3.729e+02, threshold=3.841e+02, percent-clipped=1.0 +2023-03-26 16:35:04,284 INFO [finetune.py:976] (4/7) Epoch 14, batch 100, loss[loss=0.1642, simple_loss=0.2258, pruned_loss=0.05134, over 4179.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.254, pruned_loss=0.06144, over 380579.74 frames. ], batch size: 65, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:35:08,045 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-26 16:35:08,564 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4655, 1.3026, 1.7370, 2.4550, 1.6631, 2.2701, 0.9277, 2.0628], + device='cuda:4'), covar=tensor([0.1858, 0.1466, 0.1151, 0.0801, 0.0965, 0.1142, 0.1739, 0.0775], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0115, 0.0133, 0.0164, 0.0100, 0.0137, 0.0126, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 16:35:31,060 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74595.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:35:31,615 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74596.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:35:33,162 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-26 16:35:49,899 INFO [finetune.py:976] (4/7) Epoch 14, batch 150, loss[loss=0.21, simple_loss=0.2604, pruned_loss=0.07986, over 4800.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2495, pruned_loss=0.06114, over 505680.86 frames. ], batch size: 51, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:36:20,955 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8122, 1.7577, 1.7036, 1.7185, 1.4309, 3.2633, 1.5874, 1.9255], + device='cuda:4'), covar=tensor([0.2873, 0.2050, 0.1778, 0.2053, 0.1430, 0.0248, 0.2688, 0.1138], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0116, 0.0120, 0.0124, 0.0115, 0.0098, 0.0097, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 16:36:22,702 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.552e+02 1.795e+02 2.139e+02 3.747e+02, threshold=3.589e+02, percent-clipped=0.0 +2023-03-26 16:36:33,027 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=74656.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:36:35,948 INFO [finetune.py:976] (4/7) Epoch 14, batch 200, loss[loss=0.1554, simple_loss=0.231, pruned_loss=0.03989, over 4867.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2499, pruned_loss=0.05989, over 605835.92 frames. ], batch size: 34, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:37:09,844 INFO [finetune.py:976] (4/7) Epoch 14, batch 250, loss[loss=0.1517, simple_loss=0.2146, pruned_loss=0.04442, over 4708.00 frames. ], tot_loss[loss=0.1892, simple_loss=0.2549, pruned_loss=0.06174, over 683493.73 frames. ], batch size: 23, lr: 3.57e-03, grad_scale: 32.0 +2023-03-26 16:37:15,965 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74720.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:37:27,843 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7911, 1.6647, 1.5243, 1.8768, 2.3220, 1.8849, 1.5474, 1.4330], + device='cuda:4'), covar=tensor([0.2332, 0.2135, 0.2037, 0.1745, 0.1728, 0.1229, 0.2478, 0.2151], + device='cuda:4'), in_proj_covar=tensor([0.0238, 0.0207, 0.0209, 0.0190, 0.0239, 0.0183, 0.0213, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:37:30,096 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.719e+02 2.081e+02 2.468e+02 4.342e+02, threshold=4.162e+02, percent-clipped=2.0 +2023-03-26 16:37:42,696 INFO [finetune.py:976] (4/7) Epoch 14, batch 300, loss[loss=0.2261, simple_loss=0.2909, pruned_loss=0.08058, over 4910.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2572, pruned_loss=0.06185, over 743817.17 frames. ], batch size: 36, lr: 3.56e-03, grad_scale: 32.0 +2023-03-26 16:37:48,015 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=74768.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:37:48,686 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7999, 2.4997, 1.8867, 1.1572, 2.1289, 2.3235, 1.9657, 2.1941], + device='cuda:4'), covar=tensor([0.0651, 0.0843, 0.1501, 0.1899, 0.1366, 0.1838, 0.1984, 0.0882], + device='cuda:4'), in_proj_covar=tensor([0.0167, 0.0196, 0.0199, 0.0185, 0.0212, 0.0208, 0.0223, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:38:16,373 INFO [finetune.py:976] (4/7) Epoch 14, batch 350, loss[loss=0.191, simple_loss=0.2565, pruned_loss=0.06279, over 4868.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.2599, pruned_loss=0.06345, over 790514.30 frames. ], batch size: 31, lr: 3.56e-03, grad_scale: 32.0 +2023-03-26 16:38:28,472 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-26 16:38:36,815 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.618e+01 1.648e+02 1.967e+02 2.475e+02 5.107e+02, threshold=3.933e+02, percent-clipped=3.0 +2023-03-26 16:38:49,813 INFO [finetune.py:976] (4/7) Epoch 14, batch 400, loss[loss=0.2035, simple_loss=0.2748, pruned_loss=0.06614, over 4163.00 frames. ], tot_loss[loss=0.1935, simple_loss=0.2604, pruned_loss=0.06333, over 825362.72 frames. ], batch size: 65, lr: 3.56e-03, grad_scale: 32.0 +2023-03-26 16:39:13,551 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74896.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:39:22,402 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74909.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:39:23,515 INFO [finetune.py:976] (4/7) Epoch 14, batch 450, loss[loss=0.1983, simple_loss=0.2737, pruned_loss=0.06149, over 4803.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2581, pruned_loss=0.06201, over 854359.01 frames. ], batch size: 41, lr: 3.56e-03, grad_scale: 32.0 +2023-03-26 16:39:43,597 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.587e+02 1.868e+02 2.260e+02 4.285e+02, threshold=3.737e+02, percent-clipped=2.0 +2023-03-26 16:39:45,926 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=74944.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:39:47,273 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 16:39:50,652 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74951.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:39:58,629 INFO [finetune.py:976] (4/7) Epoch 14, batch 500, loss[loss=0.1392, simple_loss=0.2118, pruned_loss=0.03323, over 4834.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2553, pruned_loss=0.061, over 878547.45 frames. ], batch size: 47, lr: 3.56e-03, grad_scale: 32.0 +2023-03-26 16:40:00,597 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74964.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:40:08,972 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=74970.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:40:10,764 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 16:40:18,694 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1457, 1.9054, 1.4193, 0.6175, 1.6882, 1.7910, 1.5218, 1.8043], + device='cuda:4'), covar=tensor([0.0744, 0.0782, 0.1161, 0.1753, 0.1160, 0.2053, 0.2023, 0.0749], + device='cuda:4'), in_proj_covar=tensor([0.0166, 0.0194, 0.0198, 0.0184, 0.0212, 0.0207, 0.0222, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:40:45,475 INFO [finetune.py:976] (4/7) Epoch 14, batch 550, loss[loss=0.1798, simple_loss=0.25, pruned_loss=0.0548, over 4807.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2531, pruned_loss=0.06041, over 896669.79 frames. ], batch size: 51, lr: 3.56e-03, grad_scale: 32.0 +2023-03-26 16:40:57,859 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=75025.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:41:09,613 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.564e+02 1.846e+02 2.204e+02 7.411e+02, threshold=3.691e+02, percent-clipped=3.0 +2023-03-26 16:41:32,953 INFO [finetune.py:976] (4/7) Epoch 14, batch 600, loss[loss=0.163, simple_loss=0.2314, pruned_loss=0.04725, over 4909.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2544, pruned_loss=0.06116, over 909582.44 frames. ], batch size: 32, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:41:48,756 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 16:41:52,037 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6822, 1.7272, 1.4711, 1.6555, 2.0821, 1.8683, 1.6949, 1.4168], + device='cuda:4'), covar=tensor([0.0281, 0.0262, 0.0532, 0.0265, 0.0167, 0.0432, 0.0301, 0.0399], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0108, 0.0140, 0.0113, 0.0100, 0.0105, 0.0095, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.2149e-05, 8.3823e-05, 1.1066e-04, 8.7543e-05, 7.7967e-05, 7.8012e-05, + 7.1511e-05, 8.2796e-05], device='cuda:4') +2023-03-26 16:41:53,918 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 16:42:08,491 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5603, 1.3331, 1.9040, 2.9573, 2.1011, 2.0959, 1.2504, 2.3993], + device='cuda:4'), covar=tensor([0.1904, 0.1669, 0.1372, 0.0714, 0.0823, 0.1418, 0.1635, 0.0649], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0116, 0.0133, 0.0164, 0.0101, 0.0137, 0.0125, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 16:42:10,223 INFO [finetune.py:976] (4/7) Epoch 14, batch 650, loss[loss=0.1455, simple_loss=0.2133, pruned_loss=0.03889, over 4714.00 frames. ], tot_loss[loss=0.1892, simple_loss=0.2565, pruned_loss=0.06097, over 920315.37 frames. ], batch size: 23, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:42:30,913 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.139e+02 1.620e+02 1.922e+02 2.248e+02 3.855e+02, threshold=3.845e+02, percent-clipped=1.0 +2023-03-26 16:42:43,797 INFO [finetune.py:976] (4/7) Epoch 14, batch 700, loss[loss=0.1695, simple_loss=0.2454, pruned_loss=0.04678, over 4877.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2583, pruned_loss=0.06117, over 927647.88 frames. ], batch size: 32, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:42:55,836 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8684, 1.6979, 1.4900, 1.2986, 1.6376, 1.6761, 1.6067, 2.1786], + device='cuda:4'), covar=tensor([0.4166, 0.4185, 0.3436, 0.3889, 0.3963, 0.2410, 0.3894, 0.1898], + device='cuda:4'), in_proj_covar=tensor([0.0288, 0.0260, 0.0225, 0.0277, 0.0247, 0.0214, 0.0249, 0.0223], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:43:16,881 INFO [finetune.py:976] (4/7) Epoch 14, batch 750, loss[loss=0.2099, simple_loss=0.2797, pruned_loss=0.07006, over 4799.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2587, pruned_loss=0.06102, over 934965.64 frames. ], batch size: 40, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:43:18,328 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 16:43:28,279 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=75228.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:43:37,703 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.931e+01 1.583e+02 1.834e+02 2.163e+02 4.783e+02, threshold=3.668e+02, percent-clipped=1.0 +2023-03-26 16:43:44,264 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75251.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:43:48,199 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 16:43:50,665 INFO [finetune.py:976] (4/7) Epoch 14, batch 800, loss[loss=0.1768, simple_loss=0.2476, pruned_loss=0.05297, over 4823.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2579, pruned_loss=0.06039, over 940635.64 frames. ], batch size: 39, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:43:53,663 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75265.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:43:56,685 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=75270.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:44:09,557 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=75289.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:44:14,897 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0037, 1.7695, 2.3406, 1.5899, 2.0007, 2.1961, 1.8035, 2.4074], + device='cuda:4'), covar=tensor([0.0912, 0.1598, 0.1073, 0.1507, 0.0625, 0.0924, 0.1917, 0.0520], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0203, 0.0191, 0.0189, 0.0176, 0.0213, 0.0214, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:44:16,561 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=75299.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:44:24,289 INFO [finetune.py:976] (4/7) Epoch 14, batch 850, loss[loss=0.2116, simple_loss=0.2677, pruned_loss=0.07769, over 4899.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2565, pruned_loss=0.06054, over 943964.51 frames. ], batch size: 36, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:44:30,260 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75320.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:44:37,463 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=75331.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 16:44:38,013 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.2513, 4.6349, 4.8769, 4.7753, 4.7256, 4.6060, 5.3917, 1.6383], + device='cuda:4'), covar=tensor([0.0972, 0.1593, 0.1225, 0.1983, 0.1623, 0.1941, 0.0768, 0.7787], + device='cuda:4'), in_proj_covar=tensor([0.0349, 0.0244, 0.0276, 0.0290, 0.0332, 0.0282, 0.0300, 0.0297], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:44:44,946 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.704e+01 1.586e+02 1.985e+02 2.275e+02 3.825e+02, threshold=3.970e+02, percent-clipped=2.0 +2023-03-26 16:44:57,420 INFO [finetune.py:976] (4/7) Epoch 14, batch 900, loss[loss=0.1816, simple_loss=0.2522, pruned_loss=0.05544, over 4822.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2526, pruned_loss=0.05877, over 947589.39 frames. ], batch size: 30, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:45:44,929 INFO [finetune.py:976] (4/7) Epoch 14, batch 950, loss[loss=0.1824, simple_loss=0.2418, pruned_loss=0.06148, over 4053.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2507, pruned_loss=0.05823, over 949410.75 frames. ], batch size: 17, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:45:50,484 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.5368, 1.4106, 1.4578, 0.8213, 1.4936, 1.6757, 1.6815, 1.3365], + device='cuda:4'), covar=tensor([0.0785, 0.0531, 0.0448, 0.0510, 0.0367, 0.0506, 0.0308, 0.0611], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0153, 0.0122, 0.0129, 0.0131, 0.0127, 0.0142, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.2945e-05, 1.1141e-04, 8.7772e-05, 9.2910e-05, 9.2785e-05, 9.1585e-05, + 1.0336e-04, 1.0588e-04], device='cuda:4') +2023-03-26 16:45:52,293 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.61 vs. limit=2.0 +2023-03-26 16:46:05,783 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.942e+01 1.566e+02 1.871e+02 2.243e+02 4.539e+02, threshold=3.743e+02, percent-clipped=1.0 +2023-03-26 16:46:18,860 INFO [finetune.py:976] (4/7) Epoch 14, batch 1000, loss[loss=0.1777, simple_loss=0.2494, pruned_loss=0.05298, over 4922.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2537, pruned_loss=0.05963, over 950860.37 frames. ], batch size: 37, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:47:07,213 INFO [finetune.py:976] (4/7) Epoch 14, batch 1050, loss[loss=0.1968, simple_loss=0.2758, pruned_loss=0.05887, over 4933.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2565, pruned_loss=0.06003, over 954979.31 frames. ], batch size: 38, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:47:31,085 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.109e+01 1.606e+02 2.003e+02 2.356e+02 8.983e+02, threshold=4.007e+02, percent-clipped=2.0 +2023-03-26 16:47:44,012 INFO [finetune.py:976] (4/7) Epoch 14, batch 1100, loss[loss=0.1825, simple_loss=0.2458, pruned_loss=0.0596, over 4863.00 frames. ], tot_loss[loss=0.189, simple_loss=0.2573, pruned_loss=0.06032, over 953763.46 frames. ], batch size: 34, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:47:47,088 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75565.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:48:00,099 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75584.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:48:18,071 INFO [finetune.py:976] (4/7) Epoch 14, batch 1150, loss[loss=0.1654, simple_loss=0.2171, pruned_loss=0.05681, over 4315.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2585, pruned_loss=0.06127, over 951850.75 frames. ], batch size: 18, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:48:19,336 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=75613.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:48:24,070 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75620.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:48:26,544 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6242, 1.6922, 1.4273, 1.6784, 2.1184, 2.0492, 1.7320, 1.5126], + device='cuda:4'), covar=tensor([0.0336, 0.0306, 0.0531, 0.0288, 0.0176, 0.0373, 0.0334, 0.0394], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0108, 0.0139, 0.0112, 0.0100, 0.0105, 0.0095, 0.0107], + device='cuda:4'), out_proj_covar=tensor([7.2545e-05, 8.3752e-05, 1.1026e-04, 8.7499e-05, 7.7743e-05, 7.7705e-05, + 7.1518e-05, 8.2276e-05], device='cuda:4') +2023-03-26 16:48:28,195 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75626.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 16:48:38,762 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.806e+01 1.586e+02 1.986e+02 2.337e+02 5.787e+02, threshold=3.972e+02, percent-clipped=2.0 +2023-03-26 16:48:51,184 INFO [finetune.py:976] (4/7) Epoch 14, batch 1200, loss[loss=0.1669, simple_loss=0.2369, pruned_loss=0.04847, over 4876.00 frames. ], tot_loss[loss=0.188, simple_loss=0.256, pruned_loss=0.06004, over 952647.41 frames. ], batch size: 34, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:48:56,399 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=75668.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:49:24,700 INFO [finetune.py:976] (4/7) Epoch 14, batch 1250, loss[loss=0.1686, simple_loss=0.2389, pruned_loss=0.0492, over 4894.00 frames. ], tot_loss[loss=0.1858, simple_loss=0.2531, pruned_loss=0.0592, over 952948.40 frames. ], batch size: 32, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:49:27,423 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 16:49:32,441 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8203, 1.7054, 2.1499, 1.3213, 1.9402, 2.1276, 1.6018, 2.3429], + device='cuda:4'), covar=tensor([0.1280, 0.2002, 0.1272, 0.2064, 0.0864, 0.1334, 0.2655, 0.0819], + device='cuda:4'), in_proj_covar=tensor([0.0194, 0.0204, 0.0191, 0.0189, 0.0175, 0.0213, 0.0214, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:49:43,527 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5742, 1.4211, 1.3154, 1.5873, 1.6086, 1.5814, 0.9325, 1.3125], + device='cuda:4'), covar=tensor([0.1955, 0.1943, 0.1798, 0.1557, 0.1512, 0.1147, 0.2516, 0.1781], + device='cuda:4'), in_proj_covar=tensor([0.0238, 0.0206, 0.0209, 0.0190, 0.0239, 0.0183, 0.0212, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:49:45,241 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.316e+01 1.497e+02 1.829e+02 2.293e+02 4.240e+02, threshold=3.659e+02, percent-clipped=2.0 +2023-03-26 16:49:57,809 INFO [finetune.py:976] (4/7) Epoch 14, batch 1300, loss[loss=0.1807, simple_loss=0.2421, pruned_loss=0.05963, over 4896.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.25, pruned_loss=0.05823, over 953495.90 frames. ], batch size: 32, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:50:05,544 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.7516, 1.5000, 1.4675, 0.7358, 1.6068, 1.6710, 1.7127, 1.4138], + device='cuda:4'), covar=tensor([0.0701, 0.0562, 0.0495, 0.0517, 0.0424, 0.0599, 0.0337, 0.0609], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0152, 0.0121, 0.0129, 0.0130, 0.0127, 0.0142, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.2939e-05, 1.1121e-04, 8.7367e-05, 9.2700e-05, 9.2411e-05, 9.1565e-05, + 1.0281e-04, 1.0537e-04], device='cuda:4') +2023-03-26 16:50:31,736 INFO [finetune.py:976] (4/7) Epoch 14, batch 1350, loss[loss=0.2022, simple_loss=0.2704, pruned_loss=0.06699, over 4906.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2518, pruned_loss=0.05926, over 954818.40 frames. ], batch size: 37, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:51:07,733 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.659e+01 1.636e+02 1.953e+02 2.256e+02 6.748e+02, threshold=3.906e+02, percent-clipped=1.0 +2023-03-26 16:51:19,719 INFO [finetune.py:976] (4/7) Epoch 14, batch 1400, loss[loss=0.2388, simple_loss=0.2974, pruned_loss=0.09007, over 4731.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2548, pruned_loss=0.05991, over 955000.95 frames. ], batch size: 59, lr: 3.56e-03, grad_scale: 16.0 +2023-03-26 16:51:35,778 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75884.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:51:43,569 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8182, 0.9838, 1.8136, 1.7451, 1.6172, 1.5191, 1.6566, 1.6703], + device='cuda:4'), covar=tensor([0.3668, 0.4007, 0.3156, 0.3466, 0.4648, 0.3498, 0.4035, 0.3195], + device='cuda:4'), in_proj_covar=tensor([0.0241, 0.0237, 0.0255, 0.0264, 0.0261, 0.0235, 0.0275, 0.0234], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:51:53,580 INFO [finetune.py:976] (4/7) Epoch 14, batch 1450, loss[loss=0.1814, simple_loss=0.2556, pruned_loss=0.05358, over 4931.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2575, pruned_loss=0.06121, over 953705.89 frames. ], batch size: 42, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:52:08,085 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75926.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 16:52:17,403 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=75932.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:52:27,769 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.664e+02 1.939e+02 2.609e+02 1.085e+03, threshold=3.877e+02, percent-clipped=5.0 +2023-03-26 16:52:44,457 INFO [finetune.py:976] (4/7) Epoch 14, batch 1500, loss[loss=0.183, simple_loss=0.252, pruned_loss=0.05701, over 4922.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2591, pruned_loss=0.06193, over 952791.08 frames. ], batch size: 42, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:52:52,977 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=75974.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:53:19,428 INFO [finetune.py:976] (4/7) Epoch 14, batch 1550, loss[loss=0.2164, simple_loss=0.2726, pruned_loss=0.08011, over 4808.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2586, pruned_loss=0.06146, over 950881.23 frames. ], batch size: 40, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:53:23,423 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 16:53:40,209 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.569e+01 1.489e+02 1.761e+02 2.263e+02 3.823e+02, threshold=3.522e+02, percent-clipped=0.0 +2023-03-26 16:53:51,517 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.5064, 1.5183, 1.6099, 0.8210, 1.5944, 1.8198, 1.8373, 1.4276], + device='cuda:4'), covar=tensor([0.0871, 0.0583, 0.0432, 0.0573, 0.0393, 0.0572, 0.0275, 0.0684], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0153, 0.0123, 0.0130, 0.0131, 0.0127, 0.0142, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.3266e-05, 1.1174e-04, 8.8314e-05, 9.3129e-05, 9.3095e-05, 9.2209e-05, + 1.0344e-04, 1.0595e-04], device='cuda:4') +2023-03-26 16:53:53,247 INFO [finetune.py:976] (4/7) Epoch 14, batch 1600, loss[loss=0.1626, simple_loss=0.2429, pruned_loss=0.04117, over 4769.00 frames. ], tot_loss[loss=0.1887, simple_loss=0.2568, pruned_loss=0.06028, over 951797.74 frames. ], batch size: 26, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:54:09,467 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4472, 1.6144, 1.2653, 1.5680, 1.8633, 1.7325, 1.4978, 1.4300], + device='cuda:4'), covar=tensor([0.0351, 0.0299, 0.0549, 0.0289, 0.0228, 0.0461, 0.0323, 0.0346], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0109, 0.0141, 0.0113, 0.0101, 0.0106, 0.0096, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.3024e-05, 8.4528e-05, 1.1124e-04, 8.8137e-05, 7.8425e-05, 7.8572e-05, + 7.2170e-05, 8.2603e-05], device='cuda:4') +2023-03-26 16:54:26,637 INFO [finetune.py:976] (4/7) Epoch 14, batch 1650, loss[loss=0.1975, simple_loss=0.2604, pruned_loss=0.06734, over 4908.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.2549, pruned_loss=0.05967, over 953382.20 frames. ], batch size: 32, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:54:47,811 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.271e+01 1.580e+02 1.872e+02 2.187e+02 4.946e+02, threshold=3.744e+02, percent-clipped=3.0 +2023-03-26 16:55:00,264 INFO [finetune.py:976] (4/7) Epoch 14, batch 1700, loss[loss=0.1608, simple_loss=0.2222, pruned_loss=0.04966, over 4826.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2525, pruned_loss=0.05912, over 953261.74 frames. ], batch size: 30, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:55:34,234 INFO [finetune.py:976] (4/7) Epoch 14, batch 1750, loss[loss=0.1853, simple_loss=0.2451, pruned_loss=0.06274, over 4718.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2549, pruned_loss=0.06059, over 953186.45 frames. ], batch size: 23, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:55:55,253 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.620e+02 1.973e+02 2.349e+02 4.562e+02, threshold=3.945e+02, percent-clipped=4.0 +2023-03-26 16:56:17,746 INFO [finetune.py:976] (4/7) Epoch 14, batch 1800, loss[loss=0.1887, simple_loss=0.2501, pruned_loss=0.06361, over 4825.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2575, pruned_loss=0.06082, over 952674.27 frames. ], batch size: 30, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:56:27,630 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76269.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 16:56:47,389 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76294.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:56:58,602 INFO [finetune.py:976] (4/7) Epoch 14, batch 1850, loss[loss=0.1856, simple_loss=0.251, pruned_loss=0.0601, over 4880.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2585, pruned_loss=0.06137, over 954430.24 frames. ], batch size: 32, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:57:07,116 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76323.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 16:57:07,807 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-03-26 16:57:11,398 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76330.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 16:57:19,099 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.514e+02 1.907e+02 2.299e+02 3.483e+02, threshold=3.815e+02, percent-clipped=0.0 +2023-03-26 16:57:20,127 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-03-26 16:57:35,323 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76355.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 16:57:38,873 INFO [finetune.py:976] (4/7) Epoch 14, batch 1900, loss[loss=0.1956, simple_loss=0.2491, pruned_loss=0.07103, over 4765.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.259, pruned_loss=0.06126, over 956012.16 frames. ], batch size: 28, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:57:57,156 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76384.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 16:58:15,527 INFO [finetune.py:976] (4/7) Epoch 14, batch 1950, loss[loss=0.2043, simple_loss=0.2695, pruned_loss=0.06953, over 4895.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2574, pruned_loss=0.06023, over 957165.02 frames. ], batch size: 43, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:58:35,772 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.938e+01 1.459e+02 1.786e+02 2.082e+02 3.715e+02, threshold=3.572e+02, percent-clipped=0.0 +2023-03-26 16:58:49,142 INFO [finetune.py:976] (4/7) Epoch 14, batch 2000, loss[loss=0.2089, simple_loss=0.2632, pruned_loss=0.07728, over 4789.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2544, pruned_loss=0.05931, over 956312.12 frames. ], batch size: 29, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:58:49,232 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1742, 3.6659, 3.8278, 3.9410, 3.9753, 3.7817, 4.3023, 1.4368], + device='cuda:4'), covar=tensor([0.0803, 0.0804, 0.0757, 0.1116, 0.1139, 0.1373, 0.0695, 0.5117], + device='cuda:4'), in_proj_covar=tensor([0.0349, 0.0244, 0.0277, 0.0292, 0.0332, 0.0281, 0.0302, 0.0297], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 16:58:53,204 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.36 vs. limit=5.0 +2023-03-26 16:58:54,820 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.47 vs. limit=5.0 +2023-03-26 16:59:22,662 INFO [finetune.py:976] (4/7) Epoch 14, batch 2050, loss[loss=0.2303, simple_loss=0.2873, pruned_loss=0.08664, over 4909.00 frames. ], tot_loss[loss=0.1838, simple_loss=0.2512, pruned_loss=0.0582, over 956792.91 frames. ], batch size: 43, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 16:59:42,966 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.833e+01 1.460e+02 1.798e+02 2.157e+02 5.136e+02, threshold=3.595e+02, percent-clipped=3.0 +2023-03-26 16:59:50,500 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-26 16:59:56,047 INFO [finetune.py:976] (4/7) Epoch 14, batch 2100, loss[loss=0.1334, simple_loss=0.208, pruned_loss=0.02939, over 4744.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.2506, pruned_loss=0.05796, over 957005.41 frames. ], batch size: 26, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:00:01,054 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-26 17:00:21,825 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 17:00:28,652 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 17:00:29,582 INFO [finetune.py:976] (4/7) Epoch 14, batch 2150, loss[loss=0.1934, simple_loss=0.2731, pruned_loss=0.05682, over 4902.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2551, pruned_loss=0.05966, over 956223.46 frames. ], batch size: 37, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:00:36,339 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1774, 2.1495, 2.3593, 1.6292, 2.1940, 2.4348, 2.3138, 1.9522], + device='cuda:4'), covar=tensor([0.0609, 0.0633, 0.0619, 0.0852, 0.0691, 0.0612, 0.0606, 0.0945], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0132, 0.0141, 0.0123, 0.0124, 0.0141, 0.0140, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:00:38,773 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=76625.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 17:00:44,040 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-26 17:00:50,411 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.680e+02 1.855e+02 2.274e+02 3.771e+02, threshold=3.710e+02, percent-clipped=2.0 +2023-03-26 17:00:54,845 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2355, 2.2210, 2.2465, 1.0454, 2.4847, 2.6512, 2.2959, 2.0696], + device='cuda:4'), covar=tensor([0.0901, 0.0673, 0.0503, 0.0732, 0.0594, 0.0850, 0.0482, 0.0688], + device='cuda:4'), in_proj_covar=tensor([0.0125, 0.0152, 0.0121, 0.0129, 0.0130, 0.0126, 0.0141, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.2606e-05, 1.1089e-04, 8.7019e-05, 9.2710e-05, 9.1965e-05, 9.0963e-05, + 1.0235e-04, 1.0504e-04], device='cuda:4') +2023-03-26 17:00:55,368 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=76650.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:01:02,483 INFO [finetune.py:976] (4/7) Epoch 14, batch 2200, loss[loss=0.2119, simple_loss=0.2663, pruned_loss=0.07873, over 4794.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2566, pruned_loss=0.05999, over 954002.73 frames. ], batch size: 26, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:01:21,648 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=76679.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 17:01:57,531 INFO [finetune.py:976] (4/7) Epoch 14, batch 2250, loss[loss=0.2116, simple_loss=0.2775, pruned_loss=0.07289, over 4741.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2585, pruned_loss=0.06102, over 952550.79 frames. ], batch size: 54, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:02:09,952 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0866, 1.8980, 1.6799, 1.8598, 1.7935, 1.8247, 1.8505, 2.5727], + device='cuda:4'), covar=tensor([0.4058, 0.4916, 0.3362, 0.4402, 0.4482, 0.2437, 0.4303, 0.1735], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0259, 0.0224, 0.0276, 0.0247, 0.0214, 0.0248, 0.0222], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:02:14,563 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-26 17:02:18,738 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.580e+02 1.848e+02 2.151e+02 3.368e+02, threshold=3.695e+02, percent-clipped=0.0 +2023-03-26 17:02:31,273 INFO [finetune.py:976] (4/7) Epoch 14, batch 2300, loss[loss=0.1892, simple_loss=0.2696, pruned_loss=0.05444, over 4771.00 frames. ], tot_loss[loss=0.1899, simple_loss=0.258, pruned_loss=0.06094, over 952389.45 frames. ], batch size: 26, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:02:36,709 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 17:03:06,756 INFO [finetune.py:976] (4/7) Epoch 14, batch 2350, loss[loss=0.1773, simple_loss=0.2436, pruned_loss=0.05553, over 4817.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2556, pruned_loss=0.05994, over 952637.15 frames. ], batch size: 41, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:03:25,951 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76839.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:03:28,177 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.075e+02 1.531e+02 1.863e+02 2.217e+02 4.521e+02, threshold=3.725e+02, percent-clipped=2.0 +2023-03-26 17:03:34,096 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6271, 1.4971, 1.5314, 1.4792, 0.9922, 2.8869, 1.1539, 1.5640], + device='cuda:4'), covar=tensor([0.3350, 0.2438, 0.2164, 0.2609, 0.1993, 0.0247, 0.2679, 0.1296], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0116, 0.0121, 0.0125, 0.0116, 0.0098, 0.0098, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 17:03:35,912 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3249, 2.1093, 2.2716, 0.9866, 2.5151, 2.6778, 2.2812, 2.0182], + device='cuda:4'), covar=tensor([0.0999, 0.0784, 0.0519, 0.0806, 0.0466, 0.0765, 0.0502, 0.0773], + device='cuda:4'), in_proj_covar=tensor([0.0127, 0.0154, 0.0123, 0.0130, 0.0132, 0.0128, 0.0143, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.4154e-05, 1.1275e-04, 8.8292e-05, 9.3793e-05, 9.3434e-05, 9.2556e-05, + 1.0347e-04, 1.0599e-04], device='cuda:4') +2023-03-26 17:03:40,659 INFO [finetune.py:976] (4/7) Epoch 14, batch 2400, loss[loss=0.199, simple_loss=0.2637, pruned_loss=0.06719, over 4753.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2527, pruned_loss=0.05891, over 952169.22 frames. ], batch size: 27, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:03:49,670 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3888, 1.3725, 1.5149, 0.8411, 1.4582, 1.6822, 1.7414, 1.3678], + device='cuda:4'), covar=tensor([0.0741, 0.0532, 0.0406, 0.0409, 0.0393, 0.0464, 0.0280, 0.0548], + device='cuda:4'), in_proj_covar=tensor([0.0127, 0.0154, 0.0123, 0.0130, 0.0132, 0.0128, 0.0143, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.4126e-05, 1.1268e-04, 8.8242e-05, 9.3719e-05, 9.3455e-05, 9.2558e-05, + 1.0345e-04, 1.0591e-04], device='cuda:4') +2023-03-26 17:03:52,108 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.71 vs. limit=2.0 +2023-03-26 17:04:06,926 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76900.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:04:14,026 INFO [finetune.py:976] (4/7) Epoch 14, batch 2450, loss[loss=0.1911, simple_loss=0.2355, pruned_loss=0.07331, over 3701.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.2503, pruned_loss=0.05821, over 952398.89 frames. ], batch size: 16, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:04:15,968 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6600, 1.5347, 1.4957, 1.7675, 1.7927, 1.8282, 1.3129, 1.4711], + device='cuda:4'), covar=tensor([0.1950, 0.1857, 0.1773, 0.1461, 0.1562, 0.0979, 0.2350, 0.1774], + device='cuda:4'), in_proj_covar=tensor([0.0240, 0.0208, 0.0212, 0.0191, 0.0242, 0.0185, 0.0215, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:04:23,109 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=76925.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 17:04:34,661 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.650e+01 1.627e+02 1.914e+02 2.448e+02 4.488e+02, threshold=3.829e+02, percent-clipped=3.0 +2023-03-26 17:04:40,118 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=76950.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:04:47,675 INFO [finetune.py:976] (4/7) Epoch 14, batch 2500, loss[loss=0.2561, simple_loss=0.3229, pruned_loss=0.09459, over 4838.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2515, pruned_loss=0.0589, over 949210.76 frames. ], batch size: 47, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:04:55,526 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=76973.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 17:04:59,667 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=76979.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 17:05:05,951 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-26 17:05:12,271 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=76998.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:05:21,618 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 17:05:21,741 INFO [finetune.py:976] (4/7) Epoch 14, batch 2550, loss[loss=0.1877, simple_loss=0.2645, pruned_loss=0.05542, over 4815.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.2548, pruned_loss=0.05928, over 951020.79 frames. ], batch size: 38, lr: 3.55e-03, grad_scale: 16.0 +2023-03-26 17:05:32,005 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77027.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 17:05:42,423 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.599e+02 1.863e+02 2.531e+02 5.028e+02, threshold=3.725e+02, percent-clipped=2.0 +2023-03-26 17:05:43,155 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6789, 1.5873, 1.4850, 1.6189, 1.0105, 3.7069, 1.4566, 1.8208], + device='cuda:4'), covar=tensor([0.3374, 0.2589, 0.2180, 0.2382, 0.1900, 0.0159, 0.2694, 0.1328], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0116, 0.0121, 0.0124, 0.0115, 0.0097, 0.0097, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 17:05:43,768 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3634, 1.6402, 1.3277, 1.4447, 1.7823, 1.7726, 1.6759, 1.6319], + device='cuda:4'), covar=tensor([0.0561, 0.0348, 0.0564, 0.0351, 0.0363, 0.0642, 0.0327, 0.0407], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0109, 0.0141, 0.0113, 0.0100, 0.0106, 0.0096, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.3022e-05, 8.4177e-05, 1.1151e-04, 8.7678e-05, 7.8349e-05, 7.8346e-05, + 7.2475e-05, 8.2201e-05], device='cuda:4') +2023-03-26 17:05:55,387 INFO [finetune.py:976] (4/7) Epoch 14, batch 2600, loss[loss=0.2075, simple_loss=0.2876, pruned_loss=0.06364, over 4828.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2557, pruned_loss=0.05939, over 951125.05 frames. ], batch size: 30, lr: 3.55e-03, grad_scale: 32.0 +2023-03-26 17:06:15,300 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-26 17:06:18,037 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77095.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:06:27,076 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1212, 1.7447, 2.4746, 1.6368, 2.2245, 2.4018, 1.7008, 2.5973], + device='cuda:4'), covar=tensor([0.1326, 0.1973, 0.1309, 0.1894, 0.0889, 0.1323, 0.2629, 0.0736], + device='cuda:4'), in_proj_covar=tensor([0.0196, 0.0206, 0.0193, 0.0190, 0.0178, 0.0214, 0.0217, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:06:30,488 INFO [finetune.py:976] (4/7) Epoch 14, batch 2650, loss[loss=0.1801, simple_loss=0.2629, pruned_loss=0.04867, over 4817.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2576, pruned_loss=0.05979, over 952973.20 frames. ], batch size: 38, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:07:08,845 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.161e+02 1.606e+02 1.948e+02 2.371e+02 3.624e+02, threshold=3.895e+02, percent-clipped=0.0 +2023-03-26 17:07:22,920 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77156.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 17:07:26,328 INFO [finetune.py:976] (4/7) Epoch 14, batch 2700, loss[loss=0.1511, simple_loss=0.2214, pruned_loss=0.04034, over 4919.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2564, pruned_loss=0.05945, over 952394.49 frames. ], batch size: 38, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:07:41,725 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 17:07:57,703 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77195.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:08:04,073 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.69 vs. limit=5.0 +2023-03-26 17:08:07,958 INFO [finetune.py:976] (4/7) Epoch 14, batch 2750, loss[loss=0.1594, simple_loss=0.2323, pruned_loss=0.04324, over 4827.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.254, pruned_loss=0.05916, over 953193.21 frames. ], batch size: 41, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:08:08,071 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77211.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:08:08,694 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5639, 1.3764, 1.2613, 1.5402, 1.6125, 1.5659, 0.8912, 1.2920], + device='cuda:4'), covar=tensor([0.2222, 0.2149, 0.2069, 0.1746, 0.1681, 0.1315, 0.2714, 0.1936], + device='cuda:4'), in_proj_covar=tensor([0.0239, 0.0207, 0.0211, 0.0191, 0.0241, 0.0184, 0.0214, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:08:15,192 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1557, 1.9940, 1.4261, 0.5095, 1.6675, 1.7944, 1.6017, 1.7235], + device='cuda:4'), covar=tensor([0.0809, 0.0716, 0.1353, 0.1894, 0.1382, 0.1858, 0.2059, 0.0925], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0197, 0.0201, 0.0186, 0.0216, 0.0210, 0.0225, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:08:18,248 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6623, 2.4326, 1.9395, 0.8375, 2.1583, 2.1030, 1.8492, 2.0718], + device='cuda:4'), covar=tensor([0.0817, 0.0879, 0.1614, 0.2166, 0.1435, 0.1979, 0.2219, 0.1052], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0198, 0.0201, 0.0186, 0.0216, 0.0210, 0.0225, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:08:28,394 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.127e+02 1.514e+02 1.871e+02 2.163e+02 3.576e+02, threshold=3.742e+02, percent-clipped=0.0 +2023-03-26 17:08:40,953 INFO [finetune.py:976] (4/7) Epoch 14, batch 2800, loss[loss=0.1381, simple_loss=0.2171, pruned_loss=0.02951, over 4774.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2502, pruned_loss=0.05747, over 954225.96 frames. ], batch size: 28, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:08:45,231 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2842, 2.1975, 1.7174, 2.0580, 2.1990, 1.9267, 2.4964, 2.2578], + device='cuda:4'), covar=tensor([0.1408, 0.2293, 0.3504, 0.2965, 0.2796, 0.1783, 0.3108, 0.1976], + device='cuda:4'), in_proj_covar=tensor([0.0180, 0.0188, 0.0233, 0.0253, 0.0244, 0.0199, 0.0213, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:08:48,274 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77272.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:08:48,853 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77273.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:08:52,185 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77277.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:09:14,630 INFO [finetune.py:976] (4/7) Epoch 14, batch 2850, loss[loss=0.1529, simple_loss=0.2196, pruned_loss=0.04313, over 4744.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.249, pruned_loss=0.05744, over 953827.04 frames. ], batch size: 27, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:09:15,016 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.03 vs. limit=5.0 +2023-03-26 17:09:20,317 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 17:09:29,758 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77334.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:09:32,162 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77338.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:09:34,910 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.563e+02 1.884e+02 2.320e+02 5.201e+02, threshold=3.768e+02, percent-clipped=2.0 +2023-03-26 17:09:42,272 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 17:09:47,965 INFO [finetune.py:976] (4/7) Epoch 14, batch 2900, loss[loss=0.2274, simple_loss=0.2949, pruned_loss=0.07996, over 4860.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2545, pruned_loss=0.05992, over 955183.72 frames. ], batch size: 44, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:10:21,782 INFO [finetune.py:976] (4/7) Epoch 14, batch 2950, loss[loss=0.2402, simple_loss=0.3062, pruned_loss=0.08714, over 4809.00 frames. ], tot_loss[loss=0.1887, simple_loss=0.2566, pruned_loss=0.06038, over 953919.53 frames. ], batch size: 45, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:10:41,988 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.681e+01 1.651e+02 1.924e+02 2.203e+02 4.754e+02, threshold=3.848e+02, percent-clipped=2.0 +2023-03-26 17:10:48,024 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77451.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 17:10:54,979 INFO [finetune.py:976] (4/7) Epoch 14, batch 3000, loss[loss=0.2036, simple_loss=0.2568, pruned_loss=0.07522, over 4727.00 frames. ], tot_loss[loss=0.1899, simple_loss=0.2578, pruned_loss=0.06095, over 953858.91 frames. ], batch size: 23, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:10:54,980 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 17:11:00,273 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7968, 1.8161, 1.8930, 1.1289, 1.9984, 1.9734, 1.9321, 1.6705], + device='cuda:4'), covar=tensor([0.0575, 0.0663, 0.0667, 0.0928, 0.0850, 0.0659, 0.0625, 0.1043], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0131, 0.0139, 0.0122, 0.0122, 0.0139, 0.0139, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:11:03,751 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8716, 1.0950, 1.9554, 1.8365, 1.6935, 1.5719, 1.6541, 1.7605], + device='cuda:4'), covar=tensor([0.3913, 0.4570, 0.3778, 0.3977, 0.5380, 0.3974, 0.5104, 0.3499], + device='cuda:4'), in_proj_covar=tensor([0.0242, 0.0237, 0.0256, 0.0265, 0.0263, 0.0236, 0.0276, 0.0235], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:11:09,361 INFO [finetune.py:1010] (4/7) Epoch 14, validation: loss=0.1563, simple_loss=0.2268, pruned_loss=0.04293, over 2265189.00 frames. +2023-03-26 17:11:09,362 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 17:11:10,421 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-03-26 17:11:34,070 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77495.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:11:44,251 INFO [finetune.py:976] (4/7) Epoch 14, batch 3050, loss[loss=0.1654, simple_loss=0.2451, pruned_loss=0.04287, over 4849.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2584, pruned_loss=0.06134, over 950960.03 frames. ], batch size: 44, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:12:02,290 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-26 17:12:13,226 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.567e+02 1.800e+02 2.244e+02 5.193e+02, threshold=3.600e+02, percent-clipped=2.0 +2023-03-26 17:12:13,792 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 17:12:13,931 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77543.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:12:29,445 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1410, 2.0125, 1.7127, 2.0785, 1.9208, 1.9038, 1.9237, 2.6635], + device='cuda:4'), covar=tensor([0.3855, 0.4627, 0.3430, 0.4056, 0.4157, 0.2513, 0.4225, 0.1585], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0259, 0.0225, 0.0277, 0.0246, 0.0214, 0.0249, 0.0223], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:12:35,684 INFO [finetune.py:976] (4/7) Epoch 14, batch 3100, loss[loss=0.16, simple_loss=0.2357, pruned_loss=0.04214, over 4858.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2566, pruned_loss=0.06048, over 953115.71 frames. ], batch size: 47, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:12:39,910 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77567.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:13:08,683 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2137, 2.3021, 2.7328, 2.6344, 2.4699, 4.8316, 2.2042, 2.5082], + device='cuda:4'), covar=tensor([0.0847, 0.1399, 0.0840, 0.0810, 0.1235, 0.0121, 0.1199, 0.1340], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0073, 0.0077, 0.0091, 0.0080, 0.0084, 0.0078], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 17:13:22,408 INFO [finetune.py:976] (4/7) Epoch 14, batch 3150, loss[loss=0.1571, simple_loss=0.224, pruned_loss=0.04511, over 4824.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2543, pruned_loss=0.06005, over 951212.78 frames. ], batch size: 25, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:13:25,544 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77616.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:13:35,442 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77629.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:13:37,886 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77633.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:13:43,294 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.167e+02 1.643e+02 1.968e+02 2.398e+02 4.679e+02, threshold=3.936e+02, percent-clipped=3.0 +2023-03-26 17:13:48,725 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0994, 1.8038, 2.6074, 4.1997, 2.8626, 2.8297, 0.7026, 3.4713], + device='cuda:4'), covar=tensor([0.1681, 0.1484, 0.1284, 0.0440, 0.0715, 0.1215, 0.2231, 0.0436], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0116, 0.0133, 0.0164, 0.0100, 0.0137, 0.0125, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 17:13:56,370 INFO [finetune.py:976] (4/7) Epoch 14, batch 3200, loss[loss=0.179, simple_loss=0.2355, pruned_loss=0.06127, over 4834.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.2508, pruned_loss=0.05882, over 953179.70 frames. ], batch size: 30, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:14:06,668 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77677.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:14:06,682 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77677.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:14:29,514 INFO [finetune.py:976] (4/7) Epoch 14, batch 3250, loss[loss=0.1966, simple_loss=0.276, pruned_loss=0.05856, over 4821.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2517, pruned_loss=0.05911, over 954529.58 frames. ], batch size: 39, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:14:36,702 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7874, 1.7715, 1.8459, 1.1631, 1.8777, 1.9551, 1.9413, 1.6147], + device='cuda:4'), covar=tensor([0.0549, 0.0662, 0.0635, 0.0876, 0.0665, 0.0661, 0.0584, 0.1110], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0132, 0.0140, 0.0123, 0.0123, 0.0140, 0.0140, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:14:47,327 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77738.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:14:49,602 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.063e+02 1.623e+02 1.885e+02 2.287e+02 7.301e+02, threshold=3.769e+02, percent-clipped=4.0 +2023-03-26 17:14:55,592 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77751.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 17:15:02,067 INFO [finetune.py:976] (4/7) Epoch 14, batch 3300, loss[loss=0.2363, simple_loss=0.2987, pruned_loss=0.08693, over 4903.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2553, pruned_loss=0.06047, over 954799.55 frames. ], batch size: 35, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:15:14,145 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4197, 1.4091, 1.4810, 0.7975, 1.5077, 1.4492, 1.5073, 1.3483], + device='cuda:4'), covar=tensor([0.0629, 0.0783, 0.0745, 0.0988, 0.0834, 0.0805, 0.0670, 0.1244], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0132, 0.0140, 0.0123, 0.0123, 0.0140, 0.0140, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:15:27,675 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77799.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:15:35,623 INFO [finetune.py:976] (4/7) Epoch 14, batch 3350, loss[loss=0.2194, simple_loss=0.2761, pruned_loss=0.08133, over 4903.00 frames. ], tot_loss[loss=0.189, simple_loss=0.2563, pruned_loss=0.06082, over 954826.80 frames. ], batch size: 35, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:15:41,514 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.6078, 3.2350, 3.3650, 3.2951, 3.1869, 3.1535, 3.7229, 1.2047], + device='cuda:4'), covar=tensor([0.1311, 0.1861, 0.1677, 0.2072, 0.2398, 0.2506, 0.1442, 0.7107], + device='cuda:4'), in_proj_covar=tensor([0.0348, 0.0244, 0.0277, 0.0294, 0.0333, 0.0284, 0.0300, 0.0298], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:15:57,267 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.144e+02 1.683e+02 1.958e+02 2.277e+02 5.309e+02, threshold=3.915e+02, percent-clipped=1.0 +2023-03-26 17:16:09,330 INFO [finetune.py:976] (4/7) Epoch 14, batch 3400, loss[loss=0.2376, simple_loss=0.301, pruned_loss=0.08712, over 4712.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2581, pruned_loss=0.06206, over 953345.38 frames. ], batch size: 59, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:16:18,552 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77867.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:16:34,121 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-26 17:16:51,413 INFO [finetune.py:976] (4/7) Epoch 14, batch 3450, loss[loss=0.2106, simple_loss=0.2782, pruned_loss=0.0715, over 4815.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.259, pruned_loss=0.062, over 953285.21 frames. ], batch size: 38, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:16:53,842 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77915.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:17:01,735 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-26 17:17:03,228 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77929.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:17:05,716 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77933.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:17:11,407 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.160e+01 1.485e+02 1.825e+02 2.093e+02 4.049e+02, threshold=3.650e+02, percent-clipped=1.0 +2023-03-26 17:17:33,219 INFO [finetune.py:976] (4/7) Epoch 14, batch 3500, loss[loss=0.1667, simple_loss=0.2367, pruned_loss=0.04833, over 4850.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2571, pruned_loss=0.06143, over 953028.02 frames. ], batch size: 49, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:17:35,605 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7405, 1.5318, 1.0768, 0.2677, 1.3242, 1.4983, 1.5013, 1.4242], + device='cuda:4'), covar=tensor([0.0896, 0.0800, 0.1339, 0.1931, 0.1456, 0.2455, 0.2273, 0.0891], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0195, 0.0200, 0.0184, 0.0214, 0.0208, 0.0223, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:17:40,922 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77972.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:17:44,502 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77977.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:17:46,960 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=77981.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:18:20,966 INFO [finetune.py:976] (4/7) Epoch 14, batch 3550, loss[loss=0.1505, simple_loss=0.2302, pruned_loss=0.03541, over 4821.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2546, pruned_loss=0.06105, over 953785.72 frames. ], batch size: 41, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:18:35,862 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78033.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:18:41,167 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.505e+02 1.960e+02 2.472e+02 4.194e+02, threshold=3.920e+02, percent-clipped=4.0 +2023-03-26 17:18:54,344 INFO [finetune.py:976] (4/7) Epoch 14, batch 3600, loss[loss=0.1496, simple_loss=0.2284, pruned_loss=0.03543, over 4752.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2519, pruned_loss=0.06, over 954570.27 frames. ], batch size: 28, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:19:28,417 INFO [finetune.py:976] (4/7) Epoch 14, batch 3650, loss[loss=0.1866, simple_loss=0.2586, pruned_loss=0.05728, over 4905.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2542, pruned_loss=0.06088, over 956131.17 frames. ], batch size: 37, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:19:38,472 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.57 vs. limit=5.0 +2023-03-26 17:19:48,729 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.050e+02 1.711e+02 2.033e+02 2.406e+02 8.151e+02, threshold=4.067e+02, percent-clipped=4.0 +2023-03-26 17:20:02,234 INFO [finetune.py:976] (4/7) Epoch 14, batch 3700, loss[loss=0.1821, simple_loss=0.2541, pruned_loss=0.05505, over 4784.00 frames. ], tot_loss[loss=0.1892, simple_loss=0.2563, pruned_loss=0.06102, over 955293.69 frames. ], batch size: 29, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:20:23,879 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.60 vs. limit=2.0 +2023-03-26 17:20:28,811 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7671, 1.0120, 1.7402, 1.7213, 1.5438, 1.4333, 1.6082, 1.6079], + device='cuda:4'), covar=tensor([0.3415, 0.3810, 0.3210, 0.3549, 0.4429, 0.3611, 0.4022, 0.3075], + device='cuda:4'), in_proj_covar=tensor([0.0244, 0.0239, 0.0256, 0.0266, 0.0263, 0.0237, 0.0278, 0.0235], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:20:35,986 INFO [finetune.py:976] (4/7) Epoch 14, batch 3750, loss[loss=0.213, simple_loss=0.2856, pruned_loss=0.07021, over 4742.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2575, pruned_loss=0.06121, over 955219.84 frames. ], batch size: 54, lr: 3.54e-03, grad_scale: 32.0 +2023-03-26 17:20:49,432 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78232.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:20:52,931 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0697, 2.0372, 1.6185, 2.0370, 2.0343, 1.7785, 2.4032, 2.1480], + device='cuda:4'), covar=tensor([0.1481, 0.2218, 0.3210, 0.2717, 0.2607, 0.1682, 0.3121, 0.1772], + device='cuda:4'), in_proj_covar=tensor([0.0181, 0.0188, 0.0236, 0.0255, 0.0245, 0.0200, 0.0213, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:20:55,808 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.186e+02 1.631e+02 1.949e+02 2.239e+02 4.423e+02, threshold=3.899e+02, percent-clipped=1.0 +2023-03-26 17:20:56,920 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-03-26 17:21:08,219 INFO [finetune.py:976] (4/7) Epoch 14, batch 3800, loss[loss=0.2061, simple_loss=0.2774, pruned_loss=0.06739, over 4841.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2592, pruned_loss=0.06194, over 955659.46 frames. ], batch size: 44, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:21:15,342 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-03-26 17:21:15,892 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=78272.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:21:20,270 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 17:21:31,047 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=78293.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 17:21:39,209 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1395, 2.0468, 1.6110, 2.2793, 2.0633, 1.8384, 2.5376, 2.1702], + device='cuda:4'), covar=tensor([0.1424, 0.2420, 0.3287, 0.2709, 0.2760, 0.1745, 0.3392, 0.1817], + device='cuda:4'), in_proj_covar=tensor([0.0182, 0.0189, 0.0237, 0.0256, 0.0247, 0.0201, 0.0214, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:21:49,336 INFO [finetune.py:976] (4/7) Epoch 14, batch 3850, loss[loss=0.1652, simple_loss=0.2297, pruned_loss=0.05031, over 4888.00 frames. ], tot_loss[loss=0.1895, simple_loss=0.2577, pruned_loss=0.06062, over 956382.20 frames. ], batch size: 35, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:21:55,867 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=78320.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:22:03,820 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=78333.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:22:10,190 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.518e+02 1.784e+02 2.158e+02 4.566e+02, threshold=3.568e+02, percent-clipped=2.0 +2023-03-26 17:22:19,334 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 17:22:21,573 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6289, 1.5216, 1.4677, 1.5203, 1.2295, 3.1433, 1.2483, 1.7126], + device='cuda:4'), covar=tensor([0.3374, 0.2495, 0.2085, 0.2338, 0.1718, 0.0220, 0.2845, 0.1233], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0115, 0.0121, 0.0124, 0.0116, 0.0097, 0.0097, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 17:22:22,704 INFO [finetune.py:976] (4/7) Epoch 14, batch 3900, loss[loss=0.1725, simple_loss=0.2469, pruned_loss=0.04901, over 4824.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2553, pruned_loss=0.06003, over 956521.40 frames. ], batch size: 25, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:22:36,856 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-26 17:22:45,627 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=78381.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:22:45,662 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2584, 2.0434, 2.4392, 4.2700, 3.0548, 2.9055, 0.9833, 3.4465], + device='cuda:4'), covar=tensor([0.1618, 0.1357, 0.1496, 0.0405, 0.0698, 0.1194, 0.2148, 0.0466], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0162, 0.0100, 0.0136, 0.0125, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 17:22:46,595 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.56 vs. limit=5.0 +2023-03-26 17:23:00,427 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-03-26 17:23:09,952 INFO [finetune.py:976] (4/7) Epoch 14, batch 3950, loss[loss=0.1744, simple_loss=0.2307, pruned_loss=0.05899, over 4868.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2518, pruned_loss=0.05897, over 954231.64 frames. ], batch size: 34, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:23:21,342 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7496, 1.5880, 1.5744, 1.6675, 1.0950, 3.3178, 1.3538, 1.7950], + device='cuda:4'), covar=tensor([0.3540, 0.2550, 0.2182, 0.2470, 0.1980, 0.0214, 0.2614, 0.1271], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0115, 0.0121, 0.0124, 0.0116, 0.0098, 0.0097, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 17:23:30,088 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4881, 1.4142, 1.5515, 1.6351, 1.5512, 2.8252, 1.3654, 1.5165], + device='cuda:4'), covar=tensor([0.0983, 0.1756, 0.1046, 0.0955, 0.1594, 0.0357, 0.1473, 0.1715], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 17:23:37,906 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.052e+02 1.587e+02 1.893e+02 2.259e+02 3.905e+02, threshold=3.786e+02, percent-clipped=1.0 +2023-03-26 17:23:50,382 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78460.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:23:50,892 INFO [finetune.py:976] (4/7) Epoch 14, batch 4000, loss[loss=0.2629, simple_loss=0.2977, pruned_loss=0.1141, over 4038.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2507, pruned_loss=0.05871, over 952575.84 frames. ], batch size: 65, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:24:10,138 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.77 vs. limit=5.0 +2023-03-26 17:24:24,848 INFO [finetune.py:976] (4/7) Epoch 14, batch 4050, loss[loss=0.1895, simple_loss=0.2641, pruned_loss=0.05748, over 4903.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2542, pruned_loss=0.05994, over 952795.12 frames. ], batch size: 43, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:24:31,534 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=78521.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:24:39,297 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 17:24:45,460 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.285e+01 1.581e+02 1.919e+02 2.315e+02 3.488e+02, threshold=3.837e+02, percent-clipped=0.0 +2023-03-26 17:24:54,883 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78556.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:24:57,774 INFO [finetune.py:976] (4/7) Epoch 14, batch 4100, loss[loss=0.1988, simple_loss=0.2658, pruned_loss=0.06586, over 4775.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2564, pruned_loss=0.06071, over 953299.71 frames. ], batch size: 29, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:25:07,988 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 17:25:09,968 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8278, 1.2165, 0.9312, 1.6380, 2.1211, 1.3559, 1.5655, 1.6255], + device='cuda:4'), covar=tensor([0.1451, 0.2203, 0.1931, 0.1188, 0.1958, 0.1960, 0.1478, 0.1948], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0111, 0.0092, 0.0119, 0.0094, 0.0100, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 17:25:16,502 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78588.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 17:25:31,553 INFO [finetune.py:976] (4/7) Epoch 14, batch 4150, loss[loss=0.1894, simple_loss=0.274, pruned_loss=0.05245, over 4868.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2577, pruned_loss=0.06113, over 953235.01 frames. ], batch size: 34, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:25:35,324 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=78617.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:25:44,798 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 17:25:52,376 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.171e+02 1.587e+02 1.869e+02 2.218e+02 3.242e+02, threshold=3.739e+02, percent-clipped=0.0 +2023-03-26 17:25:58,211 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.57 vs. limit=5.0 +2023-03-26 17:26:00,939 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 17:26:04,877 INFO [finetune.py:976] (4/7) Epoch 14, batch 4200, loss[loss=0.147, simple_loss=0.2167, pruned_loss=0.03869, over 4755.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2592, pruned_loss=0.06131, over 953344.67 frames. ], batch size: 26, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:26:33,974 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8119, 1.2553, 0.8109, 1.6219, 2.1917, 1.5255, 1.6131, 1.7823], + device='cuda:4'), covar=tensor([0.1529, 0.2188, 0.2253, 0.1335, 0.1987, 0.2064, 0.1549, 0.1936], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0111, 0.0092, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 17:26:37,999 INFO [finetune.py:976] (4/7) Epoch 14, batch 4250, loss[loss=0.1655, simple_loss=0.2332, pruned_loss=0.0489, over 4815.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.2575, pruned_loss=0.06056, over 953684.02 frames. ], batch size: 25, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:26:56,635 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7189, 0.7586, 1.6663, 1.6375, 1.5248, 1.4513, 1.5278, 1.5699], + device='cuda:4'), covar=tensor([0.3311, 0.3495, 0.3168, 0.3028, 0.4183, 0.3244, 0.3794, 0.2899], + device='cuda:4'), in_proj_covar=tensor([0.0243, 0.0239, 0.0256, 0.0266, 0.0265, 0.0238, 0.0278, 0.0236], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:27:05,887 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.090e+02 1.501e+02 1.722e+02 2.085e+02 5.543e+02, threshold=3.444e+02, percent-clipped=2.0 +2023-03-26 17:27:21,261 INFO [finetune.py:976] (4/7) Epoch 14, batch 4300, loss[loss=0.146, simple_loss=0.2267, pruned_loss=0.03265, over 4777.00 frames. ], tot_loss[loss=0.1882, simple_loss=0.2558, pruned_loss=0.0603, over 955356.38 frames. ], batch size: 29, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:27:59,989 INFO [finetune.py:976] (4/7) Epoch 14, batch 4350, loss[loss=0.1478, simple_loss=0.2214, pruned_loss=0.03716, over 4786.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2515, pruned_loss=0.05881, over 954970.41 frames. ], batch size: 29, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:28:06,793 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78816.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:28:34,426 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.179e+02 1.620e+02 1.919e+02 2.276e+02 4.946e+02, threshold=3.838e+02, percent-clipped=3.0 +2023-03-26 17:28:54,138 INFO [finetune.py:976] (4/7) Epoch 14, batch 4400, loss[loss=0.1738, simple_loss=0.2355, pruned_loss=0.05605, over 4320.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2522, pruned_loss=0.05928, over 954015.59 frames. ], batch size: 18, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:29:12,596 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=78888.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:29:27,942 INFO [finetune.py:976] (4/7) Epoch 14, batch 4450, loss[loss=0.2243, simple_loss=0.2963, pruned_loss=0.07614, over 4822.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2542, pruned_loss=0.05989, over 954822.42 frames. ], batch size: 38, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:29:28,629 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78912.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:29:33,955 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9133, 1.4259, 0.9747, 1.7774, 2.2608, 1.5196, 1.6534, 1.7750], + device='cuda:4'), covar=tensor([0.1531, 0.2165, 0.2156, 0.1311, 0.1919, 0.2017, 0.1594, 0.2072], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0111, 0.0092, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 17:29:44,125 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=78936.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:29:48,694 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.623e+02 2.075e+02 2.454e+02 4.700e+02, threshold=4.150e+02, percent-clipped=3.0 +2023-03-26 17:29:51,122 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-03-26 17:30:01,642 INFO [finetune.py:976] (4/7) Epoch 14, batch 4500, loss[loss=0.1534, simple_loss=0.2195, pruned_loss=0.04368, over 4809.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2554, pruned_loss=0.05992, over 954357.56 frames. ], batch size: 33, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:30:04,118 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78965.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:30:24,341 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2810, 1.9331, 1.9542, 0.9472, 2.1717, 2.4708, 2.0956, 1.8580], + device='cuda:4'), covar=tensor([0.0972, 0.0746, 0.0574, 0.0699, 0.0568, 0.0622, 0.0543, 0.0718], + device='cuda:4'), in_proj_covar=tensor([0.0127, 0.0153, 0.0123, 0.0130, 0.0131, 0.0127, 0.0142, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.3637e-05, 1.1182e-04, 8.8733e-05, 9.3678e-05, 9.2527e-05, 9.2049e-05, + 1.0318e-04, 1.0594e-04], device='cuda:4') +2023-03-26 17:30:34,873 INFO [finetune.py:976] (4/7) Epoch 14, batch 4550, loss[loss=0.1916, simple_loss=0.2636, pruned_loss=0.05982, over 4931.00 frames. ], tot_loss[loss=0.1895, simple_loss=0.2571, pruned_loss=0.06092, over 954317.40 frames. ], batch size: 33, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:30:44,022 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79026.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:30:54,527 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.627e+02 1.835e+02 2.182e+02 4.419e+02, threshold=3.671e+02, percent-clipped=1.0 +2023-03-26 17:31:00,895 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6667, 1.8852, 0.9869, 2.4795, 2.9310, 2.1475, 2.4163, 2.3800], + device='cuda:4'), covar=tensor([0.1296, 0.1894, 0.2222, 0.1113, 0.1600, 0.1744, 0.1326, 0.1921], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0111, 0.0093, 0.0119, 0.0095, 0.0099, 0.0090], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 17:31:08,636 INFO [finetune.py:976] (4/7) Epoch 14, batch 4600, loss[loss=0.1843, simple_loss=0.2473, pruned_loss=0.06067, over 4894.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2563, pruned_loss=0.06028, over 955161.91 frames. ], batch size: 37, lr: 3.53e-03, grad_scale: 64.0 +2023-03-26 17:31:42,446 INFO [finetune.py:976] (4/7) Epoch 14, batch 4650, loss[loss=0.2382, simple_loss=0.2934, pruned_loss=0.09151, over 4926.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2536, pruned_loss=0.05965, over 955099.57 frames. ], batch size: 38, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:31:45,567 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79116.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:32:02,950 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.138e+02 1.520e+02 1.886e+02 2.415e+02 4.094e+02, threshold=3.771e+02, percent-clipped=1.0 +2023-03-26 17:32:18,248 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 17:32:22,763 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2109, 2.1624, 2.1938, 1.5881, 2.0896, 2.2649, 2.3353, 1.8320], + device='cuda:4'), covar=tensor([0.0563, 0.0663, 0.0666, 0.0875, 0.0712, 0.0722, 0.0555, 0.1106], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0132, 0.0141, 0.0123, 0.0123, 0.0140, 0.0140, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:32:24,498 INFO [finetune.py:976] (4/7) Epoch 14, batch 4700, loss[loss=0.2189, simple_loss=0.2727, pruned_loss=0.08254, over 4918.00 frames. ], tot_loss[loss=0.1858, simple_loss=0.2525, pruned_loss=0.0595, over 954048.37 frames. ], batch size: 37, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:32:26,328 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=79164.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:32:31,917 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4692, 2.3361, 1.7652, 0.8507, 2.1073, 1.8966, 1.6827, 2.0429], + device='cuda:4'), covar=tensor([0.0842, 0.0655, 0.1550, 0.1900, 0.1219, 0.2096, 0.2084, 0.0983], + device='cuda:4'), in_proj_covar=tensor([0.0167, 0.0195, 0.0200, 0.0184, 0.0214, 0.0207, 0.0222, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:32:50,081 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 17:32:58,004 INFO [finetune.py:976] (4/7) Epoch 14, batch 4750, loss[loss=0.2004, simple_loss=0.2696, pruned_loss=0.06555, over 4774.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2507, pruned_loss=0.05926, over 953440.03 frames. ], batch size: 28, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:32:59,188 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79212.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:33:06,642 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2970, 2.1521, 1.8226, 2.3957, 2.2121, 1.9585, 2.6697, 2.3210], + device='cuda:4'), covar=tensor([0.1448, 0.2397, 0.3181, 0.2637, 0.2921, 0.1761, 0.2941, 0.1857], + device='cuda:4'), in_proj_covar=tensor([0.0180, 0.0186, 0.0234, 0.0254, 0.0243, 0.0199, 0.0212, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:33:32,018 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.558e+01 1.654e+02 1.996e+02 2.359e+02 6.861e+02, threshold=3.993e+02, percent-clipped=2.0 +2023-03-26 17:33:51,147 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=79260.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:33:51,680 INFO [finetune.py:976] (4/7) Epoch 14, batch 4800, loss[loss=0.1818, simple_loss=0.2559, pruned_loss=0.05389, over 4927.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2524, pruned_loss=0.05945, over 954105.88 frames. ], batch size: 33, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:34:10,277 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79285.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 17:34:27,365 INFO [finetune.py:976] (4/7) Epoch 14, batch 4850, loss[loss=0.1629, simple_loss=0.223, pruned_loss=0.0514, over 4725.00 frames. ], tot_loss[loss=0.189, simple_loss=0.2562, pruned_loss=0.06086, over 955100.60 frames. ], batch size: 23, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:34:35,462 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79321.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:34:36,946 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-26 17:34:42,717 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79333.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:34:46,397 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0689, 2.1357, 1.6548, 2.0437, 2.0169, 1.7487, 2.4197, 2.1215], + device='cuda:4'), covar=tensor([0.1347, 0.1807, 0.2908, 0.2659, 0.2600, 0.1684, 0.2983, 0.1724], + device='cuda:4'), in_proj_covar=tensor([0.0180, 0.0186, 0.0234, 0.0254, 0.0243, 0.0199, 0.0212, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:34:49,129 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.639e+02 1.937e+02 2.312e+02 4.640e+02, threshold=3.873e+02, percent-clipped=1.0 +2023-03-26 17:34:51,065 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79346.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 17:34:58,776 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9755, 1.7275, 2.1431, 3.7439, 2.5476, 2.5831, 0.9158, 3.0251], + device='cuda:4'), covar=tensor([0.1781, 0.1568, 0.1537, 0.0707, 0.0823, 0.1920, 0.2062, 0.0558], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0131, 0.0162, 0.0100, 0.0136, 0.0124, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 17:35:00,320 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 17:35:00,515 INFO [finetune.py:976] (4/7) Epoch 14, batch 4900, loss[loss=0.1526, simple_loss=0.2313, pruned_loss=0.03693, over 4820.00 frames. ], tot_loss[loss=0.1898, simple_loss=0.2575, pruned_loss=0.06105, over 953956.81 frames. ], batch size: 25, lr: 3.53e-03, grad_scale: 32.0 +2023-03-26 17:35:03,416 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79364.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:35:11,506 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79375.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 17:35:15,783 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.9924, 3.4523, 3.6239, 3.8482, 3.8057, 3.4926, 4.0696, 1.5283], + device='cuda:4'), covar=tensor([0.0715, 0.0841, 0.0780, 0.0828, 0.1083, 0.1445, 0.0689, 0.4656], + device='cuda:4'), in_proj_covar=tensor([0.0345, 0.0240, 0.0271, 0.0289, 0.0328, 0.0280, 0.0297, 0.0292], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:35:23,191 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79394.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:35:34,336 INFO [finetune.py:976] (4/7) Epoch 14, batch 4950, loss[loss=0.1552, simple_loss=0.2301, pruned_loss=0.04012, over 4815.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2583, pruned_loss=0.06121, over 954133.65 frames. ], batch size: 33, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:35:45,281 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79425.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:35:51,952 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79436.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 17:35:55,993 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.049e+02 1.541e+02 1.877e+02 2.434e+02 3.585e+02, threshold=3.755e+02, percent-clipped=0.0 +2023-03-26 17:36:07,914 INFO [finetune.py:976] (4/7) Epoch 14, batch 5000, loss[loss=0.1331, simple_loss=0.2056, pruned_loss=0.0303, over 4778.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2578, pruned_loss=0.06135, over 952392.02 frames. ], batch size: 25, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:36:19,359 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7550, 1.7190, 2.1684, 1.9298, 2.0361, 4.4246, 1.6060, 1.8711], + device='cuda:4'), covar=tensor([0.0926, 0.1686, 0.1199, 0.0956, 0.1361, 0.0184, 0.1431, 0.1573], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0074, 0.0077, 0.0092, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 17:36:41,414 INFO [finetune.py:976] (4/7) Epoch 14, batch 5050, loss[loss=0.1753, simple_loss=0.2495, pruned_loss=0.05062, over 4827.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2559, pruned_loss=0.06078, over 952879.14 frames. ], batch size: 33, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:37:02,696 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.152e+02 1.527e+02 1.776e+02 2.127e+02 3.568e+02, threshold=3.553e+02, percent-clipped=0.0 +2023-03-26 17:37:08,295 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8107, 1.3070, 0.8150, 1.7082, 2.1896, 1.3574, 1.6015, 1.5396], + device='cuda:4'), covar=tensor([0.1554, 0.2148, 0.2143, 0.1291, 0.2002, 0.2026, 0.1575, 0.2085], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0094, 0.0111, 0.0092, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 17:37:09,571 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.42 vs. limit=5.0 +2023-03-26 17:37:14,551 INFO [finetune.py:976] (4/7) Epoch 14, batch 5100, loss[loss=0.162, simple_loss=0.2221, pruned_loss=0.05093, over 4868.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2528, pruned_loss=0.05976, over 952272.82 frames. ], batch size: 44, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:37:19,937 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-03-26 17:37:25,968 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0786, 1.9254, 1.6338, 1.8789, 1.7693, 1.8282, 1.8803, 2.5977], + device='cuda:4'), covar=tensor([0.3923, 0.4785, 0.3521, 0.4363, 0.4529, 0.2512, 0.4023, 0.1770], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0259, 0.0224, 0.0276, 0.0247, 0.0214, 0.0249, 0.0225], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:37:38,124 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79585.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:37:57,944 INFO [finetune.py:976] (4/7) Epoch 14, batch 5150, loss[loss=0.1802, simple_loss=0.2489, pruned_loss=0.05572, over 4808.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2523, pruned_loss=0.0591, over 954545.62 frames. ], batch size: 41, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:38:04,610 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79621.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:38:20,363 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79641.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 17:38:21,529 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.601e+02 1.924e+02 2.369e+02 3.228e+02, threshold=3.849e+02, percent-clipped=0.0 +2023-03-26 17:38:23,477 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79646.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:38:38,693 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79657.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:38:41,521 INFO [finetune.py:976] (4/7) Epoch 14, batch 5200, loss[loss=0.2703, simple_loss=0.3311, pruned_loss=0.1047, over 4092.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2546, pruned_loss=0.05986, over 950956.64 frames. ], batch size: 65, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:38:50,978 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=79669.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:39:10,039 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79689.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:39:27,297 INFO [finetune.py:976] (4/7) Epoch 14, batch 5250, loss[loss=0.2023, simple_loss=0.2757, pruned_loss=0.0645, over 4818.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2569, pruned_loss=0.05992, over 953537.80 frames. ], batch size: 40, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:39:32,146 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79718.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:39:33,309 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79720.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:39:40,420 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79731.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 17:39:49,075 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.058e+02 1.683e+02 1.987e+02 2.478e+02 3.642e+02, threshold=3.974e+02, percent-clipped=0.0 +2023-03-26 17:39:59,970 INFO [finetune.py:976] (4/7) Epoch 14, batch 5300, loss[loss=0.2187, simple_loss=0.2966, pruned_loss=0.07036, over 4811.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2592, pruned_loss=0.06114, over 952844.20 frames. ], batch size: 40, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:40:00,725 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79762.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:40:03,799 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-26 17:40:27,108 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-26 17:40:33,373 INFO [finetune.py:976] (4/7) Epoch 14, batch 5350, loss[loss=0.2136, simple_loss=0.274, pruned_loss=0.07663, over 4733.00 frames. ], tot_loss[loss=0.1902, simple_loss=0.2588, pruned_loss=0.06076, over 952057.24 frames. ], batch size: 59, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:40:41,355 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79823.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:40:55,391 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.090e+02 1.503e+02 1.776e+02 2.291e+02 4.117e+02, threshold=3.553e+02, percent-clipped=3.0 +2023-03-26 17:41:06,815 INFO [finetune.py:976] (4/7) Epoch 14, batch 5400, loss[loss=0.2246, simple_loss=0.2701, pruned_loss=0.08954, over 4722.00 frames. ], tot_loss[loss=0.1893, simple_loss=0.257, pruned_loss=0.06073, over 952551.01 frames. ], batch size: 23, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:41:30,219 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2404, 1.2794, 1.5302, 1.0118, 1.1355, 1.3934, 1.2978, 1.5962], + device='cuda:4'), covar=tensor([0.1256, 0.2117, 0.1371, 0.1563, 0.1007, 0.1353, 0.2754, 0.0862], + device='cuda:4'), in_proj_covar=tensor([0.0194, 0.0204, 0.0192, 0.0190, 0.0176, 0.0214, 0.0217, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:41:35,479 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3767, 1.4341, 1.4228, 0.7587, 1.4035, 1.6783, 1.6791, 1.3716], + device='cuda:4'), covar=tensor([0.0804, 0.0499, 0.0475, 0.0436, 0.0408, 0.0550, 0.0274, 0.0573], + device='cuda:4'), in_proj_covar=tensor([0.0125, 0.0151, 0.0122, 0.0129, 0.0130, 0.0127, 0.0141, 0.0144], + device='cuda:4'), out_proj_covar=tensor([9.2730e-05, 1.1004e-04, 8.7717e-05, 9.2406e-05, 9.1943e-05, 9.1588e-05, + 1.0194e-04, 1.0456e-04], device='cuda:4') +2023-03-26 17:41:40,249 INFO [finetune.py:976] (4/7) Epoch 14, batch 5450, loss[loss=0.1723, simple_loss=0.2277, pruned_loss=0.05844, over 4829.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2538, pruned_loss=0.05954, over 955915.31 frames. ], batch size: 33, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:41:54,246 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-26 17:41:59,678 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79941.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:41:59,713 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79941.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 17:42:00,807 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.485e+02 1.773e+02 2.175e+02 4.141e+02, threshold=3.546e+02, percent-clipped=3.0 +2023-03-26 17:42:14,249 INFO [finetune.py:976] (4/7) Epoch 14, batch 5500, loss[loss=0.136, simple_loss=0.2058, pruned_loss=0.03313, over 4777.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2514, pruned_loss=0.05879, over 954662.67 frames. ], batch size: 28, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:42:32,376 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=79989.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 17:42:32,395 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79989.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:42:54,670 INFO [finetune.py:976] (4/7) Epoch 14, batch 5550, loss[loss=0.159, simple_loss=0.2188, pruned_loss=0.0496, over 3891.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2534, pruned_loss=0.0598, over 953741.11 frames. ], batch size: 17, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:42:55,987 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80013.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:43:00,252 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80020.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:43:07,416 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80031.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 17:43:11,544 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80037.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:43:15,067 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.609e+02 1.862e+02 2.441e+02 4.163e+02, threshold=3.724e+02, percent-clipped=2.0 +2023-03-26 17:43:17,906 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.82 vs. limit=5.0 +2023-03-26 17:43:25,569 INFO [finetune.py:976] (4/7) Epoch 14, batch 5600, loss[loss=0.1848, simple_loss=0.2577, pruned_loss=0.0559, over 4867.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2563, pruned_loss=0.06043, over 952766.27 frames. ], batch size: 44, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:43:29,672 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80068.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:43:36,039 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80079.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 17:43:53,127 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6238, 1.1484, 0.8881, 1.5248, 1.9891, 1.0695, 1.4300, 1.4684], + device='cuda:4'), covar=tensor([0.1501, 0.2168, 0.1986, 0.1269, 0.2086, 0.2149, 0.1513, 0.1998], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0095, 0.0111, 0.0093, 0.0120, 0.0094, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 17:44:02,497 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 17:44:02,585 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-26 17:44:12,674 INFO [finetune.py:976] (4/7) Epoch 14, batch 5650, loss[loss=0.2212, simple_loss=0.291, pruned_loss=0.07572, over 4913.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2599, pruned_loss=0.06158, over 954443.60 frames. ], batch size: 36, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:44:21,798 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80118.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:44:33,748 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 17:44:41,676 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.115e+02 1.565e+02 1.838e+02 2.201e+02 4.652e+02, threshold=3.676e+02, percent-clipped=2.0 +2023-03-26 17:44:56,307 INFO [finetune.py:976] (4/7) Epoch 14, batch 5700, loss[loss=0.1555, simple_loss=0.2144, pruned_loss=0.04828, over 4255.00 frames. ], tot_loss[loss=0.1909, simple_loss=0.2579, pruned_loss=0.06194, over 936074.83 frames. ], batch size: 18, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:45:02,927 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3887, 1.9078, 2.3097, 2.2933, 2.0569, 2.0464, 2.2146, 2.1507], + device='cuda:4'), covar=tensor([0.2860, 0.3338, 0.3122, 0.3190, 0.4418, 0.3234, 0.4081, 0.2847], + device='cuda:4'), in_proj_covar=tensor([0.0244, 0.0240, 0.0258, 0.0267, 0.0266, 0.0239, 0.0280, 0.0236], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:45:28,233 INFO [finetune.py:976] (4/7) Epoch 15, batch 0, loss[loss=0.1596, simple_loss=0.2331, pruned_loss=0.04302, over 4868.00 frames. ], tot_loss[loss=0.1596, simple_loss=0.2331, pruned_loss=0.04302, over 4868.00 frames. ], batch size: 32, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:45:28,233 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 17:45:33,001 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8403, 3.6202, 3.3152, 1.5662, 3.5778, 2.7195, 0.7256, 2.3212], + device='cuda:4'), covar=tensor([0.1827, 0.2043, 0.1748, 0.3874, 0.0991, 0.1154, 0.4198, 0.1793], + device='cuda:4'), in_proj_covar=tensor([0.0149, 0.0171, 0.0158, 0.0127, 0.0155, 0.0122, 0.0144, 0.0121], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 17:45:36,127 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6792, 1.5670, 1.5447, 1.5912, 0.9505, 2.9890, 1.1276, 1.6502], + device='cuda:4'), covar=tensor([0.3522, 0.2582, 0.2124, 0.2465, 0.1993, 0.0253, 0.2728, 0.1308], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0116, 0.0121, 0.0124, 0.0115, 0.0098, 0.0097, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 17:45:36,740 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1462, 2.0009, 1.8007, 1.9505, 2.1753, 1.8604, 2.3022, 2.1147], + device='cuda:4'), covar=tensor([0.1319, 0.2295, 0.3047, 0.2471, 0.2495, 0.1733, 0.3630, 0.1834], + device='cuda:4'), in_proj_covar=tensor([0.0180, 0.0187, 0.0234, 0.0255, 0.0245, 0.0200, 0.0213, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:45:42,542 INFO [finetune.py:1010] (4/7) Epoch 15, validation: loss=0.1586, simple_loss=0.2288, pruned_loss=0.0442, over 2265189.00 frames. +2023-03-26 17:45:42,543 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 17:45:42,687 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8794, 1.7590, 1.5779, 1.4658, 1.9136, 1.6277, 1.8552, 1.8509], + device='cuda:4'), covar=tensor([0.1380, 0.2151, 0.3073, 0.2475, 0.2613, 0.1706, 0.2737, 0.1800], + device='cuda:4'), in_proj_covar=tensor([0.0180, 0.0187, 0.0234, 0.0255, 0.0245, 0.0200, 0.0213, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:46:08,706 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4692, 2.8025, 2.4848, 1.8403, 2.5841, 2.7408, 2.6195, 2.4318], + device='cuda:4'), covar=tensor([0.0696, 0.0580, 0.0814, 0.0987, 0.0744, 0.0815, 0.0684, 0.0937], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0132, 0.0141, 0.0123, 0.0124, 0.0141, 0.0141, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:46:15,191 INFO [finetune.py:976] (4/7) Epoch 15, batch 50, loss[loss=0.2725, simple_loss=0.3165, pruned_loss=0.1143, over 4128.00 frames. ], tot_loss[loss=0.1916, simple_loss=0.2607, pruned_loss=0.06122, over 216272.83 frames. ], batch size: 66, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:46:17,635 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80241.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:46:18,756 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.025e+02 1.470e+02 1.896e+02 2.201e+02 3.299e+02, threshold=3.792e+02, percent-clipped=0.0 +2023-03-26 17:46:45,388 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80283.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:46:48,316 INFO [finetune.py:976] (4/7) Epoch 15, batch 100, loss[loss=0.1723, simple_loss=0.2317, pruned_loss=0.05639, over 4820.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2532, pruned_loss=0.05884, over 380064.91 frames. ], batch size: 41, lr: 3.52e-03, grad_scale: 32.0 +2023-03-26 17:46:48,988 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80289.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:47:05,103 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80313.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:47:05,683 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8221, 1.0927, 0.8757, 1.6574, 2.0844, 1.4688, 1.4796, 1.6348], + device='cuda:4'), covar=tensor([0.1538, 0.2356, 0.1991, 0.1350, 0.1961, 0.2014, 0.1617, 0.2031], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0095, 0.0111, 0.0093, 0.0120, 0.0094, 0.0099, 0.0090], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 17:47:21,595 INFO [finetune.py:976] (4/7) Epoch 15, batch 150, loss[loss=0.182, simple_loss=0.2542, pruned_loss=0.05488, over 4753.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2487, pruned_loss=0.05817, over 507332.85 frames. ], batch size: 27, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:47:25,132 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.875e+01 1.589e+02 1.860e+02 2.189e+02 4.694e+02, threshold=3.721e+02, percent-clipped=1.0 +2023-03-26 17:47:25,879 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80344.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:47:37,136 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80361.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:47:42,669 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0088, 2.1647, 1.8068, 1.9075, 2.4903, 2.5008, 2.1414, 1.9609], + device='cuda:4'), covar=tensor([0.0336, 0.0360, 0.0540, 0.0358, 0.0242, 0.0494, 0.0332, 0.0423], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0108, 0.0141, 0.0113, 0.0100, 0.0107, 0.0097, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.2362e-05, 8.3734e-05, 1.1152e-04, 8.7637e-05, 7.8303e-05, 7.9081e-05, + 7.2584e-05, 8.2632e-05], device='cuda:4') +2023-03-26 17:47:54,529 INFO [finetune.py:976] (4/7) Epoch 15, batch 200, loss[loss=0.2153, simple_loss=0.2783, pruned_loss=0.07612, over 4854.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.248, pruned_loss=0.05818, over 607120.03 frames. ], batch size: 44, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:47:58,250 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80393.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:48:16,681 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80418.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:48:34,170 INFO [finetune.py:976] (4/7) Epoch 15, batch 250, loss[loss=0.1811, simple_loss=0.2398, pruned_loss=0.06122, over 4712.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.25, pruned_loss=0.0583, over 682913.26 frames. ], batch size: 23, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:48:37,160 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.156e+02 1.638e+02 2.049e+02 2.410e+02 5.367e+02, threshold=4.098e+02, percent-clipped=2.0 +2023-03-26 17:48:48,456 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80454.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:48:48,800 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-03-26 17:48:58,603 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80463.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:49:00,360 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80466.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:49:20,044 INFO [finetune.py:976] (4/7) Epoch 15, batch 300, loss[loss=0.1438, simple_loss=0.2276, pruned_loss=0.03004, over 4757.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2538, pruned_loss=0.0595, over 742640.95 frames. ], batch size: 54, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:49:24,811 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80494.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:50:03,909 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80524.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:50:12,918 INFO [finetune.py:976] (4/7) Epoch 15, batch 350, loss[loss=0.1974, simple_loss=0.271, pruned_loss=0.06188, over 4914.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2558, pruned_loss=0.06013, over 789851.78 frames. ], batch size: 36, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:50:14,886 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2582, 1.8608, 1.9654, 0.8530, 2.1661, 2.2113, 2.0553, 1.8237], + device='cuda:4'), covar=tensor([0.0787, 0.0684, 0.0487, 0.0707, 0.0436, 0.0670, 0.0454, 0.0690], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0152, 0.0123, 0.0130, 0.0130, 0.0128, 0.0142, 0.0145], + device='cuda:4'), out_proj_covar=tensor([9.3312e-05, 1.1110e-04, 8.8617e-05, 9.3165e-05, 9.1907e-05, 9.2593e-05, + 1.0288e-04, 1.0513e-04], device='cuda:4') +2023-03-26 17:50:16,427 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.511e+02 1.809e+02 2.185e+02 3.892e+02, threshold=3.618e+02, percent-clipped=0.0 +2023-03-26 17:50:20,711 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7260, 1.5340, 1.4233, 1.6888, 2.4370, 1.7796, 1.6779, 1.3760], + device='cuda:4'), covar=tensor([0.2412, 0.2436, 0.2363, 0.1969, 0.1701, 0.1435, 0.2540, 0.2207], + device='cuda:4'), in_proj_covar=tensor([0.0239, 0.0207, 0.0210, 0.0191, 0.0240, 0.0185, 0.0215, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:50:24,827 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80555.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:50:47,437 INFO [finetune.py:976] (4/7) Epoch 15, batch 400, loss[loss=0.1538, simple_loss=0.2232, pruned_loss=0.04215, over 4802.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.2547, pruned_loss=0.05874, over 826000.05 frames. ], batch size: 25, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:51:29,119 INFO [finetune.py:976] (4/7) Epoch 15, batch 450, loss[loss=0.1806, simple_loss=0.2454, pruned_loss=0.05795, over 4868.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2535, pruned_loss=0.05838, over 853615.35 frames. ], batch size: 31, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:51:29,784 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80639.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:51:32,683 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.016e+02 1.580e+02 1.854e+02 2.177e+02 4.594e+02, threshold=3.707e+02, percent-clipped=2.0 +2023-03-26 17:52:03,114 INFO [finetune.py:976] (4/7) Epoch 15, batch 500, loss[loss=0.2248, simple_loss=0.2765, pruned_loss=0.08651, over 4901.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2516, pruned_loss=0.05763, over 877814.46 frames. ], batch size: 43, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:52:23,151 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2894, 2.7717, 2.7047, 1.3512, 2.8710, 2.2759, 0.9979, 2.0286], + device='cuda:4'), covar=tensor([0.2658, 0.1999, 0.1692, 0.2801, 0.1381, 0.0987, 0.3249, 0.1318], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0174, 0.0160, 0.0128, 0.0157, 0.0123, 0.0146, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 17:52:36,002 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80736.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:52:37,122 INFO [finetune.py:976] (4/7) Epoch 15, batch 550, loss[loss=0.2069, simple_loss=0.2539, pruned_loss=0.07995, over 4713.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.2492, pruned_loss=0.05717, over 895369.37 frames. ], batch size: 23, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:52:40,201 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.496e+02 1.725e+02 2.011e+02 3.976e+02, threshold=3.451e+02, percent-clipped=1.0 +2023-03-26 17:52:44,398 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80749.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:52:55,972 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3804, 1.4037, 0.7402, 2.2551, 2.5349, 1.7581, 1.8865, 2.0583], + device='cuda:4'), covar=tensor([0.1329, 0.2047, 0.2267, 0.1096, 0.1697, 0.1916, 0.1411, 0.1899], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0096, 0.0113, 0.0094, 0.0121, 0.0095, 0.0100, 0.0090], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 17:53:01,296 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80773.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:53:10,747 INFO [finetune.py:976] (4/7) Epoch 15, batch 600, loss[loss=0.1705, simple_loss=0.2415, pruned_loss=0.04977, over 4940.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2499, pruned_loss=0.05766, over 909422.23 frames. ], batch size: 38, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:53:14,575 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7179, 1.5593, 1.4896, 1.8077, 2.0881, 1.8049, 1.2849, 1.4472], + device='cuda:4'), covar=tensor([0.2371, 0.2291, 0.2173, 0.1776, 0.1684, 0.1338, 0.2648, 0.2078], + device='cuda:4'), in_proj_covar=tensor([0.0241, 0.0208, 0.0212, 0.0192, 0.0243, 0.0186, 0.0217, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:53:15,172 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1536, 1.9230, 1.6952, 1.6486, 1.9023, 1.8673, 1.8653, 2.5804], + device='cuda:4'), covar=tensor([0.3962, 0.4255, 0.3432, 0.3866, 0.3737, 0.2570, 0.3951, 0.1673], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0259, 0.0225, 0.0276, 0.0247, 0.0214, 0.0249, 0.0224], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:53:16,806 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80797.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:53:32,690 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80819.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:53:35,811 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9975, 4.0810, 3.8667, 1.9627, 4.1440, 3.1111, 0.9218, 2.9022], + device='cuda:4'), covar=tensor([0.1925, 0.1862, 0.1373, 0.3282, 0.0849, 0.0938, 0.4450, 0.1246], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0173, 0.0160, 0.0128, 0.0157, 0.0122, 0.0146, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 17:53:42,438 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-26 17:53:44,657 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80834.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:53:47,024 INFO [finetune.py:976] (4/7) Epoch 15, batch 650, loss[loss=0.1869, simple_loss=0.2598, pruned_loss=0.05701, over 4744.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2533, pruned_loss=0.05861, over 919849.70 frames. ], batch size: 27, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:53:50,577 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.196e+02 1.643e+02 1.965e+02 2.358e+02 6.399e+02, threshold=3.929e+02, percent-clipped=5.0 +2023-03-26 17:53:55,338 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80850.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:53:56,550 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2608, 2.1649, 1.8836, 0.9544, 1.9946, 1.8579, 1.7181, 1.9953], + device='cuda:4'), covar=tensor([0.1018, 0.0781, 0.1526, 0.1956, 0.1589, 0.2067, 0.1984, 0.0963], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0196, 0.0200, 0.0184, 0.0215, 0.0207, 0.0225, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:54:09,221 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-03-26 17:54:29,238 INFO [finetune.py:976] (4/7) Epoch 15, batch 700, loss[loss=0.2507, simple_loss=0.2997, pruned_loss=0.1008, over 4737.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2549, pruned_loss=0.05978, over 925696.37 frames. ], batch size: 59, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:54:33,527 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8232, 3.8487, 3.5746, 1.6926, 3.8662, 2.8779, 0.7337, 2.5945], + device='cuda:4'), covar=tensor([0.2190, 0.1512, 0.1619, 0.3378, 0.1006, 0.1000, 0.4653, 0.1408], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0174, 0.0159, 0.0128, 0.0157, 0.0122, 0.0146, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 17:54:42,714 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.0836, 3.5285, 3.7000, 3.9580, 3.8541, 3.5888, 4.1670, 1.1415], + device='cuda:4'), covar=tensor([0.0825, 0.0905, 0.0827, 0.0945, 0.1376, 0.1588, 0.0806, 0.5646], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0242, 0.0272, 0.0291, 0.0329, 0.0280, 0.0297, 0.0295], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:55:17,868 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6494, 2.4048, 2.1120, 1.0418, 2.2827, 1.9410, 1.8458, 2.1948], + device='cuda:4'), covar=tensor([0.0830, 0.0916, 0.1746, 0.2133, 0.1779, 0.2407, 0.2217, 0.1061], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0195, 0.0200, 0.0184, 0.0214, 0.0207, 0.0225, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:55:23,119 INFO [finetune.py:976] (4/7) Epoch 15, batch 750, loss[loss=0.211, simple_loss=0.2772, pruned_loss=0.07247, over 4905.00 frames. ], tot_loss[loss=0.1887, simple_loss=0.2565, pruned_loss=0.06048, over 931754.87 frames. ], batch size: 37, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:55:23,798 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80939.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:55:26,162 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.122e+02 1.628e+02 1.856e+02 2.303e+02 3.612e+02, threshold=3.712e+02, percent-clipped=0.0 +2023-03-26 17:55:32,802 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3589, 2.1502, 1.9185, 2.2074, 2.3213, 2.0212, 2.6087, 2.2693], + device='cuda:4'), covar=tensor([0.1283, 0.1941, 0.2975, 0.2223, 0.2275, 0.1670, 0.2690, 0.1815], + device='cuda:4'), in_proj_covar=tensor([0.0180, 0.0186, 0.0233, 0.0253, 0.0243, 0.0199, 0.0212, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:55:56,364 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=80987.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:55:56,912 INFO [finetune.py:976] (4/7) Epoch 15, batch 800, loss[loss=0.1936, simple_loss=0.2674, pruned_loss=0.0599, over 4907.00 frames. ], tot_loss[loss=0.189, simple_loss=0.2574, pruned_loss=0.06034, over 937581.93 frames. ], batch size: 37, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:55:59,498 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5595, 1.5756, 1.9532, 1.8843, 1.7402, 3.5099, 1.3641, 1.6077], + device='cuda:4'), covar=tensor([0.0971, 0.1712, 0.1116, 0.0947, 0.1518, 0.0259, 0.1524, 0.1693], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 17:56:05,680 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9707, 1.3449, 1.9386, 1.9328, 1.6938, 1.6643, 1.8436, 1.7385], + device='cuda:4'), covar=tensor([0.3711, 0.4162, 0.3225, 0.3686, 0.4762, 0.3651, 0.4565, 0.3227], + device='cuda:4'), in_proj_covar=tensor([0.0244, 0.0239, 0.0256, 0.0266, 0.0266, 0.0238, 0.0280, 0.0235], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 17:56:38,263 INFO [finetune.py:976] (4/7) Epoch 15, batch 850, loss[loss=0.1862, simple_loss=0.2552, pruned_loss=0.05857, over 4764.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2545, pruned_loss=0.05918, over 942880.79 frames. ], batch size: 27, lr: 3.51e-03, grad_scale: 32.0 +2023-03-26 17:56:41,289 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.679e+02 1.976e+02 2.340e+02 3.768e+02, threshold=3.952e+02, percent-clipped=2.0 +2023-03-26 17:56:44,974 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81049.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:57:06,272 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 17:57:11,963 INFO [finetune.py:976] (4/7) Epoch 15, batch 900, loss[loss=0.1737, simple_loss=0.2438, pruned_loss=0.05179, over 4729.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2524, pruned_loss=0.0585, over 946578.34 frames. ], batch size: 23, lr: 3.51e-03, grad_scale: 64.0 +2023-03-26 17:57:14,453 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81092.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:57:17,457 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81097.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:57:29,872 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81116.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:57:31,669 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81119.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:57:38,665 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81129.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:57:45,616 INFO [finetune.py:976] (4/7) Epoch 15, batch 950, loss[loss=0.1775, simple_loss=0.2473, pruned_loss=0.05381, over 4737.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2504, pruned_loss=0.05791, over 949901.02 frames. ], batch size: 59, lr: 3.51e-03, grad_scale: 64.0 +2023-03-26 17:57:48,670 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.129e+01 1.456e+02 1.848e+02 2.216e+02 5.430e+02, threshold=3.695e+02, percent-clipped=2.0 +2023-03-26 17:57:53,006 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81150.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:58:03,767 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81167.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:58:11,314 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81177.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:58:19,389 INFO [finetune.py:976] (4/7) Epoch 15, batch 1000, loss[loss=0.1584, simple_loss=0.2342, pruned_loss=0.04133, over 4783.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2513, pruned_loss=0.05836, over 950290.92 frames. ], batch size: 28, lr: 3.51e-03, grad_scale: 64.0 +2023-03-26 17:58:25,511 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81198.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:58:52,892 INFO [finetune.py:976] (4/7) Epoch 15, batch 1050, loss[loss=0.1408, simple_loss=0.2183, pruned_loss=0.0317, over 4852.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2538, pruned_loss=0.05834, over 951975.94 frames. ], batch size: 31, lr: 3.51e-03, grad_scale: 64.0 +2023-03-26 17:58:56,386 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.183e+02 1.566e+02 1.800e+02 2.282e+02 3.514e+02, threshold=3.601e+02, percent-clipped=0.0 +2023-03-26 17:59:31,696 INFO [finetune.py:976] (4/7) Epoch 15, batch 1100, loss[loss=0.1631, simple_loss=0.2459, pruned_loss=0.04019, over 4859.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2562, pruned_loss=0.05883, over 954235.15 frames. ], batch size: 44, lr: 3.51e-03, grad_scale: 64.0 +2023-03-26 17:59:43,365 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81299.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:59:49,370 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81301.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 17:59:57,457 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7239, 1.1926, 0.8655, 1.6184, 2.0289, 1.3164, 1.4147, 1.5197], + device='cuda:4'), covar=tensor([0.1427, 0.2048, 0.2014, 0.1185, 0.1957, 0.2152, 0.1453, 0.1846], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0097, 0.0113, 0.0094, 0.0122, 0.0095, 0.0101, 0.0091], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 18:00:16,333 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-03-26 18:00:16,533 INFO [finetune.py:976] (4/7) Epoch 15, batch 1150, loss[loss=0.1601, simple_loss=0.2253, pruned_loss=0.04747, over 4665.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2574, pruned_loss=0.0593, over 953868.87 frames. ], batch size: 23, lr: 3.51e-03, grad_scale: 64.0 +2023-03-26 18:00:23,242 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.122e+02 1.635e+02 2.084e+02 2.407e+02 3.907e+02, threshold=4.168e+02, percent-clipped=1.0 +2023-03-26 18:00:41,599 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81360.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:00:42,880 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81362.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:00:59,899 INFO [finetune.py:976] (4/7) Epoch 15, batch 1200, loss[loss=0.179, simple_loss=0.2309, pruned_loss=0.06358, over 4715.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2561, pruned_loss=0.05919, over 953899.20 frames. ], batch size: 59, lr: 3.51e-03, grad_scale: 64.0 +2023-03-26 18:01:03,413 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81392.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:01:04,057 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2620, 2.2457, 1.9831, 2.3875, 3.0280, 2.2347, 2.2354, 1.7308], + device='cuda:4'), covar=tensor([0.2165, 0.2037, 0.1885, 0.1504, 0.1606, 0.1187, 0.2069, 0.2044], + device='cuda:4'), in_proj_covar=tensor([0.0240, 0.0207, 0.0210, 0.0191, 0.0241, 0.0185, 0.0215, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:01:27,579 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81429.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:01:35,104 INFO [finetune.py:976] (4/7) Epoch 15, batch 1250, loss[loss=0.1556, simple_loss=0.2073, pruned_loss=0.05196, over 3976.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2532, pruned_loss=0.05848, over 954317.99 frames. ], batch size: 17, lr: 3.51e-03, grad_scale: 64.0 +2023-03-26 18:01:40,099 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81440.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:01:42,324 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.103e+01 1.546e+02 1.830e+02 2.259e+02 3.665e+02, threshold=3.660e+02, percent-clipped=0.0 +2023-03-26 18:02:01,853 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4907, 1.4614, 1.6685, 1.6145, 1.5818, 3.2683, 1.4472, 1.5062], + device='cuda:4'), covar=tensor([0.1045, 0.1861, 0.1180, 0.1059, 0.1683, 0.0240, 0.1462, 0.1773], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0078, 0.0092, 0.0081, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 18:02:05,450 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81472.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:02:08,459 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81477.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:02:15,564 INFO [finetune.py:976] (4/7) Epoch 15, batch 1300, loss[loss=0.198, simple_loss=0.2576, pruned_loss=0.06917, over 4342.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2506, pruned_loss=0.05792, over 954868.91 frames. ], batch size: 65, lr: 3.50e-03, grad_scale: 64.0 +2023-03-26 18:02:49,394 INFO [finetune.py:976] (4/7) Epoch 15, batch 1350, loss[loss=0.2088, simple_loss=0.2788, pruned_loss=0.06936, over 4866.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2508, pruned_loss=0.05799, over 953285.60 frames. ], batch size: 44, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:02:53,475 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.608e+02 1.859e+02 2.257e+02 3.880e+02, threshold=3.719e+02, percent-clipped=1.0 +2023-03-26 18:02:55,314 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7645, 3.9651, 3.6545, 1.9774, 3.9949, 3.0263, 0.7793, 2.8336], + device='cuda:4'), covar=tensor([0.2318, 0.2230, 0.1432, 0.3542, 0.0934, 0.1014, 0.4686, 0.1502], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0174, 0.0158, 0.0128, 0.0156, 0.0121, 0.0145, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 18:03:04,213 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3042, 1.3413, 1.3735, 1.5418, 1.4856, 2.8760, 1.2891, 1.4179], + device='cuda:4'), covar=tensor([0.1070, 0.1935, 0.1435, 0.1033, 0.1739, 0.0323, 0.1608, 0.1832], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0082, 0.0074, 0.0078, 0.0092, 0.0081, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 18:03:06,703 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.63 vs. limit=5.0 +2023-03-26 18:03:22,724 INFO [finetune.py:976] (4/7) Epoch 15, batch 1400, loss[loss=0.1957, simple_loss=0.2787, pruned_loss=0.05636, over 4903.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2533, pruned_loss=0.05835, over 953955.51 frames. ], batch size: 43, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:03:33,299 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5643, 2.7841, 2.5551, 1.9843, 2.8364, 3.0416, 2.8746, 2.4864], + device='cuda:4'), covar=tensor([0.0648, 0.0552, 0.0749, 0.0863, 0.0495, 0.0698, 0.0633, 0.0942], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0132, 0.0140, 0.0122, 0.0122, 0.0140, 0.0140, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:03:56,015 INFO [finetune.py:976] (4/7) Epoch 15, batch 1450, loss[loss=0.2279, simple_loss=0.2938, pruned_loss=0.08104, over 4858.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.2554, pruned_loss=0.05895, over 954342.14 frames. ], batch size: 44, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:04:00,103 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.106e+02 1.620e+02 1.887e+02 2.237e+02 3.719e+02, threshold=3.774e+02, percent-clipped=1.0 +2023-03-26 18:04:07,945 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81655.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:04:09,564 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81657.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:04:18,104 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81670.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:04:29,493 INFO [finetune.py:976] (4/7) Epoch 15, batch 1500, loss[loss=0.1698, simple_loss=0.2517, pruned_loss=0.04398, over 4787.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2561, pruned_loss=0.05898, over 956217.25 frames. ], batch size: 29, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:05:16,700 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81731.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 18:05:20,265 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5217, 3.5991, 3.4540, 1.6716, 3.7042, 2.8430, 0.7584, 2.5697], + device='cuda:4'), covar=tensor([0.2551, 0.2187, 0.1531, 0.3394, 0.1207, 0.1037, 0.4715, 0.1451], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0174, 0.0159, 0.0128, 0.0156, 0.0122, 0.0146, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 18:05:20,805 INFO [finetune.py:976] (4/7) Epoch 15, batch 1550, loss[loss=0.2002, simple_loss=0.2656, pruned_loss=0.06743, over 4919.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.2563, pruned_loss=0.05894, over 954705.54 frames. ], batch size: 38, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:05:24,955 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.530e+02 1.898e+02 2.293e+02 4.636e+02, threshold=3.795e+02, percent-clipped=1.0 +2023-03-26 18:05:50,065 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81772.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:06:03,755 INFO [finetune.py:976] (4/7) Epoch 15, batch 1600, loss[loss=0.1602, simple_loss=0.2354, pruned_loss=0.04251, over 4903.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2539, pruned_loss=0.05821, over 956306.94 frames. ], batch size: 36, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:06:12,838 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9158, 3.9909, 3.7741, 2.0881, 4.0406, 3.0856, 0.8879, 2.8755], + device='cuda:4'), covar=tensor([0.1959, 0.1754, 0.1352, 0.3069, 0.1046, 0.0944, 0.4473, 0.1415], + device='cuda:4'), in_proj_covar=tensor([0.0149, 0.0172, 0.0157, 0.0127, 0.0155, 0.0121, 0.0144, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 18:06:25,747 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=81820.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:06:34,865 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.92 vs. limit=5.0 +2023-03-26 18:06:37,141 INFO [finetune.py:976] (4/7) Epoch 15, batch 1650, loss[loss=0.1702, simple_loss=0.232, pruned_loss=0.0542, over 4933.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.2501, pruned_loss=0.05673, over 956664.23 frames. ], batch size: 38, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:06:40,763 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.362e+01 1.564e+02 1.826e+02 2.251e+02 4.924e+02, threshold=3.651e+02, percent-clipped=3.0 +2023-03-26 18:06:45,941 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 18:07:16,345 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4859, 2.1703, 2.8338, 1.6715, 2.4917, 2.7596, 1.9579, 2.7555], + device='cuda:4'), covar=tensor([0.1436, 0.2093, 0.1629, 0.2278, 0.1034, 0.1609, 0.2968, 0.1045], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0202, 0.0191, 0.0189, 0.0175, 0.0211, 0.0215, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:07:18,078 INFO [finetune.py:976] (4/7) Epoch 15, batch 1700, loss[loss=0.1593, simple_loss=0.222, pruned_loss=0.0483, over 4183.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2478, pruned_loss=0.05608, over 955345.49 frames. ], batch size: 18, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:07:51,480 INFO [finetune.py:976] (4/7) Epoch 15, batch 1750, loss[loss=0.1768, simple_loss=0.2414, pruned_loss=0.0561, over 4057.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.2511, pruned_loss=0.05774, over 953849.64 frames. ], batch size: 17, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:07:55,594 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.016e+02 1.510e+02 1.914e+02 2.293e+02 4.004e+02, threshold=3.828e+02, percent-clipped=1.0 +2023-03-26 18:07:56,320 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5627, 1.4467, 1.4309, 1.4805, 1.1745, 3.1502, 1.3095, 1.6542], + device='cuda:4'), covar=tensor([0.3186, 0.2436, 0.2068, 0.2326, 0.1679, 0.0210, 0.2729, 0.1290], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0116, 0.0120, 0.0124, 0.0115, 0.0097, 0.0097, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 18:08:02,974 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81955.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:08:03,554 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2177, 2.6703, 2.5469, 1.4287, 2.7709, 2.1588, 0.9933, 1.8671], + device='cuda:4'), covar=tensor([0.3562, 0.2152, 0.1901, 0.3180, 0.1445, 0.1043, 0.3705, 0.1582], + device='cuda:4'), in_proj_covar=tensor([0.0149, 0.0172, 0.0157, 0.0127, 0.0156, 0.0121, 0.0144, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 18:08:04,202 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81957.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:08:24,910 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81987.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:08:25,414 INFO [finetune.py:976] (4/7) Epoch 15, batch 1800, loss[loss=0.191, simple_loss=0.2647, pruned_loss=0.05868, over 4916.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2552, pruned_loss=0.05874, over 954896.78 frames. ], batch size: 37, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:08:36,206 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82003.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:08:37,855 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82005.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:08:52,964 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82026.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 18:09:00,024 INFO [finetune.py:976] (4/7) Epoch 15, batch 1850, loss[loss=0.2114, simple_loss=0.2827, pruned_loss=0.07007, over 4698.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2568, pruned_loss=0.05939, over 954395.04 frames. ], batch size: 59, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:09:03,675 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.664e+02 1.894e+02 2.440e+02 3.763e+02, threshold=3.787e+02, percent-clipped=0.0 +2023-03-26 18:09:06,714 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82048.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:09:11,016 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82055.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:09:33,281 INFO [finetune.py:976] (4/7) Epoch 15, batch 1900, loss[loss=0.1753, simple_loss=0.2473, pruned_loss=0.05168, over 4739.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2569, pruned_loss=0.05914, over 954505.48 frames. ], batch size: 27, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:09:40,543 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8304, 1.7154, 2.2545, 1.4971, 2.0127, 2.1836, 1.6201, 2.3317], + device='cuda:4'), covar=tensor([0.1468, 0.2052, 0.1327, 0.1823, 0.0896, 0.1467, 0.2884, 0.0922], + device='cuda:4'), in_proj_covar=tensor([0.0194, 0.0205, 0.0193, 0.0191, 0.0177, 0.0213, 0.0217, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:09:51,837 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82116.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:10:16,117 INFO [finetune.py:976] (4/7) Epoch 15, batch 1950, loss[loss=0.1953, simple_loss=0.2585, pruned_loss=0.06606, over 4872.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2557, pruned_loss=0.05899, over 954232.57 frames. ], batch size: 34, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:10:24,223 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.964e+01 1.525e+02 1.906e+02 2.226e+02 4.434e+02, threshold=3.812e+02, percent-clipped=2.0 +2023-03-26 18:11:01,530 INFO [finetune.py:976] (4/7) Epoch 15, batch 2000, loss[loss=0.1568, simple_loss=0.2119, pruned_loss=0.0509, over 4712.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2536, pruned_loss=0.05882, over 954549.30 frames. ], batch size: 23, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:11:16,205 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0596, 2.0133, 1.4628, 1.9890, 2.0376, 1.7165, 2.6945, 2.0845], + device='cuda:4'), covar=tensor([0.1508, 0.2267, 0.3802, 0.3461, 0.3092, 0.1823, 0.2680, 0.2142], + device='cuda:4'), in_proj_covar=tensor([0.0182, 0.0188, 0.0236, 0.0255, 0.0247, 0.0201, 0.0214, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:11:38,383 INFO [finetune.py:976] (4/7) Epoch 15, batch 2050, loss[loss=0.1743, simple_loss=0.2336, pruned_loss=0.05743, over 4254.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2513, pruned_loss=0.05841, over 954220.49 frames. ], batch size: 65, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:11:42,514 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.084e+02 1.417e+02 1.727e+02 2.274e+02 4.171e+02, threshold=3.454e+02, percent-clipped=1.0 +2023-03-26 18:12:11,847 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1920, 2.0254, 1.7635, 1.8689, 1.8840, 1.9109, 1.9564, 2.6543], + device='cuda:4'), covar=tensor([0.4060, 0.4600, 0.3510, 0.4127, 0.3995, 0.2658, 0.3897, 0.1756], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0259, 0.0226, 0.0275, 0.0247, 0.0214, 0.0248, 0.0225], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:12:24,923 INFO [finetune.py:976] (4/7) Epoch 15, batch 2100, loss[loss=0.1962, simple_loss=0.2585, pruned_loss=0.06693, over 4897.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2498, pruned_loss=0.05789, over 954728.69 frames. ], batch size: 32, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:12:31,366 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0711, 0.8891, 0.9484, 1.0413, 1.2484, 1.1318, 1.0159, 0.9525], + device='cuda:4'), covar=tensor([0.0360, 0.0362, 0.0524, 0.0330, 0.0288, 0.0464, 0.0379, 0.0392], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0108, 0.0141, 0.0112, 0.0099, 0.0106, 0.0096, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.2202e-05, 8.3721e-05, 1.1166e-04, 8.7037e-05, 7.7468e-05, 7.8528e-05, + 7.2343e-05, 8.2500e-05], device='cuda:4') +2023-03-26 18:12:54,225 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=82326.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:13:02,490 INFO [finetune.py:976] (4/7) Epoch 15, batch 2150, loss[loss=0.2125, simple_loss=0.2736, pruned_loss=0.07574, over 4841.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2531, pruned_loss=0.05888, over 954532.90 frames. ], batch size: 44, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:13:06,116 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82343.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:13:06,662 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.137e+02 1.626e+02 1.861e+02 2.291e+02 4.001e+02, threshold=3.721e+02, percent-clipped=2.0 +2023-03-26 18:13:07,990 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82346.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:13:25,987 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82374.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:13:35,326 INFO [finetune.py:976] (4/7) Epoch 15, batch 2200, loss[loss=0.1986, simple_loss=0.2663, pruned_loss=0.06544, over 4751.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2538, pruned_loss=0.05866, over 954796.86 frames. ], batch size: 54, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:13:48,069 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82407.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:13:50,412 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82411.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:14:08,110 INFO [finetune.py:976] (4/7) Epoch 15, batch 2250, loss[loss=0.214, simple_loss=0.2771, pruned_loss=0.07544, over 4888.00 frames. ], tot_loss[loss=0.1874, simple_loss=0.2559, pruned_loss=0.05945, over 954891.78 frames. ], batch size: 35, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:14:12,178 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.505e+02 1.711e+02 2.071e+02 3.892e+02, threshold=3.421e+02, percent-clipped=2.0 +2023-03-26 18:14:41,717 INFO [finetune.py:976] (4/7) Epoch 15, batch 2300, loss[loss=0.1698, simple_loss=0.2446, pruned_loss=0.04754, over 4722.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2567, pruned_loss=0.05927, over 953584.07 frames. ], batch size: 59, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:14:45,272 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82493.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 18:14:53,049 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 18:15:17,447 INFO [finetune.py:976] (4/7) Epoch 15, batch 2350, loss[loss=0.2217, simple_loss=0.2664, pruned_loss=0.08852, over 4219.00 frames. ], tot_loss[loss=0.1868, simple_loss=0.2548, pruned_loss=0.05939, over 954274.99 frames. ], batch size: 66, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:15:21,100 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.644e+02 1.984e+02 2.390e+02 4.799e+02, threshold=3.967e+02, percent-clipped=3.0 +2023-03-26 18:15:28,712 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82554.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 18:15:44,878 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82569.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:15:55,017 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3057, 2.3056, 1.8881, 2.5975, 2.3083, 2.0093, 2.9660, 2.3795], + device='cuda:4'), covar=tensor([0.1602, 0.2606, 0.3262, 0.2569, 0.2641, 0.1836, 0.2961, 0.1996], + device='cuda:4'), in_proj_covar=tensor([0.0181, 0.0187, 0.0235, 0.0254, 0.0245, 0.0200, 0.0213, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:16:00,853 INFO [finetune.py:976] (4/7) Epoch 15, batch 2400, loss[loss=0.1896, simple_loss=0.2575, pruned_loss=0.06086, over 4915.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2521, pruned_loss=0.05839, over 954253.27 frames. ], batch size: 46, lr: 3.50e-03, grad_scale: 32.0 +2023-03-26 18:16:06,623 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-26 18:16:37,308 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82630.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:16:42,547 INFO [finetune.py:976] (4/7) Epoch 15, batch 2450, loss[loss=0.1977, simple_loss=0.2608, pruned_loss=0.06733, over 4905.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2498, pruned_loss=0.05815, over 954252.83 frames. ], batch size: 36, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:16:43,891 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5065, 2.7554, 2.5271, 1.8376, 2.7701, 2.9205, 2.7747, 2.5170], + device='cuda:4'), covar=tensor([0.0725, 0.0599, 0.0829, 0.1057, 0.0643, 0.0803, 0.0658, 0.0971], + device='cuda:4'), in_proj_covar=tensor([0.0136, 0.0136, 0.0143, 0.0125, 0.0125, 0.0142, 0.0143, 0.0165], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:16:45,688 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=82643.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:16:46,163 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.277e+01 1.596e+02 1.937e+02 2.264e+02 4.235e+02, threshold=3.875e+02, percent-clipped=1.0 +2023-03-26 18:17:18,146 INFO [finetune.py:976] (4/7) Epoch 15, batch 2500, loss[loss=0.1669, simple_loss=0.2236, pruned_loss=0.05508, over 4753.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.252, pruned_loss=0.05891, over 954936.39 frames. ], batch size: 23, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:17:20,060 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82691.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:17:29,603 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4114, 1.4780, 1.8639, 1.2326, 1.4467, 1.7864, 1.3499, 1.9699], + device='cuda:4'), covar=tensor([0.1437, 0.2098, 0.1293, 0.1849, 0.1110, 0.1321, 0.2983, 0.0835], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0202, 0.0191, 0.0190, 0.0175, 0.0211, 0.0215, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:17:35,741 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82702.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:17:46,345 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=82711.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:18:03,674 INFO [finetune.py:976] (4/7) Epoch 15, batch 2550, loss[loss=0.2215, simple_loss=0.2896, pruned_loss=0.07673, over 4834.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2534, pruned_loss=0.05859, over 955658.18 frames. ], batch size: 30, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:18:07,766 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.099e+02 1.574e+02 1.850e+02 2.269e+02 4.152e+02, threshold=3.700e+02, percent-clipped=3.0 +2023-03-26 18:18:17,862 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=82759.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:18:25,914 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 18:18:36,858 INFO [finetune.py:976] (4/7) Epoch 15, batch 2600, loss[loss=0.239, simple_loss=0.2908, pruned_loss=0.09363, over 4759.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2556, pruned_loss=0.05945, over 956040.37 frames. ], batch size: 27, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:19:10,653 INFO [finetune.py:976] (4/7) Epoch 15, batch 2650, loss[loss=0.1602, simple_loss=0.2392, pruned_loss=0.0406, over 4856.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2565, pruned_loss=0.05988, over 956595.45 frames. ], batch size: 44, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:19:14,267 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.056e+02 1.579e+02 1.879e+02 2.251e+02 6.929e+02, threshold=3.759e+02, percent-clipped=2.0 +2023-03-26 18:19:17,847 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82849.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 18:19:18,475 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4859, 1.3998, 1.6825, 2.5003, 1.7255, 2.2322, 0.8086, 2.1649], + device='cuda:4'), covar=tensor([0.1711, 0.1388, 0.1140, 0.0718, 0.0864, 0.1147, 0.1654, 0.0600], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0117, 0.0133, 0.0164, 0.0100, 0.0138, 0.0124, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 18:19:36,859 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 18:19:43,179 INFO [finetune.py:976] (4/7) Epoch 15, batch 2700, loss[loss=0.185, simple_loss=0.2618, pruned_loss=0.05413, over 4858.00 frames. ], tot_loss[loss=0.1874, simple_loss=0.256, pruned_loss=0.05945, over 953890.12 frames. ], batch size: 31, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:19:56,928 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1619, 1.8835, 1.7365, 1.9486, 1.9014, 1.8565, 1.8718, 2.5669], + device='cuda:4'), covar=tensor([0.3515, 0.4637, 0.3323, 0.4091, 0.4265, 0.2431, 0.4401, 0.1733], + device='cuda:4'), in_proj_covar=tensor([0.0284, 0.0258, 0.0224, 0.0274, 0.0246, 0.0213, 0.0247, 0.0224], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:20:08,061 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82925.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:20:16,388 INFO [finetune.py:976] (4/7) Epoch 15, batch 2750, loss[loss=0.1569, simple_loss=0.2301, pruned_loss=0.04188, over 4817.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2532, pruned_loss=0.05833, over 955470.24 frames. ], batch size: 40, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:20:20,499 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.574e+02 1.758e+02 2.107e+02 4.076e+02, threshold=3.515e+02, percent-clipped=2.0 +2023-03-26 18:20:33,251 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82963.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:20:49,686 INFO [finetune.py:976] (4/7) Epoch 15, batch 2800, loss[loss=0.1791, simple_loss=0.2422, pruned_loss=0.05801, over 4925.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2497, pruned_loss=0.05682, over 957018.50 frames. ], batch size: 37, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:21:07,087 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83002.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:21:08,659 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.58 vs. limit=5.0 +2023-03-26 18:21:11,847 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1627, 2.1331, 1.8724, 2.2468, 2.0688, 2.0854, 2.0527, 2.9401], + device='cuda:4'), covar=tensor([0.3923, 0.4915, 0.3428, 0.4443, 0.4765, 0.2398, 0.4827, 0.1663], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0258, 0.0224, 0.0274, 0.0246, 0.0213, 0.0247, 0.0224], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:21:15,075 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-26 18:21:21,931 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83024.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:21:37,767 INFO [finetune.py:976] (4/7) Epoch 15, batch 2850, loss[loss=0.1501, simple_loss=0.225, pruned_loss=0.03763, over 4857.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2485, pruned_loss=0.05666, over 954987.44 frames. ], batch size: 31, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:21:41,396 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.604e+02 1.866e+02 2.227e+02 4.125e+02, threshold=3.733e+02, percent-clipped=3.0 +2023-03-26 18:21:41,611 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-26 18:21:49,030 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83050.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:21:49,319 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-26 18:21:53,941 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1364, 1.8949, 2.4661, 1.5749, 2.2042, 2.4961, 1.8685, 2.6290], + device='cuda:4'), covar=tensor([0.1339, 0.1745, 0.1413, 0.1917, 0.0960, 0.1324, 0.2424, 0.0831], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0202, 0.0192, 0.0189, 0.0174, 0.0212, 0.0216, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:22:07,018 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.54 vs. limit=2.0 +2023-03-26 18:22:15,028 INFO [finetune.py:976] (4/7) Epoch 15, batch 2900, loss[loss=0.2267, simple_loss=0.2938, pruned_loss=0.0798, over 4786.00 frames. ], tot_loss[loss=0.1838, simple_loss=0.2516, pruned_loss=0.05802, over 954088.19 frames. ], batch size: 54, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:22:57,709 INFO [finetune.py:976] (4/7) Epoch 15, batch 2950, loss[loss=0.1883, simple_loss=0.2548, pruned_loss=0.06091, over 4758.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2544, pruned_loss=0.05894, over 955334.96 frames. ], batch size: 23, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:23:01,328 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.293e+02 1.748e+02 2.030e+02 2.368e+02 3.585e+02, threshold=4.059e+02, percent-clipped=0.0 +2023-03-26 18:23:08,898 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83149.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 18:23:37,597 INFO [finetune.py:976] (4/7) Epoch 15, batch 3000, loss[loss=0.1548, simple_loss=0.2254, pruned_loss=0.04206, over 4869.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2562, pruned_loss=0.05986, over 955295.86 frames. ], batch size: 34, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:23:37,597 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 18:23:48,369 INFO [finetune.py:1010] (4/7) Epoch 15, validation: loss=0.1564, simple_loss=0.2269, pruned_loss=0.04296, over 2265189.00 frames. +2023-03-26 18:23:48,369 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 18:23:59,741 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83196.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:24:00,271 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83197.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 18:24:21,534 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83225.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:24:30,317 INFO [finetune.py:976] (4/7) Epoch 15, batch 3050, loss[loss=0.189, simple_loss=0.2709, pruned_loss=0.05351, over 4770.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2567, pruned_loss=0.05962, over 955936.39 frames. ], batch size: 28, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:24:34,916 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.542e+02 1.763e+02 2.135e+02 3.801e+02, threshold=3.526e+02, percent-clipped=0.0 +2023-03-26 18:24:44,062 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83257.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:24:54,192 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83273.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:24:59,160 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-03-26 18:25:01,279 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83284.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:25:04,097 INFO [finetune.py:976] (4/7) Epoch 15, batch 3100, loss[loss=0.176, simple_loss=0.2337, pruned_loss=0.05913, over 4869.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2552, pruned_loss=0.0588, over 955261.63 frames. ], batch size: 34, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:25:24,780 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=83319.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:25:27,711 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2056, 1.2649, 1.5912, 0.9772, 1.2096, 1.4088, 1.3152, 1.5934], + device='cuda:4'), covar=tensor([0.1066, 0.2002, 0.1095, 0.1331, 0.0893, 0.1126, 0.2523, 0.0779], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0203, 0.0193, 0.0190, 0.0176, 0.0212, 0.0216, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:25:37,275 INFO [finetune.py:976] (4/7) Epoch 15, batch 3150, loss[loss=0.1891, simple_loss=0.2558, pruned_loss=0.06116, over 4742.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2537, pruned_loss=0.05868, over 958098.90 frames. ], batch size: 26, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:25:41,381 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.111e+02 1.524e+02 1.821e+02 2.258e+02 3.585e+02, threshold=3.643e+02, percent-clipped=2.0 +2023-03-26 18:25:42,607 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83345.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:25:59,537 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-26 18:26:12,521 INFO [finetune.py:976] (4/7) Epoch 15, batch 3200, loss[loss=0.1655, simple_loss=0.23, pruned_loss=0.05045, over 4824.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2504, pruned_loss=0.05776, over 956077.99 frames. ], batch size: 30, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:26:55,877 INFO [finetune.py:976] (4/7) Epoch 15, batch 3250, loss[loss=0.1542, simple_loss=0.2357, pruned_loss=0.03632, over 4757.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2504, pruned_loss=0.05761, over 957064.37 frames. ], batch size: 28, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:27:00,085 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.648e+01 1.539e+02 1.854e+02 2.232e+02 3.646e+02, threshold=3.708e+02, percent-clipped=1.0 +2023-03-26 18:27:29,590 INFO [finetune.py:976] (4/7) Epoch 15, batch 3300, loss[loss=0.2789, simple_loss=0.3387, pruned_loss=0.1095, over 4900.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2554, pruned_loss=0.05999, over 955316.30 frames. ], batch size: 43, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:27:38,152 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 18:27:49,507 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8695, 1.2370, 1.9065, 1.8244, 1.6069, 1.5625, 1.7186, 1.6281], + device='cuda:4'), covar=tensor([0.3827, 0.4028, 0.3357, 0.3617, 0.4890, 0.3634, 0.4545, 0.3311], + device='cuda:4'), in_proj_covar=tensor([0.0245, 0.0239, 0.0257, 0.0269, 0.0268, 0.0240, 0.0281, 0.0236], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:28:07,467 INFO [finetune.py:976] (4/7) Epoch 15, batch 3350, loss[loss=0.1719, simple_loss=0.2453, pruned_loss=0.04919, over 4863.00 frames. ], tot_loss[loss=0.1897, simple_loss=0.2578, pruned_loss=0.06085, over 955066.31 frames. ], batch size: 34, lr: 3.49e-03, grad_scale: 64.0 +2023-03-26 18:28:14,162 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3133, 2.2530, 2.3994, 1.5819, 2.3079, 2.5374, 2.4492, 1.9937], + device='cuda:4'), covar=tensor([0.0620, 0.0647, 0.0687, 0.0999, 0.0635, 0.0739, 0.0667, 0.1134], + device='cuda:4'), in_proj_covar=tensor([0.0136, 0.0136, 0.0143, 0.0125, 0.0125, 0.0143, 0.0143, 0.0165], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:28:14,630 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.792e+02 2.041e+02 2.510e+02 5.102e+02, threshold=4.082e+02, percent-clipped=3.0 +2023-03-26 18:28:21,872 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=83552.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:28:54,246 INFO [finetune.py:976] (4/7) Epoch 15, batch 3400, loss[loss=0.1613, simple_loss=0.2389, pruned_loss=0.04182, over 4795.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.258, pruned_loss=0.0606, over 955507.19 frames. ], batch size: 51, lr: 3.49e-03, grad_scale: 64.0 +2023-03-26 18:29:24,887 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83619.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:29:37,311 INFO [finetune.py:976] (4/7) Epoch 15, batch 3450, loss[loss=0.2004, simple_loss=0.2663, pruned_loss=0.06722, over 4820.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2583, pruned_loss=0.06084, over 954847.99 frames. ], batch size: 39, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:29:39,047 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=83640.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:29:42,000 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.651e+02 1.928e+02 2.236e+02 3.717e+02, threshold=3.855e+02, percent-clipped=0.0 +2023-03-26 18:29:57,378 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83667.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:30:11,003 INFO [finetune.py:976] (4/7) Epoch 15, batch 3500, loss[loss=0.2083, simple_loss=0.2656, pruned_loss=0.07551, over 4833.00 frames. ], tot_loss[loss=0.1882, simple_loss=0.2559, pruned_loss=0.06023, over 954845.17 frames. ], batch size: 39, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:30:44,676 INFO [finetune.py:976] (4/7) Epoch 15, batch 3550, loss[loss=0.1751, simple_loss=0.2466, pruned_loss=0.05177, over 4803.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.252, pruned_loss=0.05876, over 955607.57 frames. ], batch size: 51, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:30:49,417 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.222e+01 1.573e+02 1.880e+02 2.102e+02 4.250e+02, threshold=3.760e+02, percent-clipped=2.0 +2023-03-26 18:30:50,172 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1982, 1.7980, 2.1627, 2.1092, 1.8779, 1.8975, 2.1142, 2.0182], + device='cuda:4'), covar=tensor([0.4405, 0.4667, 0.3651, 0.4532, 0.5808, 0.3959, 0.5388, 0.3621], + device='cuda:4'), in_proj_covar=tensor([0.0245, 0.0239, 0.0257, 0.0269, 0.0268, 0.0240, 0.0281, 0.0236], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:30:53,746 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9424, 1.7702, 2.2422, 3.6103, 2.4347, 2.6078, 1.1413, 2.8729], + device='cuda:4'), covar=tensor([0.1663, 0.1495, 0.1364, 0.0528, 0.0759, 0.1276, 0.1819, 0.0548], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0118, 0.0133, 0.0165, 0.0101, 0.0138, 0.0125, 0.0103], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 18:30:59,110 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7466, 1.6884, 1.6473, 1.7820, 1.1165, 3.8725, 1.5496, 2.0223], + device='cuda:4'), covar=tensor([0.3506, 0.2645, 0.2103, 0.2348, 0.1912, 0.0169, 0.2454, 0.1291], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0115, 0.0119, 0.0123, 0.0114, 0.0097, 0.0096, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 18:30:59,127 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83760.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:31:18,472 INFO [finetune.py:976] (4/7) Epoch 15, batch 3600, loss[loss=0.1943, simple_loss=0.2618, pruned_loss=0.0634, over 4811.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2502, pruned_loss=0.05837, over 955495.68 frames. ], batch size: 38, lr: 3.49e-03, grad_scale: 32.0 +2023-03-26 18:31:35,702 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-26 18:31:48,181 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83821.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:31:48,255 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 18:31:59,981 INFO [finetune.py:976] (4/7) Epoch 15, batch 3650, loss[loss=0.1596, simple_loss=0.2306, pruned_loss=0.04433, over 4763.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2523, pruned_loss=0.05906, over 955576.70 frames. ], batch size: 28, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:32:04,778 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.180e+02 1.603e+02 1.941e+02 2.306e+02 4.863e+02, threshold=3.882e+02, percent-clipped=1.0 +2023-03-26 18:32:09,567 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83852.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:32:33,854 INFO [finetune.py:976] (4/7) Epoch 15, batch 3700, loss[loss=0.1938, simple_loss=0.2584, pruned_loss=0.06455, over 4909.00 frames. ], tot_loss[loss=0.188, simple_loss=0.2561, pruned_loss=0.05988, over 955919.66 frames. ], batch size: 35, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:32:41,069 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6512, 3.8869, 3.7080, 2.0927, 4.0339, 3.0067, 0.7174, 2.6712], + device='cuda:4'), covar=tensor([0.2359, 0.1815, 0.1279, 0.2909, 0.0837, 0.1033, 0.4230, 0.1440], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0174, 0.0159, 0.0127, 0.0158, 0.0122, 0.0145, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 18:32:42,153 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83900.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:32:55,991 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6414, 1.5510, 1.4245, 1.8569, 2.0105, 1.7205, 1.2030, 1.4109], + device='cuda:4'), covar=tensor([0.2236, 0.2117, 0.1953, 0.1588, 0.1666, 0.1235, 0.2706, 0.1939], + device='cuda:4'), in_proj_covar=tensor([0.0240, 0.0208, 0.0210, 0.0191, 0.0242, 0.0185, 0.0216, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:32:57,381 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-26 18:33:07,588 INFO [finetune.py:976] (4/7) Epoch 15, batch 3750, loss[loss=0.2015, simple_loss=0.2699, pruned_loss=0.06657, over 4831.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2583, pruned_loss=0.06049, over 955926.75 frames. ], batch size: 49, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:33:08,907 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83940.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:33:11,800 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.596e+02 1.977e+02 2.275e+02 5.079e+02, threshold=3.955e+02, percent-clipped=1.0 +2023-03-26 18:33:14,838 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83949.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:33:53,430 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83985.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:33:55,619 INFO [finetune.py:976] (4/7) Epoch 15, batch 3800, loss[loss=0.2222, simple_loss=0.29, pruned_loss=0.0772, over 4866.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2576, pruned_loss=0.05971, over 955951.53 frames. ], batch size: 34, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:33:55,680 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=83988.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:33:58,160 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7713, 1.8927, 1.4928, 1.6843, 2.3747, 2.3721, 1.9110, 1.8665], + device='cuda:4'), covar=tensor([0.0411, 0.0345, 0.0633, 0.0335, 0.0271, 0.0385, 0.0293, 0.0360], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0108, 0.0140, 0.0112, 0.0099, 0.0106, 0.0097, 0.0107], + device='cuda:4'), out_proj_covar=tensor([7.1812e-05, 8.3370e-05, 1.1110e-04, 8.6647e-05, 7.7353e-05, 7.7958e-05, + 7.2769e-05, 8.1821e-05], device='cuda:4') +2023-03-26 18:34:11,201 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84010.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 18:34:22,277 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7053, 1.6208, 2.1966, 3.2754, 2.1448, 2.3799, 1.2202, 2.6349], + device='cuda:4'), covar=tensor([0.1866, 0.1615, 0.1358, 0.0762, 0.0928, 0.1502, 0.1889, 0.0634], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0117, 0.0132, 0.0163, 0.0100, 0.0137, 0.0124, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 18:34:36,933 INFO [finetune.py:976] (4/7) Epoch 15, batch 3850, loss[loss=0.1706, simple_loss=0.2378, pruned_loss=0.05167, over 4754.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.256, pruned_loss=0.05908, over 955467.96 frames. ], batch size: 54, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:34:41,718 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.557e+01 1.496e+02 1.862e+02 2.338e+02 3.560e+02, threshold=3.724e+02, percent-clipped=0.0 +2023-03-26 18:34:42,452 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84046.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:35:10,036 INFO [finetune.py:976] (4/7) Epoch 15, batch 3900, loss[loss=0.1976, simple_loss=0.2588, pruned_loss=0.06818, over 4888.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2534, pruned_loss=0.05863, over 956720.07 frames. ], batch size: 32, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:35:15,960 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2464, 2.1042, 1.7798, 2.1063, 2.1827, 1.9395, 2.5779, 2.2280], + device='cuda:4'), covar=tensor([0.1302, 0.2243, 0.2986, 0.2668, 0.2535, 0.1619, 0.2816, 0.1783], + device='cuda:4'), in_proj_covar=tensor([0.0181, 0.0188, 0.0235, 0.0254, 0.0245, 0.0201, 0.0212, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:35:16,537 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84097.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:35:28,909 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84116.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:35:43,584 INFO [finetune.py:976] (4/7) Epoch 15, batch 3950, loss[loss=0.2088, simple_loss=0.2584, pruned_loss=0.07956, over 4823.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2503, pruned_loss=0.05753, over 957122.78 frames. ], batch size: 40, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:35:47,770 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.469e+02 1.885e+02 2.278e+02 4.120e+02, threshold=3.770e+02, percent-clipped=1.0 +2023-03-26 18:35:51,955 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 18:35:57,233 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84158.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:36:13,534 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 18:36:16,808 INFO [finetune.py:976] (4/7) Epoch 15, batch 4000, loss[loss=0.2122, simple_loss=0.2867, pruned_loss=0.06884, over 4814.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2502, pruned_loss=0.05771, over 956367.69 frames. ], batch size: 41, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:36:24,509 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84199.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:36:26,858 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2054, 1.8606, 1.9592, 0.8568, 2.1786, 2.3308, 2.0547, 1.8027], + device='cuda:4'), covar=tensor([0.1035, 0.0813, 0.0629, 0.0750, 0.0559, 0.0683, 0.0452, 0.0805], + device='cuda:4'), in_proj_covar=tensor([0.0124, 0.0151, 0.0124, 0.0128, 0.0130, 0.0126, 0.0141, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.1550e-05, 1.0959e-04, 8.8792e-05, 9.1931e-05, 9.2165e-05, 9.0842e-05, + 1.0179e-04, 1.0597e-04], device='cuda:4') +2023-03-26 18:36:57,632 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84236.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:36:59,240 INFO [finetune.py:976] (4/7) Epoch 15, batch 4050, loss[loss=0.2149, simple_loss=0.2942, pruned_loss=0.06779, over 4729.00 frames. ], tot_loss[loss=0.1867, simple_loss=0.255, pruned_loss=0.05917, over 955509.91 frames. ], batch size: 59, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:37:07,827 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.586e+02 1.909e+02 2.268e+02 5.729e+02, threshold=3.818e+02, percent-clipped=2.0 +2023-03-26 18:37:21,558 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84260.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:37:23,904 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-03-26 18:37:39,952 INFO [finetune.py:976] (4/7) Epoch 15, batch 4100, loss[loss=0.1717, simple_loss=0.2444, pruned_loss=0.04949, over 4806.00 frames. ], tot_loss[loss=0.1898, simple_loss=0.2588, pruned_loss=0.06038, over 955914.62 frames. ], batch size: 41, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:37:46,465 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84297.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:37:52,177 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84305.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 18:38:10,151 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-26 18:38:11,861 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84336.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:38:13,431 INFO [finetune.py:976] (4/7) Epoch 15, batch 4150, loss[loss=0.2107, simple_loss=0.2829, pruned_loss=0.06922, over 4935.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2596, pruned_loss=0.06076, over 953548.94 frames. ], batch size: 42, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:38:15,327 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84341.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:38:18,113 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.618e+02 1.997e+02 2.307e+02 7.274e+02, threshold=3.993e+02, percent-clipped=3.0 +2023-03-26 18:38:48,343 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 18:38:50,308 INFO [finetune.py:976] (4/7) Epoch 15, batch 4200, loss[loss=0.1682, simple_loss=0.2347, pruned_loss=0.05079, over 4762.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2579, pruned_loss=0.05952, over 954204.29 frames. ], batch size: 28, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:39:04,800 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84397.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:39:17,870 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84416.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:39:20,217 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7378, 3.8596, 3.6135, 1.6271, 3.9970, 2.9285, 0.7714, 2.8696], + device='cuda:4'), covar=tensor([0.2481, 0.1670, 0.1543, 0.3299, 0.0927, 0.1000, 0.4388, 0.1237], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0173, 0.0159, 0.0127, 0.0157, 0.0121, 0.0145, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 18:39:31,951 INFO [finetune.py:976] (4/7) Epoch 15, batch 4250, loss[loss=0.1564, simple_loss=0.2327, pruned_loss=0.04003, over 4816.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2548, pruned_loss=0.05849, over 955064.13 frames. ], batch size: 39, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:39:36,665 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.324e+01 1.559e+02 1.825e+02 2.300e+02 4.289e+02, threshold=3.650e+02, percent-clipped=1.0 +2023-03-26 18:39:47,007 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 18:39:47,480 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84453.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:39:58,790 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84464.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:40:14,295 INFO [finetune.py:976] (4/7) Epoch 15, batch 4300, loss[loss=0.1811, simple_loss=0.2557, pruned_loss=0.05324, over 4801.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2529, pruned_loss=0.05841, over 957067.73 frames. ], batch size: 51, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:40:22,751 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2737, 2.0856, 2.8489, 1.7152, 2.5161, 2.6270, 2.0607, 2.8941], + device='cuda:4'), covar=tensor([0.1168, 0.1701, 0.1375, 0.1903, 0.0794, 0.1317, 0.2393, 0.0712], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0203, 0.0193, 0.0189, 0.0176, 0.0212, 0.0216, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:40:36,303 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.31 vs. limit=5.0 +2023-03-26 18:40:39,104 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84525.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:40:39,729 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9088, 1.9909, 1.7270, 1.7077, 2.5007, 2.3858, 2.1105, 1.9743], + device='cuda:4'), covar=tensor([0.0386, 0.0326, 0.0557, 0.0363, 0.0246, 0.0515, 0.0302, 0.0369], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0108, 0.0142, 0.0113, 0.0100, 0.0106, 0.0097, 0.0107], + device='cuda:4'), out_proj_covar=tensor([7.2130e-05, 8.3931e-05, 1.1208e-04, 8.7162e-05, 7.7899e-05, 7.8458e-05, + 7.2903e-05, 8.1888e-05], device='cuda:4') +2023-03-26 18:40:42,173 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2063, 2.2098, 1.8753, 2.3790, 2.0777, 2.0581, 2.0371, 3.0373], + device='cuda:4'), covar=tensor([0.4062, 0.5265, 0.3550, 0.4534, 0.4650, 0.2445, 0.4710, 0.1877], + device='cuda:4'), in_proj_covar=tensor([0.0284, 0.0258, 0.0225, 0.0275, 0.0247, 0.0215, 0.0248, 0.0226], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:40:47,840 INFO [finetune.py:976] (4/7) Epoch 15, batch 4350, loss[loss=0.1595, simple_loss=0.2291, pruned_loss=0.0449, over 4816.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.2491, pruned_loss=0.05728, over 956235.21 frames. ], batch size: 38, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:40:51,075 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6407, 1.6153, 2.1672, 1.7711, 1.8128, 4.0533, 1.5129, 1.7827], + device='cuda:4'), covar=tensor([0.0984, 0.1956, 0.1220, 0.1096, 0.1652, 0.0233, 0.1634, 0.1804], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0082, 0.0074, 0.0078, 0.0092, 0.0081, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 18:40:52,220 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.502e+02 1.820e+02 2.196e+02 3.984e+02, threshold=3.641e+02, percent-clipped=2.0 +2023-03-26 18:40:58,847 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84555.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:41:19,627 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84586.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:41:21,100 INFO [finetune.py:976] (4/7) Epoch 15, batch 4400, loss[loss=0.1967, simple_loss=0.2649, pruned_loss=0.06424, over 4890.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2496, pruned_loss=0.05726, over 955587.75 frames. ], batch size: 32, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:41:24,131 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84592.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:41:32,779 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84605.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:41:39,328 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4982, 1.3678, 1.2217, 1.5291, 1.5844, 1.4573, 0.8430, 1.2400], + device='cuda:4'), covar=tensor([0.2121, 0.2121, 0.1951, 0.1596, 0.1652, 0.1277, 0.2726, 0.1883], + device='cuda:4'), in_proj_covar=tensor([0.0239, 0.0206, 0.0209, 0.0191, 0.0240, 0.0184, 0.0214, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:41:51,835 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84633.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:41:54,827 INFO [finetune.py:976] (4/7) Epoch 15, batch 4450, loss[loss=0.1843, simple_loss=0.2609, pruned_loss=0.0539, over 4800.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2534, pruned_loss=0.05838, over 954924.32 frames. ], batch size: 51, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:41:57,750 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84641.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:42:01,982 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.595e+02 1.952e+02 2.292e+02 4.719e+02, threshold=3.904e+02, percent-clipped=1.0 +2023-03-26 18:42:11,294 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84653.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:42:21,342 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-03-26 18:42:30,785 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.4687, 3.9146, 4.1219, 4.2905, 4.2120, 3.9738, 4.5626, 1.3754], + device='cuda:4'), covar=tensor([0.0826, 0.0842, 0.0795, 0.1120, 0.1304, 0.1527, 0.0764, 0.5915], + device='cuda:4'), in_proj_covar=tensor([0.0351, 0.0244, 0.0276, 0.0292, 0.0334, 0.0284, 0.0300, 0.0299], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:42:46,431 INFO [finetune.py:976] (4/7) Epoch 15, batch 4500, loss[loss=0.1796, simple_loss=0.2614, pruned_loss=0.04887, over 4810.00 frames. ], tot_loss[loss=0.1871, simple_loss=0.2559, pruned_loss=0.05913, over 956527.44 frames. ], batch size: 41, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:42:47,112 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84689.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:42:49,439 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84692.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:42:51,213 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84694.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:43:20,117 INFO [finetune.py:976] (4/7) Epoch 15, batch 4550, loss[loss=0.1772, simple_loss=0.2578, pruned_loss=0.0483, over 4838.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2574, pruned_loss=0.05978, over 952393.02 frames. ], batch size: 49, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:43:25,283 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.592e+02 2.005e+02 2.406e+02 4.528e+02, threshold=4.009e+02, percent-clipped=3.0 +2023-03-26 18:43:30,246 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84753.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:43:53,704 INFO [finetune.py:976] (4/7) Epoch 15, batch 4600, loss[loss=0.1858, simple_loss=0.255, pruned_loss=0.05831, over 4860.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2575, pruned_loss=0.05985, over 951501.28 frames. ], batch size: 31, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:43:53,837 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3518, 2.2053, 1.8688, 2.5214, 2.3583, 1.9828, 2.8476, 2.3540], + device='cuda:4'), covar=tensor([0.1381, 0.2446, 0.3132, 0.2861, 0.2692, 0.1831, 0.3276, 0.1848], + device='cuda:4'), in_proj_covar=tensor([0.0181, 0.0188, 0.0234, 0.0254, 0.0245, 0.0202, 0.0213, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:44:06,845 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84800.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:44:07,397 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84801.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:44:25,140 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.2195, 1.2406, 1.3053, 0.5885, 1.1807, 1.4505, 1.5327, 1.2135], + device='cuda:4'), covar=tensor([0.0792, 0.0551, 0.0447, 0.0554, 0.0455, 0.0542, 0.0286, 0.0606], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0152, 0.0124, 0.0130, 0.0131, 0.0127, 0.0143, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.2556e-05, 1.1061e-04, 8.9163e-05, 9.2784e-05, 9.2856e-05, 9.1374e-05, + 1.0286e-04, 1.0672e-04], device='cuda:4') +2023-03-26 18:44:35,142 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6033, 1.5556, 1.9162, 1.2859, 1.7323, 1.8605, 1.5278, 2.0622], + device='cuda:4'), covar=tensor([0.1254, 0.2170, 0.1426, 0.1748, 0.0839, 0.1311, 0.2982, 0.0788], + device='cuda:4'), in_proj_covar=tensor([0.0194, 0.0205, 0.0194, 0.0191, 0.0177, 0.0214, 0.0219, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:44:36,214 INFO [finetune.py:976] (4/7) Epoch 15, batch 4650, loss[loss=0.1628, simple_loss=0.2313, pruned_loss=0.0472, over 4755.00 frames. ], tot_loss[loss=0.1873, simple_loss=0.2557, pruned_loss=0.05945, over 953121.20 frames. ], batch size: 54, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:44:40,391 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.584e+02 1.933e+02 2.372e+02 3.946e+02, threshold=3.865e+02, percent-clipped=0.0 +2023-03-26 18:44:47,570 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84855.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:44:56,242 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84861.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:45:17,972 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84881.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:45:22,915 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6279, 1.5296, 1.5022, 1.5997, 0.9793, 3.4021, 1.3623, 1.7345], + device='cuda:4'), covar=tensor([0.3438, 0.2565, 0.2161, 0.2450, 0.1982, 0.0211, 0.2506, 0.1325], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0116, 0.0120, 0.0124, 0.0115, 0.0097, 0.0097, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 18:45:25,659 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-26 18:45:26,375 INFO [finetune.py:976] (4/7) Epoch 15, batch 4700, loss[loss=0.1632, simple_loss=0.2304, pruned_loss=0.04796, over 4789.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2522, pruned_loss=0.05793, over 955329.99 frames. ], batch size: 29, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:45:29,344 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84892.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:45:36,525 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84903.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:45:39,709 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.68 vs. limit=2.0 +2023-03-26 18:45:46,247 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8779, 1.7403, 1.6623, 2.0964, 2.1286, 2.0487, 1.4300, 1.6008], + device='cuda:4'), covar=tensor([0.2199, 0.2093, 0.1900, 0.1565, 0.1888, 0.1110, 0.2668, 0.2004], + device='cuda:4'), in_proj_covar=tensor([0.0239, 0.0206, 0.0209, 0.0191, 0.0240, 0.0184, 0.0214, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:45:59,760 INFO [finetune.py:976] (4/7) Epoch 15, batch 4750, loss[loss=0.1864, simple_loss=0.2489, pruned_loss=0.06199, over 4807.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2493, pruned_loss=0.05656, over 955356.60 frames. ], batch size: 45, lr: 3.48e-03, grad_scale: 32.0 +2023-03-26 18:46:01,518 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=84940.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:46:04,960 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.000e+02 1.651e+02 1.892e+02 2.436e+02 4.596e+02, threshold=3.784e+02, percent-clipped=1.0 +2023-03-26 18:46:08,166 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0746, 1.9943, 2.0978, 1.5343, 2.1256, 2.2043, 2.2246, 1.7025], + device='cuda:4'), covar=tensor([0.0607, 0.0718, 0.0688, 0.0857, 0.0637, 0.0721, 0.0582, 0.1127], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0135, 0.0142, 0.0123, 0.0124, 0.0141, 0.0143, 0.0164], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:46:08,764 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84951.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:46:33,700 INFO [finetune.py:976] (4/7) Epoch 15, batch 4800, loss[loss=0.1785, simple_loss=0.2537, pruned_loss=0.05161, over 4895.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2526, pruned_loss=0.05825, over 955839.63 frames. ], batch size: 35, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:46:34,869 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84989.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:46:36,728 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84992.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:46:50,494 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85012.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:47:05,188 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-03-26 18:47:05,337 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2619, 2.9335, 2.7614, 1.2032, 3.0496, 2.2107, 0.6456, 1.8626], + device='cuda:4'), covar=tensor([0.2405, 0.2367, 0.1603, 0.3557, 0.1389, 0.1142, 0.4213, 0.1650], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0174, 0.0159, 0.0128, 0.0158, 0.0123, 0.0145, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 18:47:07,603 INFO [finetune.py:976] (4/7) Epoch 15, batch 4850, loss[loss=0.189, simple_loss=0.2473, pruned_loss=0.06538, over 4783.00 frames. ], tot_loss[loss=0.1884, simple_loss=0.2569, pruned_loss=0.05998, over 955242.30 frames. ], batch size: 54, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:47:09,347 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85040.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:47:10,612 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4655, 1.3861, 1.3487, 1.4940, 1.0470, 3.2634, 1.2778, 1.6980], + device='cuda:4'), covar=tensor([0.3672, 0.2747, 0.2423, 0.2648, 0.2109, 0.0229, 0.2663, 0.1379], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0115, 0.0120, 0.0124, 0.0115, 0.0097, 0.0097, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 18:47:10,635 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3334, 1.3429, 1.4151, 0.8680, 1.4499, 1.6095, 1.6943, 1.2575], + device='cuda:4'), covar=tensor([0.0887, 0.0640, 0.0552, 0.0480, 0.0447, 0.0608, 0.0386, 0.0701], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0153, 0.0125, 0.0130, 0.0132, 0.0128, 0.0144, 0.0148], + device='cuda:4'), out_proj_covar=tensor([9.3091e-05, 1.1145e-04, 8.9908e-05, 9.3404e-05, 9.3383e-05, 9.2077e-05, + 1.0377e-04, 1.0712e-04], device='cuda:4') +2023-03-26 18:47:12,312 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.155e+02 1.585e+02 1.858e+02 2.141e+02 6.123e+02, threshold=3.716e+02, percent-clipped=1.0 +2023-03-26 18:47:19,333 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 18:47:20,389 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7592, 0.7495, 1.7914, 1.6988, 1.5898, 1.5273, 1.5933, 1.7212], + device='cuda:4'), covar=tensor([0.3258, 0.3605, 0.3070, 0.3276, 0.4377, 0.3202, 0.3664, 0.2974], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0239, 0.0258, 0.0269, 0.0268, 0.0241, 0.0280, 0.0237], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:47:25,939 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 18:47:50,157 INFO [finetune.py:976] (4/7) Epoch 15, batch 4900, loss[loss=0.1751, simple_loss=0.2551, pruned_loss=0.04752, over 4795.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.2579, pruned_loss=0.06044, over 954431.74 frames. ], batch size: 51, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:47:53,466 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 18:47:59,543 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9162, 1.7047, 2.2598, 1.5299, 2.1363, 2.3187, 1.6342, 2.4389], + device='cuda:4'), covar=tensor([0.1370, 0.1971, 0.1682, 0.2019, 0.0951, 0.1380, 0.2889, 0.0890], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0203, 0.0192, 0.0190, 0.0176, 0.0213, 0.0217, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:48:06,024 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85107.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:48:11,343 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85115.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:48:11,961 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4079, 1.3136, 1.3219, 1.3561, 1.0110, 2.2437, 0.8053, 1.2520], + device='cuda:4'), covar=tensor([0.3328, 0.2491, 0.2178, 0.2431, 0.1741, 0.0349, 0.2556, 0.1331], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0115, 0.0120, 0.0124, 0.0115, 0.0097, 0.0096, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 18:48:26,708 INFO [finetune.py:976] (4/7) Epoch 15, batch 4950, loss[loss=0.1736, simple_loss=0.2466, pruned_loss=0.05027, over 4862.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2591, pruned_loss=0.06049, over 955373.51 frames. ], batch size: 34, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:48:31,439 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.624e+02 1.884e+02 2.194e+02 3.725e+02, threshold=3.769e+02, percent-clipped=1.0 +2023-03-26 18:48:39,186 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85156.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:48:47,016 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85168.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 18:48:52,399 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85176.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:48:55,803 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85181.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:49:00,448 INFO [finetune.py:976] (4/7) Epoch 15, batch 5000, loss[loss=0.1532, simple_loss=0.2228, pruned_loss=0.04185, over 4808.00 frames. ], tot_loss[loss=0.188, simple_loss=0.257, pruned_loss=0.05949, over 955146.63 frames. ], batch size: 25, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:49:26,013 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-03-26 18:49:36,187 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85229.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:49:37,746 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 18:49:41,005 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5421, 1.6048, 2.2515, 1.9053, 1.8582, 4.1801, 1.5009, 1.7853], + device='cuda:4'), covar=tensor([0.1024, 0.1747, 0.1135, 0.1010, 0.1511, 0.0209, 0.1485, 0.1719], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0073, 0.0078, 0.0092, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 18:49:42,103 INFO [finetune.py:976] (4/7) Epoch 15, batch 5050, loss[loss=0.175, simple_loss=0.2443, pruned_loss=0.05282, over 4819.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2539, pruned_loss=0.05874, over 954285.42 frames. ], batch size: 25, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:49:46,817 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.063e+02 1.593e+02 1.872e+02 2.269e+02 5.264e+02, threshold=3.745e+02, percent-clipped=1.0 +2023-03-26 18:50:21,591 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8811, 1.3748, 1.9264, 1.8950, 1.7287, 1.6956, 1.8173, 1.8148], + device='cuda:4'), covar=tensor([0.4025, 0.4036, 0.3320, 0.3631, 0.4763, 0.3405, 0.4424, 0.3139], + device='cuda:4'), in_proj_covar=tensor([0.0245, 0.0238, 0.0258, 0.0269, 0.0268, 0.0241, 0.0280, 0.0237], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:50:22,633 INFO [finetune.py:976] (4/7) Epoch 15, batch 5100, loss[loss=0.1687, simple_loss=0.2382, pruned_loss=0.04958, over 4866.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.251, pruned_loss=0.0579, over 953518.85 frames. ], batch size: 31, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:50:23,331 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85289.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:50:41,090 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2103, 2.0785, 1.8050, 2.0309, 2.1572, 1.9130, 2.4914, 2.1920], + device='cuda:4'), covar=tensor([0.1424, 0.2011, 0.3146, 0.2546, 0.2648, 0.1831, 0.2552, 0.1771], + device='cuda:4'), in_proj_covar=tensor([0.0182, 0.0188, 0.0234, 0.0254, 0.0245, 0.0202, 0.0212, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:50:42,799 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85307.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:51:02,803 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85337.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:51:03,350 INFO [finetune.py:976] (4/7) Epoch 15, batch 5150, loss[loss=0.1891, simple_loss=0.2689, pruned_loss=0.05462, over 4788.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2504, pruned_loss=0.05779, over 952766.89 frames. ], batch size: 51, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:51:08,102 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.937e+01 1.526e+02 1.888e+02 2.256e+02 3.382e+02, threshold=3.776e+02, percent-clipped=0.0 +2023-03-26 18:51:29,584 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9425, 1.6545, 1.9787, 1.9290, 1.7156, 1.6749, 1.8877, 1.8927], + device='cuda:4'), covar=tensor([0.4118, 0.4472, 0.3396, 0.4408, 0.5241, 0.4315, 0.5259, 0.3422], + device='cuda:4'), in_proj_covar=tensor([0.0245, 0.0238, 0.0257, 0.0269, 0.0267, 0.0240, 0.0279, 0.0237], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:51:35,706 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 18:51:37,048 INFO [finetune.py:976] (4/7) Epoch 15, batch 5200, loss[loss=0.2029, simple_loss=0.2783, pruned_loss=0.06373, over 4809.00 frames. ], tot_loss[loss=0.1858, simple_loss=0.254, pruned_loss=0.0588, over 953428.65 frames. ], batch size: 45, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:52:04,161 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85428.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:52:05,476 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8677, 1.2064, 1.8845, 1.8501, 1.6662, 1.6007, 1.7858, 1.7786], + device='cuda:4'), covar=tensor([0.3880, 0.4194, 0.3237, 0.3509, 0.4741, 0.3679, 0.4107, 0.3047], + device='cuda:4'), in_proj_covar=tensor([0.0245, 0.0237, 0.0257, 0.0268, 0.0267, 0.0240, 0.0279, 0.0236], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:52:10,615 INFO [finetune.py:976] (4/7) Epoch 15, batch 5250, loss[loss=0.1419, simple_loss=0.2271, pruned_loss=0.02836, over 4754.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2543, pruned_loss=0.05812, over 953879.78 frames. ], batch size: 28, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:52:15,820 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.657e+02 1.928e+02 2.523e+02 8.274e+02, threshold=3.856e+02, percent-clipped=2.0 +2023-03-26 18:52:23,067 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85456.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:52:27,687 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85463.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 18:52:33,026 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85471.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:52:40,782 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85483.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:52:44,235 INFO [finetune.py:976] (4/7) Epoch 15, batch 5300, loss[loss=0.1875, simple_loss=0.2598, pruned_loss=0.05757, over 4713.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2561, pruned_loss=0.05914, over 954634.50 frames. ], batch size: 59, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:52:44,967 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85489.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:52:54,983 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85504.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:53:19,700 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85530.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:53:24,905 INFO [finetune.py:976] (4/7) Epoch 15, batch 5350, loss[loss=0.1749, simple_loss=0.2378, pruned_loss=0.05602, over 4718.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2556, pruned_loss=0.05873, over 954989.01 frames. ], batch size: 59, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:53:28,694 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85544.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:53:29,176 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.542e+02 1.806e+02 2.197e+02 4.190e+02, threshold=3.613e+02, percent-clipped=2.0 +2023-03-26 18:53:58,023 INFO [finetune.py:976] (4/7) Epoch 15, batch 5400, loss[loss=0.226, simple_loss=0.2791, pruned_loss=0.08647, over 4828.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2547, pruned_loss=0.05838, over 956550.12 frames. ], batch size: 33, lr: 3.47e-03, grad_scale: 32.0 +2023-03-26 18:54:00,452 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85591.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:54:11,212 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85607.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:54:12,998 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8682, 4.0932, 3.8791, 1.9432, 4.1825, 3.2318, 0.7101, 3.0697], + device='cuda:4'), covar=tensor([0.2237, 0.1700, 0.1597, 0.3344, 0.1005, 0.0974, 0.4809, 0.1278], + device='cuda:4'), in_proj_covar=tensor([0.0149, 0.0172, 0.0158, 0.0127, 0.0156, 0.0121, 0.0144, 0.0121], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 18:54:23,543 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85625.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:54:31,764 INFO [finetune.py:976] (4/7) Epoch 15, batch 5450, loss[loss=0.1873, simple_loss=0.2571, pruned_loss=0.0588, over 4945.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2514, pruned_loss=0.05744, over 957250.86 frames. ], batch size: 42, lr: 3.47e-03, grad_scale: 64.0 +2023-03-26 18:54:41,083 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.884e+01 1.516e+02 1.902e+02 2.390e+02 5.288e+02, threshold=3.804e+02, percent-clipped=4.0 +2023-03-26 18:54:52,062 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85655.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:55:16,884 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85686.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 18:55:17,974 INFO [finetune.py:976] (4/7) Epoch 15, batch 5500, loss[loss=0.1469, simple_loss=0.2126, pruned_loss=0.04058, over 4775.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2478, pruned_loss=0.05579, over 957093.03 frames. ], batch size: 26, lr: 3.47e-03, grad_scale: 64.0 +2023-03-26 18:56:02,559 INFO [finetune.py:976] (4/7) Epoch 15, batch 5550, loss[loss=0.2233, simple_loss=0.2922, pruned_loss=0.07715, over 4825.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2495, pruned_loss=0.05643, over 957516.68 frames. ], batch size: 39, lr: 3.47e-03, grad_scale: 64.0 +2023-03-26 18:56:06,722 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.623e+01 1.589e+02 1.875e+02 2.150e+02 4.153e+02, threshold=3.750e+02, percent-clipped=1.0 +2023-03-26 18:56:17,603 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8803, 1.6691, 2.1085, 1.3754, 1.8795, 2.1039, 1.6658, 2.3030], + device='cuda:4'), covar=tensor([0.1402, 0.2201, 0.1424, 0.2010, 0.1080, 0.1437, 0.2654, 0.0932], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0205, 0.0193, 0.0191, 0.0177, 0.0213, 0.0218, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:56:19,301 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85763.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:56:24,635 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85771.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:56:28,977 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 18:56:32,626 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85784.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:56:34,954 INFO [finetune.py:976] (4/7) Epoch 15, batch 5600, loss[loss=0.1764, simple_loss=0.2489, pruned_loss=0.05191, over 4795.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2524, pruned_loss=0.05747, over 955818.25 frames. ], batch size: 25, lr: 3.47e-03, grad_scale: 64.0 +2023-03-26 18:56:48,430 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85811.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:56:53,124 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=85819.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:56:55,445 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7842, 1.4618, 1.0404, 1.5796, 2.0476, 1.5726, 1.5916, 1.6252], + device='cuda:4'), covar=tensor([0.1965, 0.2601, 0.2146, 0.1689, 0.2256, 0.2416, 0.1884, 0.2900], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0110, 0.0093, 0.0118, 0.0094, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 18:57:04,149 INFO [finetune.py:976] (4/7) Epoch 15, batch 5650, loss[loss=0.1877, simple_loss=0.2597, pruned_loss=0.05782, over 4816.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2543, pruned_loss=0.05784, over 953986.25 frames. ], batch size: 51, lr: 3.47e-03, grad_scale: 64.0 +2023-03-26 18:57:04,772 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85839.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:57:08,246 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.126e+02 1.563e+02 1.888e+02 2.328e+02 3.522e+02, threshold=3.776e+02, percent-clipped=0.0 +2023-03-26 18:57:26,049 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85875.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:57:32,554 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85886.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:57:33,722 INFO [finetune.py:976] (4/7) Epoch 15, batch 5700, loss[loss=0.1259, simple_loss=0.1804, pruned_loss=0.03573, over 3840.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.251, pruned_loss=0.05737, over 938031.53 frames. ], batch size: 16, lr: 3.47e-03, grad_scale: 64.0 +2023-03-26 18:57:36,257 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3302, 1.7441, 2.2731, 2.2715, 2.0534, 2.0372, 2.1863, 2.2316], + device='cuda:4'), covar=tensor([0.3684, 0.3830, 0.3265, 0.3529, 0.4939, 0.3431, 0.4455, 0.2901], + device='cuda:4'), in_proj_covar=tensor([0.0244, 0.0236, 0.0256, 0.0267, 0.0266, 0.0239, 0.0278, 0.0235], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:58:02,782 INFO [finetune.py:976] (4/7) Epoch 16, batch 0, loss[loss=0.2388, simple_loss=0.3013, pruned_loss=0.08814, over 4732.00 frames. ], tot_loss[loss=0.2388, simple_loss=0.3013, pruned_loss=0.08814, over 4732.00 frames. ], batch size: 59, lr: 3.46e-03, grad_scale: 64.0 +2023-03-26 18:58:02,782 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 18:58:09,975 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8141, 1.3814, 0.9380, 1.6018, 2.1356, 1.1370, 1.6384, 1.6084], + device='cuda:4'), covar=tensor([0.1433, 0.1904, 0.1803, 0.1230, 0.1808, 0.2000, 0.1289, 0.1931], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0110, 0.0093, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 18:58:17,921 INFO [finetune.py:1010] (4/7) Epoch 16, validation: loss=0.1572, simple_loss=0.2278, pruned_loss=0.04329, over 2265189.00 frames. +2023-03-26 18:58:17,921 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 18:58:22,809 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5125, 1.5210, 1.2601, 1.4488, 1.9161, 1.8187, 1.4903, 1.3778], + device='cuda:4'), covar=tensor([0.0342, 0.0345, 0.0620, 0.0331, 0.0199, 0.0509, 0.0345, 0.0387], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0109, 0.0143, 0.0113, 0.0100, 0.0107, 0.0097, 0.0109], + device='cuda:4'), out_proj_covar=tensor([7.2658e-05, 8.4242e-05, 1.1341e-04, 8.7361e-05, 7.8285e-05, 7.8926e-05, + 7.3140e-05, 8.2796e-05], device='cuda:4') +2023-03-26 18:58:26,615 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 18:58:26,971 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85930.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 18:58:29,350 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8390, 1.6731, 1.5631, 1.6557, 1.1406, 4.2607, 1.6647, 2.0730], + device='cuda:4'), covar=tensor([0.3091, 0.2228, 0.2025, 0.2258, 0.1728, 0.0129, 0.2390, 0.1180], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0114, 0.0118, 0.0123, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 18:58:30,585 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85936.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:58:36,428 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.566e+02 1.783e+02 2.274e+02 8.459e+02, threshold=3.567e+02, percent-clipped=4.0 +2023-03-26 18:58:42,507 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7326, 0.9868, 1.7823, 1.7065, 1.5482, 1.5319, 1.6233, 1.6807], + device='cuda:4'), covar=tensor([0.3704, 0.4070, 0.3384, 0.3573, 0.4626, 0.3502, 0.4511, 0.3239], + device='cuda:4'), in_proj_covar=tensor([0.0244, 0.0237, 0.0256, 0.0267, 0.0266, 0.0240, 0.0279, 0.0236], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 18:58:44,310 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85957.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:58:49,655 INFO [finetune.py:976] (4/7) Epoch 16, batch 50, loss[loss=0.1626, simple_loss=0.2288, pruned_loss=0.04819, over 4904.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2578, pruned_loss=0.05878, over 217832.83 frames. ], batch size: 36, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 18:58:59,780 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85981.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 18:59:05,879 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85991.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 18:59:14,742 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86002.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:59:23,677 INFO [finetune.py:976] (4/7) Epoch 16, batch 100, loss[loss=0.1344, simple_loss=0.2149, pruned_loss=0.02701, over 4758.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2526, pruned_loss=0.05773, over 382103.84 frames. ], batch size: 27, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 18:59:25,007 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86018.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 18:59:43,488 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-26 18:59:44,131 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.132e+02 1.611e+02 1.877e+02 2.147e+02 3.763e+02, threshold=3.754e+02, percent-clipped=3.0 +2023-03-26 18:59:47,903 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86052.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:00:00,138 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86063.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:00:06,631 INFO [finetune.py:976] (4/7) Epoch 16, batch 150, loss[loss=0.1544, simple_loss=0.2116, pruned_loss=0.04855, over 4814.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2491, pruned_loss=0.0572, over 510095.80 frames. ], batch size: 25, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:00:07,432 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.84 vs. limit=5.0 +2023-03-26 19:00:08,561 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86069.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:00:27,243 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86084.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:00:50,262 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86113.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:00:51,949 INFO [finetune.py:976] (4/7) Epoch 16, batch 200, loss[loss=0.2231, simple_loss=0.2649, pruned_loss=0.09065, over 4209.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2458, pruned_loss=0.05704, over 606674.19 frames. ], batch size: 65, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:01:04,024 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86130.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:01:05,146 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86132.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:01:09,828 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86139.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:01:14,011 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.089e+02 1.565e+02 1.801e+02 2.285e+02 3.660e+02, threshold=3.601e+02, percent-clipped=0.0 +2023-03-26 19:01:27,190 INFO [finetune.py:976] (4/7) Epoch 16, batch 250, loss[loss=0.1602, simple_loss=0.2335, pruned_loss=0.04351, over 4754.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2472, pruned_loss=0.05698, over 684011.57 frames. ], batch size: 27, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:01:39,657 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4266, 1.2713, 1.2527, 1.3213, 1.7127, 1.5270, 1.3262, 1.2508], + device='cuda:4'), covar=tensor([0.0286, 0.0292, 0.0610, 0.0273, 0.0197, 0.0406, 0.0316, 0.0335], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0108, 0.0143, 0.0112, 0.0100, 0.0107, 0.0097, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.2070e-05, 8.3590e-05, 1.1292e-04, 8.6837e-05, 7.7860e-05, 7.8624e-05, + 7.3071e-05, 8.2667e-05], device='cuda:4') +2023-03-26 19:01:40,838 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86186.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:01:41,402 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86187.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:02:00,656 INFO [finetune.py:976] (4/7) Epoch 16, batch 300, loss[loss=0.1885, simple_loss=0.2452, pruned_loss=0.06585, over 4810.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2527, pruned_loss=0.05967, over 744071.59 frames. ], batch size: 40, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:02:04,022 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 19:02:11,160 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86231.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:02:12,976 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86234.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:02:13,068 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2102, 2.2029, 1.8009, 2.2979, 2.1028, 2.1520, 2.1139, 3.0977], + device='cuda:4'), covar=tensor([0.3902, 0.4630, 0.3608, 0.4563, 0.4479, 0.2397, 0.4654, 0.1562], + device='cuda:4'), in_proj_covar=tensor([0.0283, 0.0259, 0.0224, 0.0273, 0.0246, 0.0213, 0.0247, 0.0225], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:02:20,698 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.190e+02 1.625e+02 1.967e+02 2.251e+02 5.649e+02, threshold=3.935e+02, percent-clipped=3.0 +2023-03-26 19:02:20,816 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86246.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:02:25,058 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8041, 1.5901, 1.7458, 1.3365, 1.7543, 1.8181, 1.7934, 1.4204], + device='cuda:4'), covar=tensor([0.0543, 0.0667, 0.0699, 0.0852, 0.0899, 0.0646, 0.0600, 0.1110], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0134, 0.0140, 0.0122, 0.0122, 0.0139, 0.0140, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:02:31,541 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.40 vs. limit=5.0 +2023-03-26 19:02:34,045 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-26 19:02:34,279 INFO [finetune.py:976] (4/7) Epoch 16, batch 350, loss[loss=0.1856, simple_loss=0.2547, pruned_loss=0.05826, over 4859.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2543, pruned_loss=0.05982, over 791306.11 frames. ], batch size: 44, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:02:44,512 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86281.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:02:47,960 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86286.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:03:01,135 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86307.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:03:05,210 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86313.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:03:07,486 INFO [finetune.py:976] (4/7) Epoch 16, batch 400, loss[loss=0.1872, simple_loss=0.2642, pruned_loss=0.05509, over 4736.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2548, pruned_loss=0.05891, over 828983.23 frames. ], batch size: 27, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:03:15,903 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86329.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:03:34,377 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.307e+01 1.538e+02 1.774e+02 2.172e+02 4.200e+02, threshold=3.548e+02, percent-clipped=1.0 +2023-03-26 19:03:45,095 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86358.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:03:50,332 INFO [finetune.py:976] (4/7) Epoch 16, batch 450, loss[loss=0.1882, simple_loss=0.2482, pruned_loss=0.06409, over 4825.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2531, pruned_loss=0.05791, over 857384.93 frames. ], batch size: 38, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:03:56,821 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7797, 1.6632, 1.4998, 1.8123, 2.0938, 1.8482, 1.4480, 1.4834], + device='cuda:4'), covar=tensor([0.2007, 0.1870, 0.1781, 0.1506, 0.1554, 0.1116, 0.2380, 0.1772], + device='cuda:4'), in_proj_covar=tensor([0.0239, 0.0207, 0.0210, 0.0190, 0.0241, 0.0184, 0.0215, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:04:12,247 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9071, 1.8251, 1.6774, 2.0089, 2.4025, 1.9689, 1.7931, 1.5884], + device='cuda:4'), covar=tensor([0.2084, 0.2022, 0.1836, 0.1571, 0.1727, 0.1175, 0.2320, 0.1859], + device='cuda:4'), in_proj_covar=tensor([0.0240, 0.0208, 0.0211, 0.0191, 0.0242, 0.0184, 0.0215, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:04:18,722 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86408.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:04:24,032 INFO [finetune.py:976] (4/7) Epoch 16, batch 500, loss[loss=0.2038, simple_loss=0.2668, pruned_loss=0.07036, over 4902.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2509, pruned_loss=0.0574, over 879201.22 frames. ], batch size: 35, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:04:30,024 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86425.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:04:38,330 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3238, 2.1746, 1.7588, 2.1511, 2.1824, 1.9267, 2.4530, 2.2924], + device='cuda:4'), covar=tensor([0.1373, 0.2015, 0.3044, 0.2536, 0.2537, 0.1719, 0.2742, 0.1878], + device='cuda:4'), in_proj_covar=tensor([0.0182, 0.0189, 0.0236, 0.0256, 0.0246, 0.0203, 0.0214, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:04:40,633 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86440.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:04:43,600 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86445.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:04:44,070 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.779e+01 1.556e+02 1.875e+02 2.226e+02 4.465e+02, threshold=3.750e+02, percent-clipped=2.0 +2023-03-26 19:04:51,675 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 19:04:57,164 INFO [finetune.py:976] (4/7) Epoch 16, batch 550, loss[loss=0.1918, simple_loss=0.2543, pruned_loss=0.06464, over 4444.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2477, pruned_loss=0.05568, over 897075.29 frames. ], batch size: 19, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:05:12,086 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 19:05:13,280 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 19:05:31,553 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86501.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:05:39,351 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-26 19:05:39,628 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86506.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:05:50,206 INFO [finetune.py:976] (4/7) Epoch 16, batch 600, loss[loss=0.2005, simple_loss=0.2704, pruned_loss=0.06534, over 4834.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2485, pruned_loss=0.05619, over 911331.64 frames. ], batch size: 33, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:06:01,804 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6127, 1.5879, 1.5260, 1.0080, 1.6519, 1.8702, 1.7937, 1.3854], + device='cuda:4'), covar=tensor([0.0981, 0.0683, 0.0555, 0.0580, 0.0477, 0.0670, 0.0361, 0.0760], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0153, 0.0124, 0.0129, 0.0131, 0.0128, 0.0143, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.2917e-05, 1.1111e-04, 8.9260e-05, 9.2568e-05, 9.2690e-05, 9.2248e-05, + 1.0293e-04, 1.0658e-04], device='cuda:4') +2023-03-26 19:06:04,185 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86531.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:06:09,401 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3780, 2.2227, 2.3864, 1.7106, 2.3278, 2.5299, 2.4911, 1.9627], + device='cuda:4'), covar=tensor([0.0539, 0.0602, 0.0616, 0.0864, 0.0619, 0.0630, 0.0544, 0.1048], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0134, 0.0140, 0.0122, 0.0123, 0.0140, 0.0141, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:06:14,604 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.101e+02 1.575e+02 1.922e+02 2.222e+02 3.111e+02, threshold=3.844e+02, percent-clipped=0.0 +2023-03-26 19:06:27,336 INFO [finetune.py:976] (4/7) Epoch 16, batch 650, loss[loss=0.1973, simple_loss=0.2805, pruned_loss=0.05708, over 4816.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.2505, pruned_loss=0.05655, over 922003.80 frames. ], batch size: 51, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:06:36,232 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86579.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:06:37,402 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86580.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:06:41,057 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86586.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 19:06:43,452 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3535, 1.2828, 1.6049, 1.1776, 1.3175, 1.4671, 1.2977, 1.6567], + device='cuda:4'), covar=tensor([0.1168, 0.2067, 0.1181, 0.1388, 0.0900, 0.1134, 0.2715, 0.0758], + device='cuda:4'), in_proj_covar=tensor([0.0195, 0.0207, 0.0194, 0.0192, 0.0179, 0.0216, 0.0220, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:06:52,272 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86602.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:06:59,429 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86613.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:07:01,142 INFO [finetune.py:976] (4/7) Epoch 16, batch 700, loss[loss=0.1971, simple_loss=0.269, pruned_loss=0.0626, over 4872.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2534, pruned_loss=0.05805, over 927758.15 frames. ], batch size: 31, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:07:13,496 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86634.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:07:17,803 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86641.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:07:18,429 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8660, 1.6297, 1.5011, 1.2643, 1.6231, 1.7041, 1.6415, 2.1872], + device='cuda:4'), covar=tensor([0.3583, 0.4364, 0.3295, 0.3724, 0.3685, 0.2319, 0.3656, 0.1793], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0260, 0.0225, 0.0274, 0.0248, 0.0214, 0.0249, 0.0226], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:07:21,644 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.513e+02 1.866e+02 2.326e+02 3.823e+02, threshold=3.732e+02, percent-clipped=0.0 +2023-03-26 19:07:29,554 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86658.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:07:31,332 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86661.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:07:34,799 INFO [finetune.py:976] (4/7) Epoch 16, batch 750, loss[loss=0.1623, simple_loss=0.2515, pruned_loss=0.03652, over 4841.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2552, pruned_loss=0.05881, over 932874.94 frames. ], batch size: 49, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:08:02,109 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86706.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:08:03,364 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86708.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:08:07,755 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-03-26 19:08:08,108 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.1523, 2.8245, 2.5587, 1.3519, 2.7278, 2.2294, 2.2288, 2.3921], + device='cuda:4'), covar=tensor([0.1054, 0.0881, 0.1703, 0.2229, 0.1849, 0.2219, 0.2060, 0.1268], + device='cuda:4'), in_proj_covar=tensor([0.0167, 0.0195, 0.0199, 0.0182, 0.0211, 0.0206, 0.0222, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:08:08,617 INFO [finetune.py:976] (4/7) Epoch 16, batch 800, loss[loss=0.1554, simple_loss=0.2287, pruned_loss=0.04107, over 4823.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2544, pruned_loss=0.05791, over 938472.24 frames. ], batch size: 38, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:08:14,167 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86725.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:08:26,128 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 19:08:28,782 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.501e+02 1.840e+02 2.208e+02 4.378e+02, threshold=3.681e+02, percent-clipped=4.0 +2023-03-26 19:08:40,240 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86756.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:08:49,459 INFO [finetune.py:976] (4/7) Epoch 16, batch 850, loss[loss=0.2174, simple_loss=0.2916, pruned_loss=0.07157, over 4704.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2527, pruned_loss=0.05794, over 942248.47 frames. ], batch size: 59, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:08:54,218 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86773.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:09:09,589 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86796.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:09:13,592 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86801.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:09:21,907 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86814.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:09:23,014 INFO [finetune.py:976] (4/7) Epoch 16, batch 900, loss[loss=0.1821, simple_loss=0.2451, pruned_loss=0.05953, over 4843.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2492, pruned_loss=0.0569, over 943871.31 frames. ], batch size: 49, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:09:43,094 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.048e+02 1.502e+02 1.904e+02 2.205e+02 3.944e+02, threshold=3.808e+02, percent-clipped=2.0 +2023-03-26 19:09:56,619 INFO [finetune.py:976] (4/7) Epoch 16, batch 950, loss[loss=0.1892, simple_loss=0.2687, pruned_loss=0.05487, over 4817.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2479, pruned_loss=0.0567, over 945939.34 frames. ], batch size: 41, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:10:02,706 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86875.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:10:04,478 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86878.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:10:20,383 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86902.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:10:40,046 INFO [finetune.py:976] (4/7) Epoch 16, batch 1000, loss[loss=0.1255, simple_loss=0.194, pruned_loss=0.02844, over 4734.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.248, pruned_loss=0.05623, over 946218.56 frames. ], batch size: 23, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:11:01,992 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86936.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:11:08,315 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86939.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:11:12,979 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.150e+02 1.633e+02 1.920e+02 2.320e+02 4.350e+02, threshold=3.840e+02, percent-clipped=2.0 +2023-03-26 19:11:13,133 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1291, 2.0322, 1.7094, 1.6705, 2.5851, 2.5132, 2.0785, 2.0100], + device='cuda:4'), covar=tensor([0.0329, 0.0360, 0.0556, 0.0374, 0.0212, 0.0566, 0.0327, 0.0419], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0108, 0.0143, 0.0113, 0.0099, 0.0106, 0.0097, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.2572e-05, 8.3815e-05, 1.1299e-04, 8.7070e-05, 7.7397e-05, 7.8312e-05, + 7.3141e-05, 8.2343e-05], device='cuda:4') +2023-03-26 19:11:20,312 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=86950.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:11:34,825 INFO [finetune.py:976] (4/7) Epoch 16, batch 1050, loss[loss=0.1848, simple_loss=0.2644, pruned_loss=0.05257, over 4929.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2509, pruned_loss=0.0568, over 948212.72 frames. ], batch size: 33, lr: 3.46e-03, grad_scale: 32.0 +2023-03-26 19:11:58,520 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87002.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:12:05,582 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-26 19:12:08,331 INFO [finetune.py:976] (4/7) Epoch 16, batch 1100, loss[loss=0.2331, simple_loss=0.2974, pruned_loss=0.0844, over 4898.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2523, pruned_loss=0.057, over 951753.09 frames. ], batch size: 43, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:12:27,409 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.131e+02 1.613e+02 1.835e+02 2.273e+02 4.124e+02, threshold=3.670e+02, percent-clipped=1.0 +2023-03-26 19:12:31,623 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7734, 1.8359, 1.5530, 1.9359, 2.4114, 2.0178, 1.6472, 1.4472], + device='cuda:4'), covar=tensor([0.2188, 0.1905, 0.1873, 0.1557, 0.1587, 0.1129, 0.2268, 0.1984], + device='cuda:4'), in_proj_covar=tensor([0.0240, 0.0208, 0.0211, 0.0191, 0.0242, 0.0185, 0.0214, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:12:39,656 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87063.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:12:41,762 INFO [finetune.py:976] (4/7) Epoch 16, batch 1150, loss[loss=0.1717, simple_loss=0.2411, pruned_loss=0.0511, over 4896.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2538, pruned_loss=0.05774, over 950556.24 frames. ], batch size: 46, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:13:01,426 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87096.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:13:04,425 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87101.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:13:15,214 INFO [finetune.py:976] (4/7) Epoch 16, batch 1200, loss[loss=0.1986, simple_loss=0.2587, pruned_loss=0.06926, over 4912.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2536, pruned_loss=0.05796, over 952304.66 frames. ], batch size: 37, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:13:16,019 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 19:13:31,480 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9433, 1.7778, 1.4682, 1.4974, 1.7065, 1.7046, 1.7352, 2.3491], + device='cuda:4'), covar=tensor([0.3795, 0.3923, 0.3340, 0.3742, 0.3879, 0.2378, 0.3745, 0.1963], + device='cuda:4'), in_proj_covar=tensor([0.0283, 0.0259, 0.0224, 0.0273, 0.0246, 0.0213, 0.0248, 0.0226], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:13:33,239 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87144.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:13:34,371 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.093e+01 1.544e+02 1.890e+02 2.251e+02 4.242e+02, threshold=3.781e+02, percent-clipped=1.0 +2023-03-26 19:13:36,664 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87149.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:13:50,132 INFO [finetune.py:976] (4/7) Epoch 16, batch 1250, loss[loss=0.168, simple_loss=0.2326, pruned_loss=0.05173, over 4932.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2511, pruned_loss=0.05694, over 953196.57 frames. ], batch size: 33, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:13:53,124 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87170.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:14:24,755 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.2206, 1.3647, 1.4042, 0.7402, 1.3921, 1.5692, 1.5890, 1.3250], + device='cuda:4'), covar=tensor([0.0845, 0.0556, 0.0488, 0.0515, 0.0454, 0.0565, 0.0342, 0.0601], + device='cuda:4'), in_proj_covar=tensor([0.0125, 0.0152, 0.0124, 0.0128, 0.0131, 0.0128, 0.0143, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.2325e-05, 1.1052e-04, 8.9083e-05, 9.1680e-05, 9.2526e-05, 9.2351e-05, + 1.0284e-04, 1.0619e-04], device='cuda:4') +2023-03-26 19:14:31,066 INFO [finetune.py:976] (4/7) Epoch 16, batch 1300, loss[loss=0.2066, simple_loss=0.2685, pruned_loss=0.07239, over 4747.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2475, pruned_loss=0.05576, over 954009.14 frames. ], batch size: 26, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:14:41,193 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87230.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:14:43,587 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87234.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:14:44,802 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87236.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:14:51,269 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.006e+02 1.578e+02 1.944e+02 2.250e+02 4.130e+02, threshold=3.887e+02, percent-clipped=1.0 +2023-03-26 19:15:04,398 INFO [finetune.py:976] (4/7) Epoch 16, batch 1350, loss[loss=0.1781, simple_loss=0.2589, pruned_loss=0.04863, over 4912.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2487, pruned_loss=0.05701, over 953697.14 frames. ], batch size: 38, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:15:16,952 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87284.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:15:21,850 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87291.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:15:41,646 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.46 vs. limit=5.0 +2023-03-26 19:15:44,164 INFO [finetune.py:976] (4/7) Epoch 16, batch 1400, loss[loss=0.2524, simple_loss=0.3115, pruned_loss=0.0966, over 4809.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2537, pruned_loss=0.0585, over 953361.92 frames. ], batch size: 39, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:16:19,255 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.049e+02 1.619e+02 1.873e+02 2.376e+02 3.982e+02, threshold=3.745e+02, percent-clipped=1.0 +2023-03-26 19:16:31,530 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87358.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:16:38,511 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87362.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:16:41,376 INFO [finetune.py:976] (4/7) Epoch 16, batch 1450, loss[loss=0.1901, simple_loss=0.2551, pruned_loss=0.06253, over 4891.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2556, pruned_loss=0.05912, over 955459.94 frames. ], batch size: 35, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:16:42,932 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-26 19:17:17,384 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8017, 1.4537, 1.8933, 1.9853, 1.5840, 3.4125, 1.3954, 1.5572], + device='cuda:4'), covar=tensor([0.0842, 0.1806, 0.1264, 0.0851, 0.1574, 0.0210, 0.1447, 0.1810], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0082, 0.0075, 0.0078, 0.0093, 0.0081, 0.0086, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 19:17:18,468 INFO [finetune.py:976] (4/7) Epoch 16, batch 1500, loss[loss=0.1972, simple_loss=0.2589, pruned_loss=0.06773, over 4830.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2568, pruned_loss=0.05925, over 956645.00 frames. ], batch size: 47, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:17:22,338 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.89 vs. limit=5.0 +2023-03-26 19:17:23,389 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87423.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:17:39,130 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.635e+02 2.033e+02 2.421e+02 4.092e+02, threshold=4.066e+02, percent-clipped=1.0 +2023-03-26 19:17:48,791 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87461.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:17:52,196 INFO [finetune.py:976] (4/7) Epoch 16, batch 1550, loss[loss=0.201, simple_loss=0.2661, pruned_loss=0.06795, over 4778.00 frames. ], tot_loss[loss=0.1866, simple_loss=0.2557, pruned_loss=0.0588, over 955942.96 frames. ], batch size: 51, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:17:55,188 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87470.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:17:58,868 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3075, 1.4201, 1.5093, 0.8128, 1.5493, 1.7505, 1.7739, 1.4283], + device='cuda:4'), covar=tensor([0.1041, 0.0713, 0.0500, 0.0642, 0.0477, 0.0632, 0.0313, 0.0851], + device='cuda:4'), in_proj_covar=tensor([0.0125, 0.0151, 0.0124, 0.0128, 0.0131, 0.0127, 0.0142, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.2254e-05, 1.1001e-04, 8.8976e-05, 9.1606e-05, 9.2567e-05, 9.1573e-05, + 1.0259e-04, 1.0622e-04], device='cuda:4') +2023-03-26 19:18:06,907 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 19:18:09,219 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3524, 1.4071, 1.4324, 1.5685, 1.6416, 2.9059, 1.3969, 1.5609], + device='cuda:4'), covar=tensor([0.0970, 0.1749, 0.1135, 0.0950, 0.1481, 0.0277, 0.1368, 0.1638], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0074, 0.0078, 0.0092, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 19:18:18,816 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87505.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:18:25,485 INFO [finetune.py:976] (4/7) Epoch 16, batch 1600, loss[loss=0.1675, simple_loss=0.2322, pruned_loss=0.05135, over 4805.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2537, pruned_loss=0.058, over 957460.63 frames. ], batch size: 25, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:18:27,238 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87518.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:18:30,228 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87522.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:18:33,950 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7924, 0.6551, 1.7474, 1.6284, 1.5651, 1.5001, 1.5719, 1.7230], + device='cuda:4'), covar=tensor([0.3302, 0.3756, 0.3277, 0.3461, 0.4376, 0.3688, 0.4175, 0.3037], + device='cuda:4'), in_proj_covar=tensor([0.0247, 0.0239, 0.0257, 0.0270, 0.0269, 0.0241, 0.0281, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:18:38,563 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87534.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:18:46,617 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.707e+01 1.405e+02 1.628e+02 1.991e+02 3.372e+02, threshold=3.256e+02, percent-clipped=0.0 +2023-03-26 19:18:57,037 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0812, 2.0782, 2.2010, 1.7504, 2.0783, 2.6396, 2.4185, 1.6945], + device='cuda:4'), covar=tensor([0.0654, 0.0690, 0.0719, 0.1016, 0.1496, 0.0523, 0.0571, 0.1413], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0134, 0.0141, 0.0123, 0.0123, 0.0140, 0.0142, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:18:59,343 INFO [finetune.py:976] (4/7) Epoch 16, batch 1650, loss[loss=0.1859, simple_loss=0.2528, pruned_loss=0.05946, over 4726.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2502, pruned_loss=0.0564, over 956352.92 frames. ], batch size: 59, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:18:59,474 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87566.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:19:10,145 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87582.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:19:11,438 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87584.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:19:13,569 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87586.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:19:17,442 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87589.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:19:36,471 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87605.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:19:43,482 INFO [finetune.py:976] (4/7) Epoch 16, batch 1700, loss[loss=0.1995, simple_loss=0.2823, pruned_loss=0.05833, over 4852.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2482, pruned_loss=0.05577, over 956039.41 frames. ], batch size: 44, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:20:03,757 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87645.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:20:04,232 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.164e+02 1.606e+02 1.879e+02 2.372e+02 9.403e+02, threshold=3.758e+02, percent-clipped=6.0 +2023-03-26 19:20:06,847 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87650.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:20:08,009 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5063, 3.4527, 3.2524, 1.6547, 3.6010, 2.7025, 0.8259, 2.4182], + device='cuda:4'), covar=tensor([0.2474, 0.2749, 0.1966, 0.3558, 0.1387, 0.1028, 0.4447, 0.1593], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0173, 0.0158, 0.0127, 0.0156, 0.0122, 0.0145, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 19:20:11,626 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87658.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:20:16,843 INFO [finetune.py:976] (4/7) Epoch 16, batch 1750, loss[loss=0.1669, simple_loss=0.2374, pruned_loss=0.04818, over 4897.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.251, pruned_loss=0.05704, over 954685.66 frames. ], batch size: 35, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:20:16,972 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87666.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 19:20:44,154 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87706.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:20:50,641 INFO [finetune.py:976] (4/7) Epoch 16, batch 1800, loss[loss=0.2041, simple_loss=0.278, pruned_loss=0.06507, over 4841.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2542, pruned_loss=0.05815, over 955490.68 frames. ], batch size: 47, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:20:51,920 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87718.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:21:13,300 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.163e+02 1.635e+02 1.907e+02 2.399e+02 5.758e+02, threshold=3.813e+02, percent-clipped=1.0 +2023-03-26 19:21:36,192 INFO [finetune.py:976] (4/7) Epoch 16, batch 1850, loss[loss=0.1752, simple_loss=0.2517, pruned_loss=0.04931, over 4785.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2567, pruned_loss=0.05938, over 956408.89 frames. ], batch size: 51, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:21:39,396 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.47 vs. limit=5.0 +2023-03-26 19:22:22,285 INFO [finetune.py:976] (4/7) Epoch 16, batch 1900, loss[loss=0.1797, simple_loss=0.2521, pruned_loss=0.05362, over 4905.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.2573, pruned_loss=0.05887, over 957408.01 frames. ], batch size: 37, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:22:22,409 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8481, 1.7830, 1.6903, 1.7982, 1.3112, 3.2567, 1.5233, 1.9853], + device='cuda:4'), covar=tensor([0.2968, 0.2105, 0.1912, 0.2114, 0.1640, 0.0245, 0.2279, 0.1075], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0116, 0.0121, 0.0125, 0.0115, 0.0097, 0.0097, 0.0097], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 19:22:23,005 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87817.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:22:41,851 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.250e+02 1.530e+02 1.792e+02 2.215e+02 4.706e+02, threshold=3.584e+02, percent-clipped=3.0 +2023-03-26 19:22:53,010 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87861.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 19:22:55,975 INFO [finetune.py:976] (4/7) Epoch 16, batch 1950, loss[loss=0.1444, simple_loss=0.2088, pruned_loss=0.04002, over 3947.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2543, pruned_loss=0.05779, over 957591.17 frames. ], batch size: 17, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:23:09,168 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87886.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:23:29,595 INFO [finetune.py:976] (4/7) Epoch 16, batch 2000, loss[loss=0.1781, simple_loss=0.252, pruned_loss=0.05209, over 4851.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2514, pruned_loss=0.05679, over 956774.79 frames. ], batch size: 49, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:23:41,100 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=87934.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:23:44,687 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8093, 1.3777, 0.8553, 1.7407, 2.1802, 1.3196, 1.6371, 1.6983], + device='cuda:4'), covar=tensor([0.1410, 0.1944, 0.1943, 0.1114, 0.1791, 0.1892, 0.1322, 0.1905], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0092, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 19:23:45,287 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87940.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 19:23:48,826 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87945.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:23:49,306 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.562e+02 1.812e+02 2.199e+02 5.123e+02, threshold=3.624e+02, percent-clipped=1.0 +2023-03-26 19:23:59,898 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87961.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:24:02,865 INFO [finetune.py:976] (4/7) Epoch 16, batch 2050, loss[loss=0.1475, simple_loss=0.2226, pruned_loss=0.0362, over 4028.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2485, pruned_loss=0.05595, over 957601.94 frames. ], batch size: 17, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:24:32,041 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1901, 1.9894, 2.2697, 1.4434, 2.2092, 2.3018, 2.1677, 1.7943], + device='cuda:4'), covar=tensor([0.0558, 0.0690, 0.0621, 0.0891, 0.0608, 0.0687, 0.0601, 0.1179], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0134, 0.0141, 0.0123, 0.0124, 0.0139, 0.0142, 0.0164], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:24:37,647 INFO [finetune.py:976] (4/7) Epoch 16, batch 2100, loss[loss=0.1883, simple_loss=0.2658, pruned_loss=0.05534, over 4902.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2485, pruned_loss=0.0561, over 958114.15 frames. ], batch size: 35, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:24:41,372 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88018.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:24:46,849 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6276, 1.2571, 0.8442, 1.5752, 2.0179, 1.0390, 1.4439, 1.5004], + device='cuda:4'), covar=tensor([0.1558, 0.2112, 0.1830, 0.1201, 0.1880, 0.1983, 0.1406, 0.1935], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0092, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 19:24:48,084 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0505, 0.9769, 1.0027, 0.3657, 0.8387, 1.1415, 1.1658, 0.9575], + device='cuda:4'), covar=tensor([0.0890, 0.0566, 0.0538, 0.0531, 0.0548, 0.0563, 0.0379, 0.0680], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0152, 0.0125, 0.0129, 0.0132, 0.0128, 0.0143, 0.0148], + device='cuda:4'), out_proj_covar=tensor([9.2898e-05, 1.1056e-04, 8.9561e-05, 9.2050e-05, 9.3606e-05, 9.2169e-05, + 1.0341e-04, 1.0668e-04], device='cuda:4') +2023-03-26 19:24:59,880 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.339e+01 1.671e+02 1.963e+02 2.403e+02 4.597e+02, threshold=3.926e+02, percent-clipped=2.0 +2023-03-26 19:25:13,436 INFO [finetune.py:976] (4/7) Epoch 16, batch 2150, loss[loss=0.2144, simple_loss=0.2806, pruned_loss=0.07411, over 4821.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2531, pruned_loss=0.0578, over 957390.56 frames. ], batch size: 33, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:25:13,503 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88066.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:25:35,476 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=88100.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:25:46,508 INFO [finetune.py:976] (4/7) Epoch 16, batch 2200, loss[loss=0.2044, simple_loss=0.2676, pruned_loss=0.07059, over 4294.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2554, pruned_loss=0.0585, over 955772.48 frames. ], batch size: 66, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:25:47,210 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88117.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:26:01,678 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6535, 1.5273, 1.0825, 0.2945, 1.3427, 1.4590, 1.4201, 1.5065], + device='cuda:4'), covar=tensor([0.0990, 0.0810, 0.1492, 0.2033, 0.1362, 0.2382, 0.2388, 0.0860], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0196, 0.0201, 0.0183, 0.0212, 0.0208, 0.0224, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:26:05,787 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.658e+02 1.940e+02 2.471e+02 6.986e+02, threshold=3.880e+02, percent-clipped=3.0 +2023-03-26 19:26:15,409 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88161.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:26:15,433 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=88161.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:26:18,733 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88165.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:26:19,282 INFO [finetune.py:976] (4/7) Epoch 16, batch 2250, loss[loss=0.1935, simple_loss=0.261, pruned_loss=0.06296, over 4841.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2563, pruned_loss=0.05904, over 955746.12 frames. ], batch size: 49, lr: 3.45e-03, grad_scale: 32.0 +2023-03-26 19:26:31,474 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2333, 1.7428, 2.1912, 2.1606, 1.8547, 1.8507, 2.1300, 1.9951], + device='cuda:4'), covar=tensor([0.3667, 0.4122, 0.3195, 0.3759, 0.5156, 0.3734, 0.4551, 0.3142], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0236, 0.0256, 0.0268, 0.0268, 0.0241, 0.0279, 0.0236], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:26:56,476 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88209.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:27:05,822 INFO [finetune.py:976] (4/7) Epoch 16, batch 2300, loss[loss=0.1811, simple_loss=0.2506, pruned_loss=0.05582, over 4780.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2553, pruned_loss=0.05824, over 955732.89 frames. ], batch size: 26, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:27:29,743 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88240.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:27:33,298 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88245.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:27:33,919 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6622, 1.6328, 1.4097, 1.6224, 2.0522, 1.9111, 1.6445, 1.5146], + device='cuda:4'), covar=tensor([0.0336, 0.0298, 0.0613, 0.0287, 0.0198, 0.0456, 0.0356, 0.0373], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0108, 0.0143, 0.0112, 0.0100, 0.0106, 0.0097, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.2848e-05, 8.3566e-05, 1.1287e-04, 8.6731e-05, 7.7695e-05, 7.8389e-05, + 7.2657e-05, 8.2235e-05], device='cuda:4') +2023-03-26 19:27:34,391 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.462e+02 1.840e+02 2.103e+02 3.666e+02, threshold=3.679e+02, percent-clipped=0.0 +2023-03-26 19:27:43,883 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88261.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:27:47,285 INFO [finetune.py:976] (4/7) Epoch 16, batch 2350, loss[loss=0.2043, simple_loss=0.2734, pruned_loss=0.06756, over 4920.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2528, pruned_loss=0.05702, over 957840.98 frames. ], batch size: 37, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:27:50,965 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1067, 1.9959, 1.7339, 2.0467, 1.9052, 1.9254, 1.9450, 2.6691], + device='cuda:4'), covar=tensor([0.3727, 0.4443, 0.3227, 0.3946, 0.3966, 0.2410, 0.3929, 0.1659], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0260, 0.0225, 0.0274, 0.0247, 0.0214, 0.0249, 0.0226], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:28:01,663 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88288.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:28:03,566 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=88291.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:28:04,699 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88293.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:28:07,107 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2383, 2.1288, 1.7040, 2.1704, 2.2402, 1.8604, 2.5713, 2.2617], + device='cuda:4'), covar=tensor([0.1321, 0.2198, 0.3137, 0.2740, 0.2567, 0.1721, 0.3132, 0.1815], + device='cuda:4'), in_proj_covar=tensor([0.0182, 0.0187, 0.0234, 0.0253, 0.0244, 0.0201, 0.0211, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:28:15,423 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88309.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:28:20,093 INFO [finetune.py:976] (4/7) Epoch 16, batch 2400, loss[loss=0.1919, simple_loss=0.2543, pruned_loss=0.06476, over 4863.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2499, pruned_loss=0.05634, over 958262.05 frames. ], batch size: 44, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:28:40,399 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.006e+02 1.525e+02 1.766e+02 2.106e+02 3.774e+02, threshold=3.532e+02, percent-clipped=1.0 +2023-03-26 19:28:44,070 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=88352.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:28:52,864 INFO [finetune.py:976] (4/7) Epoch 16, batch 2450, loss[loss=0.2274, simple_loss=0.2732, pruned_loss=0.09077, over 4830.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2469, pruned_loss=0.05557, over 958112.96 frames. ], batch size: 30, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:29:26,834 INFO [finetune.py:976] (4/7) Epoch 16, batch 2500, loss[loss=0.224, simple_loss=0.3017, pruned_loss=0.07318, over 4858.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2492, pruned_loss=0.05644, over 958094.92 frames. ], batch size: 44, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:29:48,234 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.743e+02 2.000e+02 2.601e+02 5.270e+02, threshold=4.000e+02, percent-clipped=5.0 +2023-03-26 19:29:54,227 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=88456.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:30:00,709 INFO [finetune.py:976] (4/7) Epoch 16, batch 2550, loss[loss=0.1544, simple_loss=0.2278, pruned_loss=0.04051, over 4789.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2528, pruned_loss=0.05773, over 954519.72 frames. ], batch size: 26, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:30:17,737 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2668, 2.0047, 1.5116, 0.5920, 1.7599, 1.8522, 1.7056, 1.9235], + device='cuda:4'), covar=tensor([0.0864, 0.0755, 0.1393, 0.1908, 0.1231, 0.2144, 0.2062, 0.0785], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0196, 0.0199, 0.0183, 0.0212, 0.0207, 0.0223, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:30:33,907 INFO [finetune.py:976] (4/7) Epoch 16, batch 2600, loss[loss=0.1824, simple_loss=0.2549, pruned_loss=0.05492, over 4809.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2546, pruned_loss=0.05781, over 956542.94 frames. ], batch size: 38, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:30:34,882 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-26 19:30:55,467 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.436e+01 1.666e+02 1.940e+02 2.279e+02 3.712e+02, threshold=3.880e+02, percent-clipped=0.0 +2023-03-26 19:30:59,282 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1652, 2.0580, 1.6589, 2.0707, 2.0480, 1.7735, 2.3638, 2.1972], + device='cuda:4'), covar=tensor([0.1242, 0.2159, 0.2997, 0.2666, 0.2645, 0.1667, 0.3522, 0.1691], + device='cuda:4'), in_proj_covar=tensor([0.0182, 0.0187, 0.0234, 0.0252, 0.0244, 0.0201, 0.0212, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:31:07,484 INFO [finetune.py:976] (4/7) Epoch 16, batch 2650, loss[loss=0.2583, simple_loss=0.3118, pruned_loss=0.1024, over 4813.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2552, pruned_loss=0.05769, over 956527.64 frames. ], batch size: 40, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:31:12,341 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2474, 3.7003, 3.8628, 4.1064, 3.9943, 3.7575, 4.3246, 1.2563], + device='cuda:4'), covar=tensor([0.0653, 0.0833, 0.0820, 0.0940, 0.1114, 0.1482, 0.0642, 0.5532], + device='cuda:4'), in_proj_covar=tensor([0.0349, 0.0245, 0.0277, 0.0292, 0.0335, 0.0283, 0.0298, 0.0295], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:31:41,339 INFO [finetune.py:976] (4/7) Epoch 16, batch 2700, loss[loss=0.1889, simple_loss=0.2682, pruned_loss=0.05483, over 4809.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2545, pruned_loss=0.05737, over 955277.58 frames. ], batch size: 40, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:32:05,540 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.504e+02 1.815e+02 2.296e+02 4.078e+02, threshold=3.631e+02, percent-clipped=1.0 +2023-03-26 19:32:05,630 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=88647.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:32:26,939 INFO [finetune.py:976] (4/7) Epoch 16, batch 2750, loss[loss=0.1739, simple_loss=0.2421, pruned_loss=0.05283, over 4897.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2519, pruned_loss=0.05711, over 955409.19 frames. ], batch size: 35, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:32:42,724 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-03-26 19:32:51,377 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 19:33:17,052 INFO [finetune.py:976] (4/7) Epoch 16, batch 2800, loss[loss=0.1913, simple_loss=0.2644, pruned_loss=0.05912, over 4793.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2492, pruned_loss=0.0565, over 953999.42 frames. ], batch size: 26, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:33:24,317 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4587, 1.3876, 1.3328, 1.5013, 1.8024, 1.6249, 1.4503, 1.3568], + device='cuda:4'), covar=tensor([0.0336, 0.0298, 0.0568, 0.0271, 0.0179, 0.0486, 0.0333, 0.0351], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0108, 0.0143, 0.0112, 0.0099, 0.0106, 0.0097, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.2829e-05, 8.3457e-05, 1.1276e-04, 8.6469e-05, 7.7267e-05, 7.8471e-05, + 7.2883e-05, 8.2299e-05], device='cuda:4') +2023-03-26 19:33:37,857 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.207e+02 1.516e+02 1.815e+02 2.286e+02 3.246e+02, threshold=3.631e+02, percent-clipped=0.0 +2023-03-26 19:33:43,839 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-26 19:33:44,962 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88756.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:33:50,911 INFO [finetune.py:976] (4/7) Epoch 16, batch 2850, loss[loss=0.1701, simple_loss=0.2447, pruned_loss=0.0478, over 4787.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2492, pruned_loss=0.0571, over 951865.78 frames. ], batch size: 29, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:33:58,300 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.10 vs. limit=5.0 +2023-03-26 19:34:17,619 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88804.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:34:24,822 INFO [finetune.py:976] (4/7) Epoch 16, batch 2900, loss[loss=0.1594, simple_loss=0.2281, pruned_loss=0.04534, over 4787.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.2519, pruned_loss=0.05822, over 954236.87 frames. ], batch size: 25, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:34:45,211 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.025e+02 1.632e+02 1.968e+02 2.500e+02 4.348e+02, threshold=3.936e+02, percent-clipped=6.0 +2023-03-26 19:34:45,349 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5227, 1.5234, 1.2747, 1.5591, 1.8861, 1.7125, 1.5077, 1.4485], + device='cuda:4'), covar=tensor([0.0364, 0.0346, 0.0591, 0.0300, 0.0238, 0.0588, 0.0363, 0.0412], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0107, 0.0142, 0.0111, 0.0098, 0.0106, 0.0097, 0.0107], + device='cuda:4'), out_proj_covar=tensor([7.2315e-05, 8.3011e-05, 1.1238e-04, 8.5885e-05, 7.6785e-05, 7.8081e-05, + 7.2487e-05, 8.1845e-05], device='cuda:4') +2023-03-26 19:34:58,809 INFO [finetune.py:976] (4/7) Epoch 16, batch 2950, loss[loss=0.1997, simple_loss=0.2719, pruned_loss=0.0638, over 4811.00 frames. ], tot_loss[loss=0.1866, simple_loss=0.255, pruned_loss=0.05907, over 953725.14 frames. ], batch size: 40, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:35:04,891 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3765, 1.4068, 1.7039, 1.6284, 1.4846, 3.2586, 1.3005, 1.4889], + device='cuda:4'), covar=tensor([0.1026, 0.1841, 0.1137, 0.1049, 0.1694, 0.0241, 0.1532, 0.1791], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0073, 0.0077, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 19:35:32,641 INFO [finetune.py:976] (4/7) Epoch 16, batch 3000, loss[loss=0.2264, simple_loss=0.2825, pruned_loss=0.08516, over 4801.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2571, pruned_loss=0.05997, over 955218.79 frames. ], batch size: 51, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:35:32,642 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 19:35:38,753 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4440, 1.3444, 1.3296, 1.4632, 1.7418, 1.5631, 1.3703, 1.3093], + device='cuda:4'), covar=tensor([0.0341, 0.0312, 0.0562, 0.0289, 0.0228, 0.0444, 0.0341, 0.0355], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0108, 0.0143, 0.0112, 0.0099, 0.0107, 0.0097, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.2988e-05, 8.3872e-05, 1.1301e-04, 8.6599e-05, 7.7392e-05, 7.8828e-05, + 7.3023e-05, 8.2483e-05], device='cuda:4') +2023-03-26 19:35:41,773 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1329, 2.0335, 1.7858, 1.8952, 2.1549, 1.8970, 2.3091, 2.1177], + device='cuda:4'), covar=tensor([0.1401, 0.2259, 0.3232, 0.2653, 0.2620, 0.1779, 0.3190, 0.2059], + device='cuda:4'), in_proj_covar=tensor([0.0181, 0.0186, 0.0233, 0.0251, 0.0242, 0.0200, 0.0210, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:35:49,227 INFO [finetune.py:1010] (4/7) Epoch 16, validation: loss=0.1563, simple_loss=0.2263, pruned_loss=0.04316, over 2265189.00 frames. +2023-03-26 19:35:49,228 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 19:35:58,841 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9210, 1.6608, 1.5270, 1.3005, 1.6925, 1.6962, 1.6460, 2.2053], + device='cuda:4'), covar=tensor([0.3814, 0.4525, 0.3263, 0.3986, 0.3879, 0.2420, 0.3658, 0.1809], + device='cuda:4'), in_proj_covar=tensor([0.0288, 0.0262, 0.0227, 0.0276, 0.0249, 0.0217, 0.0251, 0.0229], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:36:10,238 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0146, 1.8663, 1.5602, 1.7826, 1.7436, 1.7286, 1.7717, 2.4251], + device='cuda:4'), covar=tensor([0.3650, 0.4209, 0.3269, 0.3703, 0.4134, 0.2332, 0.3973, 0.1668], + device='cuda:4'), in_proj_covar=tensor([0.0288, 0.0262, 0.0227, 0.0276, 0.0249, 0.0217, 0.0251, 0.0229], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:36:10,682 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.661e+02 1.990e+02 2.439e+02 3.546e+02, threshold=3.980e+02, percent-clipped=0.0 +2023-03-26 19:36:11,252 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88947.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:36:23,198 INFO [finetune.py:976] (4/7) Epoch 16, batch 3050, loss[loss=0.1978, simple_loss=0.2738, pruned_loss=0.06088, over 4790.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2571, pruned_loss=0.05914, over 954496.15 frames. ], batch size: 45, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:36:43,520 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=88995.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:36:44,774 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6964, 1.6309, 1.4130, 1.5811, 1.9892, 1.9834, 1.6312, 1.4914], + device='cuda:4'), covar=tensor([0.0393, 0.0437, 0.0665, 0.0396, 0.0228, 0.0595, 0.0416, 0.0501], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0108, 0.0143, 0.0112, 0.0099, 0.0107, 0.0097, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.3022e-05, 8.3670e-05, 1.1323e-04, 8.6609e-05, 7.7337e-05, 7.8762e-05, + 7.2858e-05, 8.2250e-05], device='cuda:4') +2023-03-26 19:36:57,478 INFO [finetune.py:976] (4/7) Epoch 16, batch 3100, loss[loss=0.195, simple_loss=0.2683, pruned_loss=0.06085, over 4903.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2545, pruned_loss=0.05785, over 955410.71 frames. ], batch size: 36, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:37:06,391 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89027.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:37:07,348 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 19:37:20,956 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.966e+01 1.506e+02 1.838e+02 2.198e+02 3.411e+02, threshold=3.676e+02, percent-clipped=0.0 +2023-03-26 19:37:33,684 INFO [finetune.py:976] (4/7) Epoch 16, batch 3150, loss[loss=0.1715, simple_loss=0.2352, pruned_loss=0.0539, over 4901.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.2525, pruned_loss=0.05797, over 957489.35 frames. ], batch size: 32, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:37:56,582 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89088.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:38:25,683 INFO [finetune.py:976] (4/7) Epoch 16, batch 3200, loss[loss=0.1864, simple_loss=0.2488, pruned_loss=0.06201, over 4821.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2488, pruned_loss=0.0563, over 956933.70 frames. ], batch size: 41, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:38:34,603 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-26 19:38:50,102 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.129e+02 1.609e+02 1.908e+02 2.339e+02 4.086e+02, threshold=3.816e+02, percent-clipped=1.0 +2023-03-26 19:38:50,249 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2706, 2.1464, 1.7220, 0.8468, 1.8798, 1.7699, 1.6650, 1.9890], + device='cuda:4'), covar=tensor([0.0825, 0.0659, 0.1396, 0.1831, 0.1336, 0.2144, 0.2042, 0.0829], + device='cuda:4'), in_proj_covar=tensor([0.0166, 0.0194, 0.0198, 0.0181, 0.0210, 0.0204, 0.0221, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:38:50,314 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 19:38:53,276 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.5175, 1.5008, 1.6055, 0.9488, 1.5847, 1.8548, 1.9217, 1.3957], + device='cuda:4'), covar=tensor([0.1039, 0.0721, 0.0482, 0.0597, 0.0453, 0.0545, 0.0319, 0.0735], + device='cuda:4'), in_proj_covar=tensor([0.0124, 0.0150, 0.0123, 0.0126, 0.0130, 0.0127, 0.0142, 0.0146], + device='cuda:4'), out_proj_covar=tensor([9.1430e-05, 1.0925e-04, 8.7801e-05, 8.9911e-05, 9.2157e-05, 9.1760e-05, + 1.0192e-04, 1.0514e-04], device='cuda:4') +2023-03-26 19:38:59,814 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5985, 1.4767, 1.4008, 1.6461, 1.5857, 1.6379, 1.0157, 1.3856], + device='cuda:4'), covar=tensor([0.1894, 0.1817, 0.1682, 0.1484, 0.1428, 0.1074, 0.2273, 0.1783], + device='cuda:4'), in_proj_covar=tensor([0.0242, 0.0209, 0.0212, 0.0193, 0.0245, 0.0186, 0.0215, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:39:02,082 INFO [finetune.py:976] (4/7) Epoch 16, batch 3250, loss[loss=0.2668, simple_loss=0.3174, pruned_loss=0.1081, over 4824.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.2498, pruned_loss=0.05693, over 954900.53 frames. ], batch size: 45, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:39:04,499 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89169.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:39:21,689 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3445, 1.5705, 1.0084, 2.1409, 2.4885, 2.0953, 1.9689, 2.0192], + device='cuda:4'), covar=tensor([0.1270, 0.2007, 0.1884, 0.1116, 0.1716, 0.1602, 0.1339, 0.1898], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0095, 0.0111, 0.0093, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 19:39:28,305 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5647, 1.4964, 2.1077, 1.8382, 1.8983, 3.9176, 1.4517, 1.7195], + device='cuda:4'), covar=tensor([0.1041, 0.2009, 0.1447, 0.1099, 0.1652, 0.0229, 0.1726, 0.1984], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0074, 0.0078, 0.0092, 0.0081, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 19:39:35,941 INFO [finetune.py:976] (4/7) Epoch 16, batch 3300, loss[loss=0.172, simple_loss=0.2525, pruned_loss=0.04577, over 4774.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.2525, pruned_loss=0.05761, over 956036.18 frames. ], batch size: 54, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:39:45,109 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89230.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:39:56,763 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.235e+02 1.765e+02 2.004e+02 2.308e+02 3.942e+02, threshold=4.007e+02, percent-clipped=1.0 +2023-03-26 19:40:09,186 INFO [finetune.py:976] (4/7) Epoch 16, batch 3350, loss[loss=0.1861, simple_loss=0.2577, pruned_loss=0.05721, over 4779.00 frames. ], tot_loss[loss=0.1858, simple_loss=0.2548, pruned_loss=0.05839, over 956327.88 frames. ], batch size: 51, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:40:26,864 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0750, 2.9441, 2.7627, 1.2089, 3.0463, 2.2195, 0.5481, 1.8899], + device='cuda:4'), covar=tensor([0.2665, 0.1980, 0.1877, 0.3330, 0.1341, 0.1090, 0.4156, 0.1583], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0176, 0.0160, 0.0129, 0.0158, 0.0123, 0.0147, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 19:40:42,690 INFO [finetune.py:976] (4/7) Epoch 16, batch 3400, loss[loss=0.1974, simple_loss=0.2721, pruned_loss=0.06139, over 4822.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2555, pruned_loss=0.05846, over 955143.89 frames. ], batch size: 33, lr: 3.44e-03, grad_scale: 32.0 +2023-03-26 19:40:56,405 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3547, 1.4094, 1.5234, 1.6166, 1.5661, 2.9873, 1.3409, 1.5325], + device='cuda:4'), covar=tensor([0.1019, 0.1862, 0.1092, 0.0926, 0.1570, 0.0289, 0.1497, 0.1752], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0081, 0.0074, 0.0077, 0.0092, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 19:41:12,554 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.581e+02 1.832e+02 2.219e+02 5.301e+02, threshold=3.664e+02, percent-clipped=1.0 +2023-03-26 19:41:24,424 INFO [finetune.py:976] (4/7) Epoch 16, batch 3450, loss[loss=0.1702, simple_loss=0.233, pruned_loss=0.05366, over 4273.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2542, pruned_loss=0.0579, over 954483.93 frames. ], batch size: 65, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:41:29,837 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89374.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:41:35,740 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89383.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:41:48,195 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2728, 2.9051, 2.7844, 1.2787, 3.0426, 2.2400, 0.7607, 1.9732], + device='cuda:4'), covar=tensor([0.2361, 0.2285, 0.1817, 0.3689, 0.1306, 0.1137, 0.3997, 0.1648], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0175, 0.0159, 0.0128, 0.0158, 0.0123, 0.0147, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 19:41:58,330 INFO [finetune.py:976] (4/7) Epoch 16, batch 3500, loss[loss=0.1674, simple_loss=0.2366, pruned_loss=0.04906, over 4758.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2519, pruned_loss=0.05752, over 954967.22 frames. ], batch size: 27, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:42:04,412 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89425.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:42:10,963 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89435.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:42:13,380 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89439.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:42:18,648 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.936e+01 1.517e+02 1.946e+02 2.225e+02 4.216e+02, threshold=3.891e+02, percent-clipped=3.0 +2023-03-26 19:42:19,941 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7456, 1.5621, 2.2400, 3.4357, 2.2896, 2.4999, 1.2370, 2.7992], + device='cuda:4'), covar=tensor([0.2048, 0.1813, 0.1658, 0.0910, 0.1021, 0.1707, 0.2246, 0.0641], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0116, 0.0134, 0.0165, 0.0101, 0.0139, 0.0125, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 19:42:31,134 INFO [finetune.py:976] (4/7) Epoch 16, batch 3550, loss[loss=0.1388, simple_loss=0.2057, pruned_loss=0.03593, over 4791.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2497, pruned_loss=0.05688, over 955374.68 frames. ], batch size: 26, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:42:38,750 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.91 vs. limit=5.0 +2023-03-26 19:42:40,863 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.6077, 3.9265, 4.1910, 4.4705, 4.3410, 4.0828, 4.6765, 1.5349], + device='cuda:4'), covar=tensor([0.0795, 0.1029, 0.0788, 0.0970, 0.1372, 0.1386, 0.0672, 0.5759], + device='cuda:4'), in_proj_covar=tensor([0.0352, 0.0246, 0.0277, 0.0295, 0.0335, 0.0282, 0.0299, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:42:43,928 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89486.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:42:53,371 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89500.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 19:42:59,275 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89509.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:43:01,425 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-03-26 19:43:06,068 INFO [finetune.py:976] (4/7) Epoch 16, batch 3600, loss[loss=0.1858, simple_loss=0.2523, pruned_loss=0.05962, over 4101.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2479, pruned_loss=0.05627, over 952880.46 frames. ], batch size: 65, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:43:14,214 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89525.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:43:37,734 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.924e+01 1.526e+02 1.890e+02 2.215e+02 3.895e+02, threshold=3.780e+02, percent-clipped=1.0 +2023-03-26 19:43:55,671 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5436, 2.1066, 1.7447, 0.7585, 2.0218, 1.9376, 1.6777, 1.9843], + device='cuda:4'), covar=tensor([0.0800, 0.1187, 0.1747, 0.2315, 0.1694, 0.2464, 0.2517, 0.1120], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0196, 0.0200, 0.0182, 0.0213, 0.0207, 0.0223, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:44:03,076 INFO [finetune.py:976] (4/7) Epoch 16, batch 3650, loss[loss=0.1995, simple_loss=0.2743, pruned_loss=0.06236, over 4912.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.251, pruned_loss=0.05744, over 953421.37 frames. ], batch size: 43, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:44:06,127 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89570.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:44:36,728 INFO [finetune.py:976] (4/7) Epoch 16, batch 3700, loss[loss=0.1906, simple_loss=0.2646, pruned_loss=0.05826, over 4827.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.253, pruned_loss=0.05744, over 953671.97 frames. ], batch size: 47, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:44:47,567 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3591, 2.0994, 2.7416, 1.6223, 2.3585, 2.6623, 1.9112, 2.9027], + device='cuda:4'), covar=tensor([0.1292, 0.1925, 0.1567, 0.2225, 0.1025, 0.1448, 0.2502, 0.0821], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0203, 0.0190, 0.0189, 0.0176, 0.0211, 0.0216, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:44:50,582 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5967, 1.6321, 1.8551, 1.8293, 1.7294, 2.9552, 1.4799, 1.6235], + device='cuda:4'), covar=tensor([0.0865, 0.1475, 0.1258, 0.0836, 0.1382, 0.0301, 0.1330, 0.1543], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0073, 0.0077, 0.0092, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 19:44:56,694 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-26 19:44:57,073 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.548e+01 1.593e+02 1.994e+02 2.376e+02 3.738e+02, threshold=3.989e+02, percent-clipped=0.0 +2023-03-26 19:45:03,063 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4690, 2.1639, 2.8889, 1.6121, 2.5001, 2.6837, 1.9818, 2.8611], + device='cuda:4'), covar=tensor([0.1280, 0.1777, 0.1369, 0.2372, 0.0933, 0.1395, 0.2533, 0.0773], + device='cuda:4'), in_proj_covar=tensor([0.0191, 0.0203, 0.0190, 0.0189, 0.0175, 0.0211, 0.0216, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:45:04,287 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2489, 2.3028, 2.3180, 1.8344, 2.0045, 2.6089, 2.5638, 1.9717], + device='cuda:4'), covar=tensor([0.0553, 0.0503, 0.0663, 0.0976, 0.1642, 0.0510, 0.0429, 0.0976], + device='cuda:4'), in_proj_covar=tensor([0.0137, 0.0137, 0.0145, 0.0126, 0.0126, 0.0143, 0.0144, 0.0168], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:45:10,212 INFO [finetune.py:976] (4/7) Epoch 16, batch 3750, loss[loss=0.1449, simple_loss=0.2185, pruned_loss=0.03564, over 4698.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2541, pruned_loss=0.05774, over 954095.27 frames. ], batch size: 23, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:45:19,312 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89680.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:45:21,103 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=89683.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:45:43,312 INFO [finetune.py:976] (4/7) Epoch 16, batch 3800, loss[loss=0.1821, simple_loss=0.263, pruned_loss=0.05059, over 4715.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2558, pruned_loss=0.05826, over 954525.97 frames. ], batch size: 59, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:45:52,763 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89730.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:45:53,367 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=89731.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:46:00,039 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89741.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:46:03,525 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.549e+02 1.882e+02 2.353e+02 4.344e+02, threshold=3.764e+02, percent-clipped=2.0 +2023-03-26 19:46:09,907 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-26 19:46:19,027 INFO [finetune.py:976] (4/7) Epoch 16, batch 3850, loss[loss=0.1538, simple_loss=0.2202, pruned_loss=0.04373, over 4762.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2541, pruned_loss=0.05796, over 953215.13 frames. ], batch size: 27, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:46:29,111 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89781.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:46:38,086 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89795.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:46:38,110 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5834, 1.6155, 2.0887, 3.3881, 2.3064, 2.5169, 0.8529, 2.7623], + device='cuda:4'), covar=tensor([0.1854, 0.1366, 0.1388, 0.0487, 0.0840, 0.1322, 0.2067, 0.0471], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0117, 0.0135, 0.0167, 0.0102, 0.0140, 0.0126, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 19:46:38,748 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3202, 2.3491, 2.1083, 2.5510, 2.9383, 2.4151, 2.4260, 1.7882], + device='cuda:4'), covar=tensor([0.2174, 0.1845, 0.1832, 0.1507, 0.1717, 0.1057, 0.1876, 0.1977], + device='cuda:4'), in_proj_covar=tensor([0.0240, 0.0208, 0.0210, 0.0191, 0.0242, 0.0184, 0.0215, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:46:39,351 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3751, 2.1731, 1.6899, 0.7397, 1.9361, 1.9067, 1.7817, 2.0645], + device='cuda:4'), covar=tensor([0.0851, 0.0752, 0.1452, 0.2019, 0.1298, 0.1987, 0.1836, 0.0800], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0195, 0.0200, 0.0182, 0.0213, 0.0207, 0.0223, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:46:44,629 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.9193, 4.2134, 4.4173, 4.6829, 4.6140, 4.4067, 5.0242, 1.6214], + device='cuda:4'), covar=tensor([0.0696, 0.0805, 0.0717, 0.0829, 0.1130, 0.1374, 0.0500, 0.5287], + device='cuda:4'), in_proj_covar=tensor([0.0349, 0.0245, 0.0276, 0.0293, 0.0332, 0.0281, 0.0298, 0.0294], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:46:52,162 INFO [finetune.py:976] (4/7) Epoch 16, batch 3900, loss[loss=0.2027, simple_loss=0.2686, pruned_loss=0.06835, over 4845.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2519, pruned_loss=0.05725, over 952674.51 frames. ], batch size: 47, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:46:58,255 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=89825.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:47:12,472 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.084e+02 1.500e+02 1.857e+02 2.274e+02 5.172e+02, threshold=3.715e+02, percent-clipped=1.0 +2023-03-26 19:47:15,640 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8279, 2.0967, 1.8090, 1.7769, 2.6551, 2.5309, 2.0931, 2.1162], + device='cuda:4'), covar=tensor([0.0503, 0.0377, 0.0523, 0.0367, 0.0201, 0.0544, 0.0485, 0.0405], + device='cuda:4'), in_proj_covar=tensor([0.0096, 0.0109, 0.0145, 0.0114, 0.0101, 0.0109, 0.0100, 0.0109], + device='cuda:4'), out_proj_covar=tensor([7.4242e-05, 8.4452e-05, 1.1452e-04, 8.7791e-05, 7.8347e-05, 8.0138e-05, + 7.4841e-05, 8.3367e-05], device='cuda:4') +2023-03-26 19:47:23,888 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89865.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:47:23,924 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89865.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:47:24,440 INFO [finetune.py:976] (4/7) Epoch 16, batch 3950, loss[loss=0.1507, simple_loss=0.2219, pruned_loss=0.03976, over 4914.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2481, pruned_loss=0.05593, over 951715.65 frames. ], batch size: 36, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:47:29,793 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=89873.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:47:30,483 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0721, 2.0498, 1.6742, 1.8832, 2.0989, 1.7926, 2.4385, 2.0829], + device='cuda:4'), covar=tensor([0.1481, 0.2102, 0.2992, 0.2541, 0.2543, 0.1730, 0.2809, 0.1917], + device='cuda:4'), in_proj_covar=tensor([0.0181, 0.0187, 0.0233, 0.0252, 0.0243, 0.0201, 0.0211, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:47:47,665 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1043, 1.9937, 2.1489, 1.6072, 2.2363, 2.2988, 2.2849, 1.7167], + device='cuda:4'), covar=tensor([0.0521, 0.0612, 0.0566, 0.0801, 0.0630, 0.0512, 0.0465, 0.1049], + device='cuda:4'), in_proj_covar=tensor([0.0136, 0.0137, 0.0144, 0.0126, 0.0126, 0.0143, 0.0144, 0.0167], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:47:57,666 INFO [finetune.py:976] (4/7) Epoch 16, batch 4000, loss[loss=0.1712, simple_loss=0.2395, pruned_loss=0.05145, over 4904.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2477, pruned_loss=0.05602, over 953105.65 frames. ], batch size: 36, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:48:04,742 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89926.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:48:08,787 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.0206, 3.1991, 2.9416, 2.2133, 2.9938, 3.2092, 3.1269, 2.8080], + device='cuda:4'), covar=tensor([0.0555, 0.0502, 0.0664, 0.0860, 0.0646, 0.0619, 0.0587, 0.0864], + device='cuda:4'), in_proj_covar=tensor([0.0136, 0.0136, 0.0143, 0.0125, 0.0125, 0.0143, 0.0143, 0.0167], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:48:17,813 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8168, 1.7081, 2.0608, 2.0213, 1.9743, 3.5183, 1.7159, 1.8711], + device='cuda:4'), covar=tensor([0.0897, 0.1746, 0.0986, 0.0873, 0.1366, 0.0278, 0.1379, 0.1637], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0082, 0.0074, 0.0078, 0.0092, 0.0081, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 19:48:18,309 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.466e+01 1.676e+02 1.974e+02 2.469e+02 4.779e+02, threshold=3.947e+02, percent-clipped=6.0 +2023-03-26 19:48:32,941 INFO [finetune.py:976] (4/7) Epoch 16, batch 4050, loss[loss=0.2057, simple_loss=0.2627, pruned_loss=0.07435, over 4769.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2513, pruned_loss=0.05757, over 950895.27 frames. ], batch size: 26, lr: 3.43e-03, grad_scale: 64.0 +2023-03-26 19:49:01,484 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 19:49:29,802 INFO [finetune.py:976] (4/7) Epoch 16, batch 4100, loss[loss=0.2167, simple_loss=0.2474, pruned_loss=0.09298, over 4360.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2549, pruned_loss=0.05886, over 950615.30 frames. ], batch size: 18, lr: 3.43e-03, grad_scale: 64.0 +2023-03-26 19:49:30,517 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6840, 1.4717, 1.9091, 1.2856, 1.7019, 1.7914, 1.4447, 2.0199], + device='cuda:4'), covar=tensor([0.1132, 0.2010, 0.1142, 0.1592, 0.0859, 0.1303, 0.2569, 0.0827], + device='cuda:4'), in_proj_covar=tensor([0.0194, 0.0206, 0.0192, 0.0192, 0.0177, 0.0214, 0.0220, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:49:32,992 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8411, 1.7864, 1.6036, 1.9323, 2.1466, 1.8920, 1.6254, 1.6072], + device='cuda:4'), covar=tensor([0.1706, 0.1670, 0.1549, 0.1338, 0.1537, 0.1034, 0.2209, 0.1515], + device='cuda:4'), in_proj_covar=tensor([0.0242, 0.0209, 0.0212, 0.0192, 0.0244, 0.0186, 0.0216, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:49:43,228 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90030.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:49:46,963 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90036.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:49:51,494 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6968, 1.5736, 2.2930, 3.2537, 2.2563, 2.2993, 0.9812, 2.6772], + device='cuda:4'), covar=tensor([0.1661, 0.1434, 0.1221, 0.0558, 0.0777, 0.1816, 0.1884, 0.0479], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0117, 0.0134, 0.0166, 0.0101, 0.0140, 0.0125, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 19:49:54,440 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.559e+02 1.839e+02 2.160e+02 6.359e+02, threshold=3.678e+02, percent-clipped=1.0 +2023-03-26 19:49:58,247 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7011, 1.6580, 1.9382, 1.2895, 1.7600, 1.8925, 1.5801, 2.1055], + device='cuda:4'), covar=tensor([0.1232, 0.2011, 0.1382, 0.1695, 0.0977, 0.1460, 0.2642, 0.0887], + device='cuda:4'), in_proj_covar=tensor([0.0194, 0.0206, 0.0193, 0.0192, 0.0178, 0.0214, 0.0220, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:50:06,386 INFO [finetune.py:976] (4/7) Epoch 16, batch 4150, loss[loss=0.2296, simple_loss=0.2965, pruned_loss=0.08129, over 4812.00 frames. ], tot_loss[loss=0.1866, simple_loss=0.2555, pruned_loss=0.05885, over 950478.61 frames. ], batch size: 47, lr: 3.43e-03, grad_scale: 64.0 +2023-03-26 19:50:14,191 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90078.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:50:16,551 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90081.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:50:26,448 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90095.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 19:50:39,512 INFO [finetune.py:976] (4/7) Epoch 16, batch 4200, loss[loss=0.1603, simple_loss=0.2272, pruned_loss=0.04668, over 4837.00 frames. ], tot_loss[loss=0.1876, simple_loss=0.2566, pruned_loss=0.05926, over 950122.47 frames. ], batch size: 47, lr: 3.43e-03, grad_scale: 64.0 +2023-03-26 19:50:47,974 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90129.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:50:48,626 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90130.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:50:56,641 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1001, 3.5742, 3.7613, 3.9002, 3.8610, 3.6782, 4.1721, 1.3777], + device='cuda:4'), covar=tensor([0.0706, 0.0817, 0.0788, 0.0916, 0.1120, 0.1313, 0.0749, 0.5127], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0246, 0.0278, 0.0296, 0.0335, 0.0282, 0.0300, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:50:57,836 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90143.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 19:51:00,649 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.547e+02 1.785e+02 2.134e+02 3.751e+02, threshold=3.570e+02, percent-clipped=1.0 +2023-03-26 19:51:12,325 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90165.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:51:12,427 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-26 19:51:12,839 INFO [finetune.py:976] (4/7) Epoch 16, batch 4250, loss[loss=0.1521, simple_loss=0.2257, pruned_loss=0.03924, over 4867.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.253, pruned_loss=0.05772, over 952406.79 frames. ], batch size: 34, lr: 3.43e-03, grad_scale: 64.0 +2023-03-26 19:51:29,547 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90191.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:51:43,668 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90213.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:51:45,899 INFO [finetune.py:976] (4/7) Epoch 16, batch 4300, loss[loss=0.2293, simple_loss=0.2845, pruned_loss=0.08699, over 4743.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2512, pruned_loss=0.05724, over 952946.12 frames. ], batch size: 27, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:51:48,959 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90221.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 19:51:56,161 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90232.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:51:59,206 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8640, 1.3984, 0.7799, 1.6154, 2.2479, 1.4432, 1.6290, 1.8042], + device='cuda:4'), covar=tensor([0.1474, 0.2147, 0.2238, 0.1282, 0.1974, 0.1986, 0.1413, 0.1979], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0096, 0.0112, 0.0094, 0.0120, 0.0096, 0.0099, 0.0090], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 19:52:07,166 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.180e+02 1.551e+02 1.797e+02 2.190e+02 3.764e+02, threshold=3.594e+02, percent-clipped=1.0 +2023-03-26 19:52:18,555 INFO [finetune.py:976] (4/7) Epoch 16, batch 4350, loss[loss=0.2285, simple_loss=0.2938, pruned_loss=0.08157, over 4902.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2491, pruned_loss=0.05701, over 951411.08 frames. ], batch size: 37, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:52:36,976 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90293.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:52:47,684 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1289, 3.5650, 3.7265, 3.8662, 3.8863, 3.6977, 4.1976, 1.5094], + device='cuda:4'), covar=tensor([0.0740, 0.1009, 0.0941, 0.1136, 0.1102, 0.1491, 0.0686, 0.5484], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0246, 0.0277, 0.0295, 0.0335, 0.0283, 0.0299, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 19:52:51,897 INFO [finetune.py:976] (4/7) Epoch 16, batch 4400, loss[loss=0.1898, simple_loss=0.2583, pruned_loss=0.06065, over 4840.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2496, pruned_loss=0.05668, over 951713.39 frames. ], batch size: 30, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:52:53,194 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5948, 1.5260, 1.4405, 1.5104, 1.1828, 3.0033, 1.2425, 1.6104], + device='cuda:4'), covar=tensor([0.3997, 0.3070, 0.2428, 0.2998, 0.1811, 0.0340, 0.2485, 0.1257], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0115, 0.0120, 0.0123, 0.0113, 0.0097, 0.0096, 0.0096], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 19:53:04,944 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90336.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:53:13,598 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.586e+02 1.845e+02 2.241e+02 4.760e+02, threshold=3.689e+02, percent-clipped=4.0 +2023-03-26 19:53:15,493 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90351.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:53:25,408 INFO [finetune.py:976] (4/7) Epoch 16, batch 4450, loss[loss=0.3039, simple_loss=0.34, pruned_loss=0.1339, over 4805.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.253, pruned_loss=0.05771, over 952199.45 frames. ], batch size: 51, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:53:32,089 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5072, 1.4907, 1.9138, 1.7129, 1.5947, 3.3614, 1.3865, 1.5488], + device='cuda:4'), covar=tensor([0.0975, 0.1803, 0.1069, 0.0976, 0.1593, 0.0250, 0.1486, 0.1829], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0078, 0.0092, 0.0080, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 19:53:34,822 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 19:53:37,365 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90384.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:53:42,217 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3504, 1.4024, 1.4993, 0.8542, 1.4712, 1.7634, 1.7685, 1.2856], + device='cuda:4'), covar=tensor([0.0917, 0.0616, 0.0427, 0.0504, 0.0467, 0.0529, 0.0295, 0.0682], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0151, 0.0124, 0.0127, 0.0131, 0.0129, 0.0143, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.2502e-05, 1.0999e-04, 8.8710e-05, 9.0567e-05, 9.2700e-05, 9.2904e-05, + 1.0268e-04, 1.0651e-04], device='cuda:4') +2023-03-26 19:54:05,563 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90412.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:54:07,882 INFO [finetune.py:976] (4/7) Epoch 16, batch 4500, loss[loss=0.1943, simple_loss=0.2667, pruned_loss=0.06093, over 4844.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.2534, pruned_loss=0.05749, over 952407.28 frames. ], batch size: 49, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:54:40,947 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.141e+02 1.659e+02 2.056e+02 2.631e+02 3.688e+02, threshold=4.111e+02, percent-clipped=0.0 +2023-03-26 19:54:53,433 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90459.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:55:01,924 INFO [finetune.py:976] (4/7) Epoch 16, batch 4550, loss[loss=0.175, simple_loss=0.2474, pruned_loss=0.05127, over 4841.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2541, pruned_loss=0.05698, over 954042.64 frames. ], batch size: 30, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:55:18,094 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90486.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:55:24,699 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90496.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:55:38,628 INFO [finetune.py:976] (4/7) Epoch 16, batch 4600, loss[loss=0.1927, simple_loss=0.257, pruned_loss=0.0642, over 4826.00 frames. ], tot_loss[loss=0.1838, simple_loss=0.2537, pruned_loss=0.0569, over 953417.65 frames. ], batch size: 30, lr: 3.43e-03, grad_scale: 32.0 +2023-03-26 19:55:41,650 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90520.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:55:42,202 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90521.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 19:55:59,294 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.556e+01 1.469e+02 1.715e+02 2.012e+02 4.010e+02, threshold=3.429e+02, percent-clipped=0.0 +2023-03-26 19:56:06,291 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90557.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:56:11,546 INFO [finetune.py:976] (4/7) Epoch 16, batch 4650, loss[loss=0.1853, simple_loss=0.245, pruned_loss=0.06279, over 4853.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2521, pruned_loss=0.05712, over 954523.33 frames. ], batch size: 47, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 19:56:14,406 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90569.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:56:26,174 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90588.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:56:45,055 INFO [finetune.py:976] (4/7) Epoch 16, batch 4700, loss[loss=0.1484, simple_loss=0.2238, pruned_loss=0.03653, over 4757.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2494, pruned_loss=0.05619, over 956210.96 frames. ], batch size: 28, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 19:57:05,663 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.103e+02 1.593e+02 1.858e+02 2.163e+02 3.767e+02, threshold=3.717e+02, percent-clipped=1.0 +2023-03-26 19:57:18,494 INFO [finetune.py:976] (4/7) Epoch 16, batch 4750, loss[loss=0.2127, simple_loss=0.272, pruned_loss=0.07673, over 4832.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2485, pruned_loss=0.05622, over 956973.08 frames. ], batch size: 33, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 19:57:45,641 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90707.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:57:52,457 INFO [finetune.py:976] (4/7) Epoch 16, batch 4800, loss[loss=0.2726, simple_loss=0.3363, pruned_loss=0.1044, over 4777.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2507, pruned_loss=0.05703, over 954873.95 frames. ], batch size: 59, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 19:58:13,286 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.231e+02 1.629e+02 1.979e+02 2.321e+02 4.531e+02, threshold=3.957e+02, percent-clipped=1.0 +2023-03-26 19:58:25,073 INFO [finetune.py:976] (4/7) Epoch 16, batch 4850, loss[loss=0.2054, simple_loss=0.2711, pruned_loss=0.06984, over 4737.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2547, pruned_loss=0.05795, over 956134.68 frames. ], batch size: 59, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 19:58:34,040 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.21 vs. limit=5.0 +2023-03-26 19:58:39,071 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90786.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:58:57,843 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90815.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:58:58,404 INFO [finetune.py:976] (4/7) Epoch 16, batch 4900, loss[loss=0.1568, simple_loss=0.2044, pruned_loss=0.05465, over 4317.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2553, pruned_loss=0.05853, over 953210.48 frames. ], batch size: 18, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 19:59:11,175 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90834.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:59:24,121 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.543e+02 1.926e+02 2.205e+02 3.945e+02, threshold=3.852e+02, percent-clipped=0.0 +2023-03-26 19:59:25,345 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90849.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:59:27,133 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90852.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 19:59:42,912 INFO [finetune.py:976] (4/7) Epoch 16, batch 4950, loss[loss=0.2, simple_loss=0.2663, pruned_loss=0.06681, over 4920.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2577, pruned_loss=0.05918, over 955239.37 frames. ], batch size: 33, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 19:59:45,156 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=6.17 vs. limit=5.0 +2023-03-26 20:00:12,915 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90888.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:00:31,917 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 20:00:34,909 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90910.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:00:38,894 INFO [finetune.py:976] (4/7) Epoch 16, batch 5000, loss[loss=0.1902, simple_loss=0.2563, pruned_loss=0.06209, over 4816.00 frames. ], tot_loss[loss=0.186, simple_loss=0.2554, pruned_loss=0.05831, over 952791.23 frames. ], batch size: 38, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:00:53,036 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=90936.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:00:56,734 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8506, 1.9435, 1.6976, 1.6826, 2.4707, 2.3062, 2.0756, 1.9648], + device='cuda:4'), covar=tensor([0.0419, 0.0333, 0.0498, 0.0350, 0.0250, 0.0531, 0.0357, 0.0350], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0108, 0.0143, 0.0112, 0.0099, 0.0107, 0.0098, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.3215e-05, 8.3587e-05, 1.1327e-04, 8.6741e-05, 7.7176e-05, 7.9306e-05, + 7.3623e-05, 8.2679e-05], device='cuda:4') +2023-03-26 20:00:59,824 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.67 vs. limit=5.0 +2023-03-26 20:01:00,205 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.882e+01 1.531e+02 1.843e+02 2.134e+02 5.620e+02, threshold=3.687e+02, percent-clipped=2.0 +2023-03-26 20:01:11,975 INFO [finetune.py:976] (4/7) Epoch 16, batch 5050, loss[loss=0.1479, simple_loss=0.217, pruned_loss=0.03941, over 4740.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.252, pruned_loss=0.05713, over 953169.42 frames. ], batch size: 59, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:01:28,982 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-03-26 20:01:34,420 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 20:01:37,250 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.2887, 3.2155, 3.1845, 2.5170, 3.0767, 3.4645, 3.4444, 2.9456], + device='cuda:4'), covar=tensor([0.0412, 0.0459, 0.0581, 0.0817, 0.0584, 0.0559, 0.0464, 0.0787], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0133, 0.0140, 0.0122, 0.0122, 0.0139, 0.0140, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:01:40,207 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91007.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:01:45,572 INFO [finetune.py:976] (4/7) Epoch 16, batch 5100, loss[loss=0.1967, simple_loss=0.2557, pruned_loss=0.06883, over 4822.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2487, pruned_loss=0.05601, over 952278.48 frames. ], batch size: 39, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:01:46,919 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8847, 1.2872, 1.8525, 1.8909, 1.6854, 1.6018, 1.7948, 1.7365], + device='cuda:4'), covar=tensor([0.3611, 0.3523, 0.3239, 0.3198, 0.4555, 0.3582, 0.4046, 0.2970], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0239, 0.0258, 0.0271, 0.0270, 0.0245, 0.0281, 0.0237], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:02:07,770 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.054e+02 1.471e+02 1.778e+02 2.096e+02 3.940e+02, threshold=3.556e+02, percent-clipped=2.0 +2023-03-26 20:02:12,114 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91055.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:02:16,308 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3718, 2.3036, 1.7964, 0.9645, 2.0211, 1.8184, 1.6307, 2.0433], + device='cuda:4'), covar=tensor([0.0944, 0.0792, 0.1587, 0.2051, 0.1449, 0.2582, 0.2357, 0.1121], + device='cuda:4'), in_proj_covar=tensor([0.0166, 0.0193, 0.0197, 0.0181, 0.0210, 0.0204, 0.0221, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:02:19,222 INFO [finetune.py:976] (4/7) Epoch 16, batch 5150, loss[loss=0.2179, simple_loss=0.2829, pruned_loss=0.07645, over 4838.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2491, pruned_loss=0.0565, over 953231.37 frames. ], batch size: 47, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:02:31,816 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3526, 1.8480, 2.3228, 2.2887, 2.1159, 2.0380, 2.2365, 2.1677], + device='cuda:4'), covar=tensor([0.3335, 0.3714, 0.3366, 0.3420, 0.4437, 0.3355, 0.3964, 0.3027], + device='cuda:4'), in_proj_covar=tensor([0.0247, 0.0239, 0.0259, 0.0271, 0.0270, 0.0245, 0.0282, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:02:50,664 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5467, 0.6651, 1.6486, 1.5376, 1.4355, 1.3368, 1.4576, 1.5668], + device='cuda:4'), covar=tensor([0.3547, 0.3633, 0.2981, 0.3215, 0.4146, 0.3327, 0.3942, 0.2749], + device='cuda:4'), in_proj_covar=tensor([0.0247, 0.0238, 0.0258, 0.0271, 0.0270, 0.0245, 0.0281, 0.0237], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:02:52,434 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91115.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:02:52,959 INFO [finetune.py:976] (4/7) Epoch 16, batch 5200, loss[loss=0.1747, simple_loss=0.2421, pruned_loss=0.05368, over 4875.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2525, pruned_loss=0.05799, over 952651.64 frames. ], batch size: 31, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:03:14,342 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.594e+02 1.966e+02 2.465e+02 4.658e+02, threshold=3.932e+02, percent-clipped=3.0 +2023-03-26 20:03:17,846 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91152.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:03:24,389 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91163.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:03:26,649 INFO [finetune.py:976] (4/7) Epoch 16, batch 5250, loss[loss=0.2088, simple_loss=0.2853, pruned_loss=0.0662, over 4719.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2547, pruned_loss=0.05826, over 951987.94 frames. ], batch size: 59, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:03:44,135 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.59 vs. limit=5.0 +2023-03-26 20:03:49,258 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91200.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:03:53,266 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=91205.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:03:57,202 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.85 vs. limit=5.0 +2023-03-26 20:03:59,841 INFO [finetune.py:976] (4/7) Epoch 16, batch 5300, loss[loss=0.1603, simple_loss=0.24, pruned_loss=0.04029, over 4784.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2562, pruned_loss=0.05877, over 952469.79 frames. ], batch size: 29, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:04:21,117 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.597e+02 1.843e+02 2.222e+02 3.769e+02, threshold=3.686e+02, percent-clipped=0.0 +2023-03-26 20:04:25,678 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.60 vs. limit=5.0 +2023-03-26 20:04:33,502 INFO [finetune.py:976] (4/7) Epoch 16, batch 5350, loss[loss=0.1938, simple_loss=0.2561, pruned_loss=0.06575, over 4761.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2563, pruned_loss=0.05801, over 954300.89 frames. ], batch size: 27, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:04:41,155 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5537, 1.1270, 0.7355, 1.2979, 2.0093, 0.7363, 1.2649, 1.4393], + device='cuda:4'), covar=tensor([0.1522, 0.2154, 0.1844, 0.1293, 0.1857, 0.2090, 0.1495, 0.1933], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0096, 0.0112, 0.0093, 0.0120, 0.0095, 0.0099, 0.0090], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 20:04:50,623 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91285.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:05:08,773 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0812, 1.9252, 1.6574, 1.7705, 1.8550, 1.8257, 1.8513, 2.6318], + device='cuda:4'), covar=tensor([0.3877, 0.4557, 0.3295, 0.4236, 0.4323, 0.2483, 0.3920, 0.1750], + device='cuda:4'), in_proj_covar=tensor([0.0286, 0.0261, 0.0227, 0.0276, 0.0249, 0.0217, 0.0250, 0.0229], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:05:21,320 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.4509, 1.6111, 1.6395, 0.8744, 1.6785, 1.9402, 1.8784, 1.4748], + device='cuda:4'), covar=tensor([0.0977, 0.0575, 0.0628, 0.0639, 0.0540, 0.0691, 0.0381, 0.1013], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0151, 0.0124, 0.0128, 0.0131, 0.0129, 0.0143, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.2695e-05, 1.0959e-04, 8.8706e-05, 9.0921e-05, 9.2306e-05, 9.2757e-05, + 1.0297e-04, 1.0639e-04], device='cuda:4') +2023-03-26 20:05:29,810 INFO [finetune.py:976] (4/7) Epoch 16, batch 5400, loss[loss=0.1811, simple_loss=0.2426, pruned_loss=0.05979, over 4184.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2545, pruned_loss=0.05748, over 955033.30 frames. ], batch size: 65, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:06:01,763 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=91346.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:06:03,346 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.781e+01 1.583e+02 1.812e+02 2.291e+02 3.767e+02, threshold=3.624e+02, percent-clipped=1.0 +2023-03-26 20:06:15,750 INFO [finetune.py:976] (4/7) Epoch 16, batch 5450, loss[loss=0.1585, simple_loss=0.2331, pruned_loss=0.04194, over 4739.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2506, pruned_loss=0.05628, over 952847.19 frames. ], batch size: 23, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:06:19,990 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7893, 1.4191, 0.7309, 1.6680, 2.1448, 1.5408, 1.6330, 1.7499], + device='cuda:4'), covar=tensor([0.1460, 0.1977, 0.2198, 0.1190, 0.1893, 0.2066, 0.1331, 0.1802], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0095, 0.0111, 0.0093, 0.0119, 0.0095, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 20:06:28,021 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 20:06:49,417 INFO [finetune.py:976] (4/7) Epoch 16, batch 5500, loss[loss=0.2281, simple_loss=0.2925, pruned_loss=0.08181, over 4736.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2478, pruned_loss=0.05553, over 952310.79 frames. ], batch size: 54, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:07:07,351 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3206, 2.2624, 1.7554, 2.1184, 2.1793, 1.9203, 2.4928, 2.2479], + device='cuda:4'), covar=tensor([0.1282, 0.2170, 0.3176, 0.2611, 0.2675, 0.1808, 0.3177, 0.1731], + device='cuda:4'), in_proj_covar=tensor([0.0183, 0.0188, 0.0234, 0.0254, 0.0245, 0.0202, 0.0212, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:07:10,223 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.789e+01 1.460e+02 1.744e+02 2.187e+02 6.443e+02, threshold=3.488e+02, percent-clipped=1.0 +2023-03-26 20:07:11,589 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7398, 0.9791, 1.7772, 1.7412, 1.5890, 1.5427, 1.6610, 1.6833], + device='cuda:4'), covar=tensor([0.3780, 0.3933, 0.3068, 0.3418, 0.4419, 0.3522, 0.4096, 0.2968], + device='cuda:4'), in_proj_covar=tensor([0.0247, 0.0240, 0.0259, 0.0272, 0.0270, 0.0245, 0.0282, 0.0239], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:07:22,094 INFO [finetune.py:976] (4/7) Epoch 16, batch 5550, loss[loss=0.1876, simple_loss=0.2552, pruned_loss=0.05999, over 4874.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.25, pruned_loss=0.05683, over 953230.66 frames. ], batch size: 34, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:07:47,586 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91505.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:07:53,905 INFO [finetune.py:976] (4/7) Epoch 16, batch 5600, loss[loss=0.2544, simple_loss=0.311, pruned_loss=0.09889, over 4862.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.2531, pruned_loss=0.05732, over 955660.77 frames. ], batch size: 31, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:08:08,980 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91542.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 20:08:13,218 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.075e+02 1.603e+02 1.969e+02 2.458e+02 5.397e+02, threshold=3.938e+02, percent-clipped=5.0 +2023-03-26 20:08:16,177 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91553.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:08:23,627 INFO [finetune.py:976] (4/7) Epoch 16, batch 5650, loss[loss=0.1634, simple_loss=0.2372, pruned_loss=0.04479, over 4824.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2555, pruned_loss=0.05785, over 956889.60 frames. ], batch size: 33, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:08:34,797 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7745, 1.3184, 0.7991, 1.6626, 2.1770, 1.2728, 1.5589, 1.7019], + device='cuda:4'), covar=tensor([0.1602, 0.2156, 0.2139, 0.1251, 0.1878, 0.2191, 0.1480, 0.2038], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0095, 0.0112, 0.0092, 0.0119, 0.0095, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 20:08:43,136 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2236, 2.1052, 2.2585, 1.4113, 2.2244, 2.2714, 2.2249, 1.9075], + device='cuda:4'), covar=tensor([0.0472, 0.0662, 0.0546, 0.0868, 0.0750, 0.0577, 0.0551, 0.1034], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0134, 0.0141, 0.0124, 0.0124, 0.0140, 0.0142, 0.0165], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:08:45,670 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=91603.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 20:08:53,176 INFO [finetune.py:976] (4/7) Epoch 16, batch 5700, loss[loss=0.171, simple_loss=0.2283, pruned_loss=0.05686, over 4063.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2519, pruned_loss=0.05743, over 938106.96 frames. ], batch size: 17, lr: 3.42e-03, grad_scale: 32.0 +2023-03-26 20:09:03,625 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.85 vs. limit=5.0 +2023-03-26 20:09:07,893 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=91641.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:09:08,023 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.03 vs. limit=5.0 +2023-03-26 20:09:08,327 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 20:09:21,370 INFO [finetune.py:976] (4/7) Epoch 17, batch 0, loss[loss=0.1864, simple_loss=0.2513, pruned_loss=0.06077, over 4905.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.2513, pruned_loss=0.06077, over 4905.00 frames. ], batch size: 38, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:09:21,370 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 20:09:23,577 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8534, 1.0892, 1.9270, 1.8119, 1.6960, 1.6366, 1.7003, 1.8317], + device='cuda:4'), covar=tensor([0.3852, 0.4350, 0.3857, 0.4212, 0.5204, 0.3988, 0.4995, 0.3380], + device='cuda:4'), in_proj_covar=tensor([0.0247, 0.0240, 0.0259, 0.0271, 0.0270, 0.0245, 0.0282, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:09:23,643 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.8916, 3.4960, 3.6027, 3.7524, 3.6680, 3.5041, 3.9546, 1.4058], + device='cuda:4'), covar=tensor([0.0948, 0.0872, 0.0922, 0.1127, 0.1468, 0.1651, 0.0818, 0.4907], + device='cuda:4'), in_proj_covar=tensor([0.0345, 0.0241, 0.0272, 0.0290, 0.0330, 0.0278, 0.0296, 0.0291], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:09:24,163 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4479, 1.2730, 1.3221, 1.4361, 1.7035, 1.5600, 1.3475, 1.2261], + device='cuda:4'), covar=tensor([0.0319, 0.0308, 0.0637, 0.0302, 0.0223, 0.0435, 0.0382, 0.0389], + device='cuda:4'), in_proj_covar=tensor([0.0095, 0.0108, 0.0144, 0.0113, 0.0100, 0.0108, 0.0098, 0.0109], + device='cuda:4'), out_proj_covar=tensor([7.3808e-05, 8.3862e-05, 1.1375e-04, 8.6997e-05, 7.7713e-05, 7.9723e-05, + 7.3573e-05, 8.3274e-05], device='cuda:4') +2023-03-26 20:09:32,016 INFO [finetune.py:1010] (4/7) Epoch 17, validation: loss=0.1591, simple_loss=0.2283, pruned_loss=0.04492, over 2265189.00 frames. +2023-03-26 20:09:32,017 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 20:09:35,490 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.774e+01 1.479e+02 1.757e+02 2.057e+02 5.096e+02, threshold=3.514e+02, percent-clipped=1.0 +2023-03-26 20:10:00,074 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-26 20:10:07,336 INFO [finetune.py:976] (4/7) Epoch 17, batch 50, loss[loss=0.2062, simple_loss=0.2822, pruned_loss=0.06508, over 4901.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2516, pruned_loss=0.05499, over 216416.67 frames. ], batch size: 46, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:10:16,237 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7061, 2.5552, 1.9849, 0.9100, 2.1704, 2.0703, 1.8772, 2.1934], + device='cuda:4'), covar=tensor([0.0775, 0.0679, 0.1433, 0.2102, 0.1295, 0.2112, 0.2044, 0.1005], + device='cuda:4'), in_proj_covar=tensor([0.0167, 0.0194, 0.0198, 0.0181, 0.0211, 0.0205, 0.0222, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:10:23,534 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-03-26 20:10:32,217 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8948, 1.6095, 2.3031, 3.5477, 2.4673, 2.6618, 1.4325, 2.9143], + device='cuda:4'), covar=tensor([0.1873, 0.1886, 0.1558, 0.0916, 0.0909, 0.1364, 0.1850, 0.0677], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0116, 0.0132, 0.0164, 0.0100, 0.0137, 0.0123, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 20:10:52,791 INFO [finetune.py:976] (4/7) Epoch 17, batch 100, loss[loss=0.1762, simple_loss=0.2518, pruned_loss=0.05035, over 4793.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2453, pruned_loss=0.05356, over 379694.86 frames. ], batch size: 51, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:11:01,252 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.600e+02 1.810e+02 2.096e+02 3.529e+02, threshold=3.620e+02, percent-clipped=1.0 +2023-03-26 20:11:09,373 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-03-26 20:11:37,625 INFO [finetune.py:976] (4/7) Epoch 17, batch 150, loss[loss=0.1717, simple_loss=0.2396, pruned_loss=0.05187, over 4821.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.244, pruned_loss=0.05394, over 506699.63 frames. ], batch size: 41, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:12:11,014 INFO [finetune.py:976] (4/7) Epoch 17, batch 200, loss[loss=0.1896, simple_loss=0.2524, pruned_loss=0.06339, over 4827.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2452, pruned_loss=0.05563, over 605447.99 frames. ], batch size: 30, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:12:11,699 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3050, 1.3590, 1.5473, 1.5548, 1.5554, 2.9720, 1.2954, 1.4770], + device='cuda:4'), covar=tensor([0.1087, 0.2011, 0.1128, 0.1018, 0.1665, 0.0313, 0.1626, 0.1923], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 20:12:14,524 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.595e+02 1.938e+02 2.273e+02 4.627e+02, threshold=3.876e+02, percent-clipped=4.0 +2023-03-26 20:12:43,714 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91891.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:12:44,795 INFO [finetune.py:976] (4/7) Epoch 17, batch 250, loss[loss=0.227, simple_loss=0.2989, pruned_loss=0.0776, over 4813.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.249, pruned_loss=0.05629, over 684299.75 frames. ], batch size: 45, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:12:48,434 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=91898.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 20:12:50,141 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8179, 1.6496, 1.5196, 1.7516, 2.1873, 2.0855, 1.7205, 1.5821], + device='cuda:4'), covar=tensor([0.0270, 0.0321, 0.0599, 0.0307, 0.0184, 0.0408, 0.0395, 0.0372], + device='cuda:4'), in_proj_covar=tensor([0.0095, 0.0108, 0.0144, 0.0113, 0.0100, 0.0108, 0.0098, 0.0109], + device='cuda:4'), out_proj_covar=tensor([7.4044e-05, 8.3918e-05, 1.1392e-04, 8.7083e-05, 7.7791e-05, 7.9603e-05, + 7.3818e-05, 8.3000e-05], device='cuda:4') +2023-03-26 20:13:14,610 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6461, 2.5912, 2.6831, 2.0457, 2.8679, 2.8721, 2.9583, 2.0691], + device='cuda:4'), covar=tensor([0.0920, 0.0830, 0.0799, 0.1053, 0.0546, 0.0908, 0.0777, 0.1776], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0134, 0.0141, 0.0124, 0.0124, 0.0140, 0.0141, 0.0165], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:13:17,058 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91941.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:13:18,231 INFO [finetune.py:976] (4/7) Epoch 17, batch 300, loss[loss=0.1769, simple_loss=0.2541, pruned_loss=0.04982, over 4822.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2505, pruned_loss=0.05573, over 744624.26 frames. ], batch size: 33, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:13:21,756 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.873e+01 1.605e+02 2.003e+02 2.239e+02 3.510e+02, threshold=4.006e+02, percent-clipped=0.0 +2023-03-26 20:13:24,437 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=91952.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:13:27,188 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91955.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:13:30,076 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1264, 1.3213, 1.4500, 0.7758, 1.2988, 1.6068, 1.6545, 1.2949], + device='cuda:4'), covar=tensor([0.0975, 0.0611, 0.0550, 0.0523, 0.0496, 0.0632, 0.0337, 0.0758], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0151, 0.0123, 0.0128, 0.0131, 0.0128, 0.0143, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.2653e-05, 1.0950e-04, 8.8382e-05, 9.0729e-05, 9.2254e-05, 9.2388e-05, + 1.0296e-04, 1.0613e-04], device='cuda:4') +2023-03-26 20:13:42,648 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3238, 2.3474, 1.9323, 2.4011, 2.1958, 2.2594, 2.2298, 3.0878], + device='cuda:4'), covar=tensor([0.3637, 0.4984, 0.3424, 0.4327, 0.4590, 0.2487, 0.4460, 0.1743], + device='cuda:4'), in_proj_covar=tensor([0.0287, 0.0262, 0.0228, 0.0277, 0.0251, 0.0218, 0.0250, 0.0231], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:13:44,456 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91981.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 20:13:49,292 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=91989.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:13:52,146 INFO [finetune.py:976] (4/7) Epoch 17, batch 350, loss[loss=0.1783, simple_loss=0.2478, pruned_loss=0.0544, over 4801.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2518, pruned_loss=0.05612, over 787811.52 frames. ], batch size: 39, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:13:54,110 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5704, 1.5086, 1.4825, 1.4836, 1.0580, 3.0061, 1.1068, 1.5362], + device='cuda:4'), covar=tensor([0.3590, 0.2691, 0.2295, 0.2594, 0.1938, 0.0277, 0.2805, 0.1397], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0115, 0.0120, 0.0123, 0.0113, 0.0096, 0.0096, 0.0096], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 20:14:09,814 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92016.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:14:11,114 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 20:14:26,156 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92042.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 20:14:26,649 INFO [finetune.py:976] (4/7) Epoch 17, batch 400, loss[loss=0.1991, simple_loss=0.2603, pruned_loss=0.06891, over 4859.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2532, pruned_loss=0.05678, over 825981.50 frames. ], batch size: 34, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:14:30,187 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.886e+01 1.544e+02 1.847e+02 2.163e+02 3.487e+02, threshold=3.695e+02, percent-clipped=0.0 +2023-03-26 20:14:48,521 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 20:15:00,227 INFO [finetune.py:976] (4/7) Epoch 17, batch 450, loss[loss=0.1697, simple_loss=0.2251, pruned_loss=0.05711, over 4826.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2516, pruned_loss=0.05668, over 854767.74 frames. ], batch size: 30, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:15:06,684 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0983, 1.6923, 2.1714, 3.9076, 2.7020, 2.7422, 0.7360, 3.2550], + device='cuda:4'), covar=tensor([0.1549, 0.1466, 0.1453, 0.0541, 0.0734, 0.1330, 0.2039, 0.0405], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0116, 0.0133, 0.0165, 0.0101, 0.0137, 0.0124, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 20:15:10,826 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9057, 1.8420, 1.5714, 1.9860, 2.3415, 2.0288, 1.6387, 1.5592], + device='cuda:4'), covar=tensor([0.2103, 0.1995, 0.1904, 0.1605, 0.1677, 0.1189, 0.2453, 0.1937], + device='cuda:4'), in_proj_covar=tensor([0.0243, 0.0209, 0.0213, 0.0192, 0.0244, 0.0186, 0.0217, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:15:28,974 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1981, 3.6469, 3.8507, 4.0503, 3.9608, 3.7113, 4.2744, 1.3698], + device='cuda:4'), covar=tensor([0.0813, 0.0867, 0.0888, 0.0962, 0.1340, 0.1609, 0.0774, 0.5482], + device='cuda:4'), in_proj_covar=tensor([0.0351, 0.0246, 0.0278, 0.0295, 0.0336, 0.0283, 0.0302, 0.0298], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:15:33,735 INFO [finetune.py:976] (4/7) Epoch 17, batch 500, loss[loss=0.1797, simple_loss=0.2471, pruned_loss=0.05609, over 4818.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2496, pruned_loss=0.05608, over 878504.11 frames. ], batch size: 39, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:15:37,216 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.578e+02 1.863e+02 2.269e+02 4.074e+02, threshold=3.727e+02, percent-clipped=2.0 +2023-03-26 20:16:30,567 INFO [finetune.py:976] (4/7) Epoch 17, batch 550, loss[loss=0.213, simple_loss=0.2815, pruned_loss=0.0722, over 4871.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2473, pruned_loss=0.05552, over 893983.08 frames. ], batch size: 34, lr: 3.41e-03, grad_scale: 32.0 +2023-03-26 20:16:33,717 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92198.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 20:16:58,528 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.80 vs. limit=2.0 +2023-03-26 20:17:13,335 INFO [finetune.py:976] (4/7) Epoch 17, batch 600, loss[loss=0.1975, simple_loss=0.2654, pruned_loss=0.06476, over 4829.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2482, pruned_loss=0.05582, over 907046.73 frames. ], batch size: 33, lr: 3.41e-03, grad_scale: 64.0 +2023-03-26 20:17:15,216 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=92246.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 20:17:15,827 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92247.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:17:16,353 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.142e+02 1.709e+02 1.980e+02 2.349e+02 5.069e+02, threshold=3.960e+02, percent-clipped=5.0 +2023-03-26 20:17:47,090 INFO [finetune.py:976] (4/7) Epoch 17, batch 650, loss[loss=0.1867, simple_loss=0.2752, pruned_loss=0.0491, over 4835.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2515, pruned_loss=0.0567, over 917926.33 frames. ], batch size: 47, lr: 3.41e-03, grad_scale: 64.0 +2023-03-26 20:17:58,671 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92311.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:18:16,782 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92337.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 20:18:20,698 INFO [finetune.py:976] (4/7) Epoch 17, batch 700, loss[loss=0.1611, simple_loss=0.2406, pruned_loss=0.04085, over 4742.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2539, pruned_loss=0.05731, over 924907.38 frames. ], batch size: 54, lr: 3.41e-03, grad_scale: 64.0 +2023-03-26 20:18:23,727 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.116e+02 1.553e+02 1.844e+02 2.135e+02 3.970e+02, threshold=3.688e+02, percent-clipped=1.0 +2023-03-26 20:18:54,358 INFO [finetune.py:976] (4/7) Epoch 17, batch 750, loss[loss=0.2121, simple_loss=0.2684, pruned_loss=0.07791, over 4896.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2553, pruned_loss=0.05756, over 930761.88 frames. ], batch size: 36, lr: 3.41e-03, grad_scale: 64.0 +2023-03-26 20:19:28,150 INFO [finetune.py:976] (4/7) Epoch 17, batch 800, loss[loss=0.1728, simple_loss=0.2377, pruned_loss=0.05398, over 4920.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2547, pruned_loss=0.05729, over 935751.76 frames. ], batch size: 33, lr: 3.41e-03, grad_scale: 64.0 +2023-03-26 20:19:31,195 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.199e+02 1.753e+02 1.963e+02 2.342e+02 4.288e+02, threshold=3.926e+02, percent-clipped=2.0 +2023-03-26 20:19:38,890 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-26 20:20:01,479 INFO [finetune.py:976] (4/7) Epoch 17, batch 850, loss[loss=0.1804, simple_loss=0.2521, pruned_loss=0.05436, over 4855.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2526, pruned_loss=0.05666, over 938499.99 frames. ], batch size: 49, lr: 3.41e-03, grad_scale: 64.0 +2023-03-26 20:20:11,365 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-26 20:20:13,114 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6188, 0.6800, 1.6258, 1.5523, 1.4383, 1.3569, 1.5338, 1.5646], + device='cuda:4'), covar=tensor([0.3060, 0.3484, 0.2939, 0.3071, 0.4000, 0.3061, 0.3613, 0.2773], + device='cuda:4'), in_proj_covar=tensor([0.0247, 0.0240, 0.0257, 0.0271, 0.0270, 0.0245, 0.0282, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:20:25,474 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4764, 1.2696, 1.3768, 1.3946, 1.6979, 1.6663, 1.4146, 1.3020], + device='cuda:4'), covar=tensor([0.0318, 0.0324, 0.0550, 0.0346, 0.0242, 0.0410, 0.0346, 0.0385], + device='cuda:4'), in_proj_covar=tensor([0.0095, 0.0108, 0.0144, 0.0113, 0.0100, 0.0108, 0.0098, 0.0109], + device='cuda:4'), out_proj_covar=tensor([7.4046e-05, 8.3648e-05, 1.1374e-04, 8.7156e-05, 7.7625e-05, 7.9571e-05, + 7.3760e-05, 8.3147e-05], device='cuda:4') +2023-03-26 20:20:35,311 INFO [finetune.py:976] (4/7) Epoch 17, batch 900, loss[loss=0.2167, simple_loss=0.2626, pruned_loss=0.08539, over 4725.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2506, pruned_loss=0.05627, over 943636.11 frames. ], batch size: 59, lr: 3.41e-03, grad_scale: 64.0 +2023-03-26 20:20:38,307 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92547.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:20:38,827 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.037e+02 1.480e+02 1.791e+02 2.296e+02 4.324e+02, threshold=3.582e+02, percent-clipped=2.0 +2023-03-26 20:21:15,095 INFO [finetune.py:976] (4/7) Epoch 17, batch 950, loss[loss=0.213, simple_loss=0.2763, pruned_loss=0.07483, over 4841.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2492, pruned_loss=0.05617, over 947458.56 frames. ], batch size: 44, lr: 3.40e-03, grad_scale: 64.0 +2023-03-26 20:21:16,913 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=92595.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:21:37,256 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92611.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:21:54,783 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92625.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:22:05,420 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92637.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 20:22:13,801 INFO [finetune.py:976] (4/7) Epoch 17, batch 1000, loss[loss=0.1549, simple_loss=0.2382, pruned_loss=0.03578, over 4926.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2506, pruned_loss=0.05661, over 950026.40 frames. ], batch size: 38, lr: 3.40e-03, grad_scale: 64.0 +2023-03-26 20:22:20,426 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.492e+01 1.711e+02 2.074e+02 2.603e+02 6.251e+02, threshold=4.148e+02, percent-clipped=4.0 +2023-03-26 20:22:27,818 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=92659.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:22:45,203 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=92685.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 20:22:45,849 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92686.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:22:50,929 INFO [finetune.py:976] (4/7) Epoch 17, batch 1050, loss[loss=0.1989, simple_loss=0.2733, pruned_loss=0.0623, over 4895.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2533, pruned_loss=0.05679, over 952343.86 frames. ], batch size: 35, lr: 3.40e-03, grad_scale: 64.0 +2023-03-26 20:22:58,634 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.56 vs. limit=5.0 +2023-03-26 20:23:08,348 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92720.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:23:19,651 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7216, 1.6469, 2.1840, 3.4846, 2.3412, 2.4259, 0.8814, 2.8918], + device='cuda:4'), covar=tensor([0.1687, 0.1386, 0.1236, 0.0534, 0.0785, 0.1397, 0.1838, 0.0469], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0114, 0.0131, 0.0162, 0.0100, 0.0135, 0.0122, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 20:23:23,714 INFO [finetune.py:976] (4/7) Epoch 17, batch 1100, loss[loss=0.1723, simple_loss=0.2448, pruned_loss=0.04993, over 4877.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2553, pruned_loss=0.05776, over 954729.54 frames. ], batch size: 32, lr: 3.40e-03, grad_scale: 64.0 +2023-03-26 20:23:27,193 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.694e+02 2.013e+02 2.338e+02 4.806e+02, threshold=4.026e+02, percent-clipped=2.0 +2023-03-26 20:23:48,938 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92781.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:23:56,364 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.52 vs. limit=5.0 +2023-03-26 20:23:57,178 INFO [finetune.py:976] (4/7) Epoch 17, batch 1150, loss[loss=0.1789, simple_loss=0.2438, pruned_loss=0.05697, over 4831.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.2562, pruned_loss=0.05807, over 956182.37 frames. ], batch size: 47, lr: 3.40e-03, grad_scale: 64.0 +2023-03-26 20:24:22,948 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92831.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:24:31,080 INFO [finetune.py:976] (4/7) Epoch 17, batch 1200, loss[loss=0.1493, simple_loss=0.23, pruned_loss=0.03433, over 4808.00 frames. ], tot_loss[loss=0.1838, simple_loss=0.2539, pruned_loss=0.05685, over 957111.82 frames. ], batch size: 40, lr: 3.40e-03, grad_scale: 64.0 +2023-03-26 20:24:34,572 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.909e+01 1.547e+02 1.742e+02 2.125e+02 5.044e+02, threshold=3.483e+02, percent-clipped=2.0 +2023-03-26 20:24:52,027 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 20:24:54,821 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8347, 1.2682, 0.8963, 1.7233, 2.1434, 1.4257, 1.5840, 1.6580], + device='cuda:4'), covar=tensor([0.1454, 0.1972, 0.1882, 0.1207, 0.1746, 0.1941, 0.1289, 0.1937], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0110, 0.0092, 0.0118, 0.0094, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 20:25:04,213 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92892.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:25:04,690 INFO [finetune.py:976] (4/7) Epoch 17, batch 1250, loss[loss=0.1643, simple_loss=0.2355, pruned_loss=0.04655, over 4916.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2525, pruned_loss=0.05724, over 957605.80 frames. ], batch size: 37, lr: 3.40e-03, grad_scale: 64.0 +2023-03-26 20:25:08,402 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92899.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:25:29,732 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92931.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:25:37,457 INFO [finetune.py:976] (4/7) Epoch 17, batch 1300, loss[loss=0.1515, simple_loss=0.2163, pruned_loss=0.04337, over 4911.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2497, pruned_loss=0.05653, over 958839.82 frames. ], batch size: 36, lr: 3.40e-03, grad_scale: 64.0 +2023-03-26 20:25:41,344 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.503e+02 1.790e+02 2.154e+02 4.064e+02, threshold=3.581e+02, percent-clipped=1.0 +2023-03-26 20:25:46,883 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 20:25:49,699 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92960.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:25:59,950 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92975.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:26:00,595 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9634, 1.9644, 1.8653, 2.1907, 2.4566, 2.0636, 1.9139, 1.5599], + device='cuda:4'), covar=tensor([0.2449, 0.2089, 0.1901, 0.1545, 0.2111, 0.1215, 0.2335, 0.2027], + device='cuda:4'), in_proj_covar=tensor([0.0240, 0.0207, 0.0210, 0.0190, 0.0240, 0.0184, 0.0214, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:26:01,181 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0169, 0.8712, 0.9470, 1.1202, 1.1737, 1.1500, 0.9602, 0.9314], + device='cuda:4'), covar=tensor([0.0361, 0.0326, 0.0692, 0.0321, 0.0299, 0.0507, 0.0346, 0.0411], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0107, 0.0143, 0.0112, 0.0098, 0.0107, 0.0098, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.3238e-05, 8.2923e-05, 1.1289e-04, 8.6591e-05, 7.6617e-05, 7.8750e-05, + 7.3339e-05, 8.2664e-05], device='cuda:4') +2023-03-26 20:26:03,543 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92981.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:26:10,758 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92992.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:26:11,265 INFO [finetune.py:976] (4/7) Epoch 17, batch 1350, loss[loss=0.1911, simple_loss=0.2723, pruned_loss=0.05494, over 4836.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2497, pruned_loss=0.05662, over 956483.26 frames. ], batch size: 49, lr: 3.40e-03, grad_scale: 64.0 +2023-03-26 20:26:23,865 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93010.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:26:51,548 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93036.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:26:51,558 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93036.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:27:00,931 INFO [finetune.py:976] (4/7) Epoch 17, batch 1400, loss[loss=0.1626, simple_loss=0.2487, pruned_loss=0.03826, over 4819.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2514, pruned_loss=0.05683, over 956816.99 frames. ], batch size: 38, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:27:08,967 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.549e+02 1.883e+02 2.310e+02 4.523e+02, threshold=3.767e+02, percent-clipped=3.0 +2023-03-26 20:27:28,100 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2984, 2.1831, 1.8584, 2.2741, 2.0724, 2.1030, 2.0916, 3.0548], + device='cuda:4'), covar=tensor([0.3937, 0.4966, 0.3414, 0.4265, 0.4963, 0.2419, 0.4424, 0.1607], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0260, 0.0225, 0.0275, 0.0249, 0.0217, 0.0248, 0.0229], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:27:33,627 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-03-26 20:27:34,726 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93071.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:27:39,788 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93076.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:27:50,100 INFO [finetune.py:976] (4/7) Epoch 17, batch 1450, loss[loss=0.1787, simple_loss=0.2441, pruned_loss=0.05662, over 4764.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2532, pruned_loss=0.05692, over 957688.06 frames. ], batch size: 28, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:27:53,125 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93097.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 20:28:05,993 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93115.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:28:23,787 INFO [finetune.py:976] (4/7) Epoch 17, batch 1500, loss[loss=0.2336, simple_loss=0.2907, pruned_loss=0.08821, over 4815.00 frames. ], tot_loss[loss=0.1866, simple_loss=0.2559, pruned_loss=0.0586, over 954438.73 frames. ], batch size: 40, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:28:27,863 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.132e+02 1.651e+02 1.993e+02 2.270e+02 5.642e+02, threshold=3.987e+02, percent-clipped=1.0 +2023-03-26 20:28:47,134 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93176.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:28:53,712 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93187.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:28:57,281 INFO [finetune.py:976] (4/7) Epoch 17, batch 1550, loss[loss=0.2004, simple_loss=0.2701, pruned_loss=0.06533, over 4920.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2544, pruned_loss=0.05744, over 953801.22 frames. ], batch size: 38, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:29:30,932 INFO [finetune.py:976] (4/7) Epoch 17, batch 1600, loss[loss=0.1585, simple_loss=0.234, pruned_loss=0.04149, over 4813.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2527, pruned_loss=0.05708, over 955125.84 frames. ], batch size: 40, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:29:34,598 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.059e+02 1.546e+02 1.807e+02 2.216e+02 3.989e+02, threshold=3.613e+02, percent-clipped=1.0 +2023-03-26 20:29:38,729 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93255.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:29:57,387 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93281.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:30:01,026 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93287.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:30:04,615 INFO [finetune.py:976] (4/7) Epoch 17, batch 1650, loss[loss=0.1721, simple_loss=0.2432, pruned_loss=0.05053, over 4776.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2496, pruned_loss=0.05622, over 956158.07 frames. ], batch size: 27, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:30:28,980 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93329.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:30:30,700 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93331.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:30:38,314 INFO [finetune.py:976] (4/7) Epoch 17, batch 1700, loss[loss=0.1759, simple_loss=0.2224, pruned_loss=0.06473, over 4193.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2476, pruned_loss=0.0558, over 955134.18 frames. ], batch size: 17, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:30:41,945 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.487e+02 1.694e+02 2.142e+02 3.933e+02, threshold=3.388e+02, percent-clipped=2.0 +2023-03-26 20:30:53,819 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93366.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:30:56,302 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2280, 2.0763, 1.7827, 1.9750, 1.9247, 1.9435, 1.9449, 2.7466], + device='cuda:4'), covar=tensor([0.3728, 0.4439, 0.3319, 0.4207, 0.4203, 0.2496, 0.4044, 0.1778], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0260, 0.0227, 0.0276, 0.0250, 0.0217, 0.0250, 0.0230], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:31:00,381 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93376.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:31:11,633 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93392.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 20:31:12,186 INFO [finetune.py:976] (4/7) Epoch 17, batch 1750, loss[loss=0.1844, simple_loss=0.2554, pruned_loss=0.05669, over 4764.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2503, pruned_loss=0.05682, over 955426.36 frames. ], batch size: 26, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:31:33,283 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93424.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:31:48,779 INFO [finetune.py:976] (4/7) Epoch 17, batch 1800, loss[loss=0.2044, simple_loss=0.2663, pruned_loss=0.07127, over 4881.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2542, pruned_loss=0.05784, over 954158.43 frames. ], batch size: 32, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:31:56,894 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.552e+02 1.846e+02 2.179e+02 3.576e+02, threshold=3.692e+02, percent-clipped=3.0 +2023-03-26 20:32:20,903 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93471.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:32:41,047 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93487.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:32:49,595 INFO [finetune.py:976] (4/7) Epoch 17, batch 1850, loss[loss=0.1961, simple_loss=0.2756, pruned_loss=0.05825, over 4861.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.2553, pruned_loss=0.05843, over 954954.15 frames. ], batch size: 34, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:33:20,424 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93535.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:33:26,142 INFO [finetune.py:976] (4/7) Epoch 17, batch 1900, loss[loss=0.1875, simple_loss=0.2632, pruned_loss=0.05594, over 4840.00 frames. ], tot_loss[loss=0.1858, simple_loss=0.2554, pruned_loss=0.05808, over 955936.78 frames. ], batch size: 44, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:33:30,348 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.241e+02 1.618e+02 1.925e+02 2.327e+02 3.543e+02, threshold=3.851e+02, percent-clipped=0.0 +2023-03-26 20:33:34,112 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93555.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:33:55,450 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93587.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:33:59,456 INFO [finetune.py:976] (4/7) Epoch 17, batch 1950, loss[loss=0.1539, simple_loss=0.2383, pruned_loss=0.03476, over 4761.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2539, pruned_loss=0.05747, over 953330.02 frames. ], batch size: 28, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:34:06,615 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93603.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:34:25,025 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93631.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:34:27,390 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93635.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:34:32,611 INFO [finetune.py:976] (4/7) Epoch 17, batch 2000, loss[loss=0.2169, simple_loss=0.2707, pruned_loss=0.08154, over 4903.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2518, pruned_loss=0.05656, over 954897.14 frames. ], batch size: 35, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:34:37,205 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.489e+01 1.528e+02 1.753e+02 2.103e+02 5.258e+02, threshold=3.506e+02, percent-clipped=1.0 +2023-03-26 20:34:43,929 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6576, 1.5367, 2.1956, 3.5267, 2.4634, 2.5488, 1.0613, 2.8254], + device='cuda:4'), covar=tensor([0.1811, 0.1594, 0.1367, 0.0495, 0.0761, 0.1182, 0.1917, 0.0594], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0116, 0.0134, 0.0165, 0.0101, 0.0136, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 20:34:48,135 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93666.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:34:56,474 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93679.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:35:01,054 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.5875, 3.9824, 4.1669, 4.4478, 4.2835, 4.1110, 4.6774, 1.6100], + device='cuda:4'), covar=tensor([0.0691, 0.0842, 0.0741, 0.0808, 0.1278, 0.1541, 0.0642, 0.5353], + device='cuda:4'), in_proj_covar=tensor([0.0349, 0.0244, 0.0275, 0.0292, 0.0333, 0.0281, 0.0300, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:35:05,804 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93692.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:35:06,327 INFO [finetune.py:976] (4/7) Epoch 17, batch 2050, loss[loss=0.1346, simple_loss=0.2118, pruned_loss=0.02872, over 4776.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2477, pruned_loss=0.05502, over 953397.70 frames. ], batch size: 28, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:35:08,299 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 20:35:20,566 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93714.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:35:29,711 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-26 20:35:37,804 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93740.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:35:39,566 INFO [finetune.py:976] (4/7) Epoch 17, batch 2100, loss[loss=0.1917, simple_loss=0.2551, pruned_loss=0.06415, over 4808.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2482, pruned_loss=0.05559, over 953261.09 frames. ], batch size: 38, lr: 3.40e-03, grad_scale: 32.0 +2023-03-26 20:35:43,622 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.402e+01 1.568e+02 1.860e+02 2.232e+02 5.340e+02, threshold=3.720e+02, percent-clipped=4.0 +2023-03-26 20:35:57,994 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93770.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:35:58,584 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93771.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:36:12,209 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.3115, 2.8641, 3.0232, 3.2231, 3.0843, 2.8463, 3.3196, 1.0274], + device='cuda:4'), covar=tensor([0.0996, 0.1053, 0.1068, 0.1069, 0.1546, 0.1852, 0.1164, 0.5282], + device='cuda:4'), in_proj_covar=tensor([0.0349, 0.0244, 0.0275, 0.0292, 0.0334, 0.0282, 0.0301, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:36:13,280 INFO [finetune.py:976] (4/7) Epoch 17, batch 2150, loss[loss=0.2466, simple_loss=0.3058, pruned_loss=0.09374, over 4903.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2521, pruned_loss=0.05733, over 951434.25 frames. ], batch size: 37, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:36:24,188 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-26 20:36:26,354 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5574, 1.4116, 2.2053, 3.3129, 2.2915, 2.3543, 1.0635, 2.6413], + device='cuda:4'), covar=tensor([0.1729, 0.1459, 0.1188, 0.0512, 0.0744, 0.1477, 0.1722, 0.0503], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0116, 0.0134, 0.0165, 0.0101, 0.0137, 0.0125, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 20:36:31,174 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=93819.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:36:38,633 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93831.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:36:47,328 INFO [finetune.py:976] (4/7) Epoch 17, batch 2200, loss[loss=0.2007, simple_loss=0.2692, pruned_loss=0.06608, over 4843.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2542, pruned_loss=0.05821, over 952651.06 frames. ], batch size: 44, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:36:51,482 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.198e+02 1.569e+02 1.869e+02 2.308e+02 4.137e+02, threshold=3.738e+02, percent-clipped=1.0 +2023-03-26 20:37:01,715 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6125, 1.4512, 1.9085, 1.7819, 1.6136, 3.4391, 1.4086, 1.5548], + device='cuda:4'), covar=tensor([0.0896, 0.1822, 0.1102, 0.1040, 0.1679, 0.0225, 0.1558, 0.1800], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0082, 0.0074, 0.0079, 0.0092, 0.0081, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 20:37:36,162 INFO [finetune.py:976] (4/7) Epoch 17, batch 2250, loss[loss=0.189, simple_loss=0.2632, pruned_loss=0.05738, over 4924.00 frames. ], tot_loss[loss=0.1864, simple_loss=0.256, pruned_loss=0.0584, over 955464.04 frames. ], batch size: 33, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:38:30,054 INFO [finetune.py:976] (4/7) Epoch 17, batch 2300, loss[loss=0.1724, simple_loss=0.2516, pruned_loss=0.0466, over 4798.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2562, pruned_loss=0.05811, over 953889.09 frames. ], batch size: 45, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:38:34,185 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.189e+02 1.601e+02 1.890e+02 2.328e+02 3.292e+02, threshold=3.781e+02, percent-clipped=0.0 +2023-03-26 20:38:56,115 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93981.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:39:03,825 INFO [finetune.py:976] (4/7) Epoch 17, batch 2350, loss[loss=0.1671, simple_loss=0.2399, pruned_loss=0.04714, over 4824.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2532, pruned_loss=0.05716, over 954635.54 frames. ], batch size: 39, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:39:35,520 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6527, 0.7254, 1.7072, 1.6195, 1.5321, 1.4285, 1.5638, 1.6338], + device='cuda:4'), covar=tensor([0.3476, 0.3616, 0.3156, 0.3371, 0.4112, 0.3455, 0.3856, 0.2938], + device='cuda:4'), in_proj_covar=tensor([0.0248, 0.0240, 0.0259, 0.0273, 0.0272, 0.0247, 0.0284, 0.0240], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:39:37,907 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=94042.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:39:38,385 INFO [finetune.py:976] (4/7) Epoch 17, batch 2400, loss[loss=0.1078, simple_loss=0.1801, pruned_loss=0.01778, over 4804.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2489, pruned_loss=0.05514, over 955277.51 frames. ], batch size: 25, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:39:42,505 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.024e+02 1.475e+02 1.781e+02 2.110e+02 4.538e+02, threshold=3.563e+02, percent-clipped=2.0 +2023-03-26 20:40:06,881 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1297, 1.3046, 1.2741, 1.2982, 1.4628, 2.4217, 1.2856, 1.4062], + device='cuda:4'), covar=tensor([0.1004, 0.1824, 0.1089, 0.0989, 0.1661, 0.0392, 0.1564, 0.1720], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0078, 0.0091, 0.0081, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 20:40:11,638 INFO [finetune.py:976] (4/7) Epoch 17, batch 2450, loss[loss=0.1692, simple_loss=0.2491, pruned_loss=0.04469, over 4836.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2461, pruned_loss=0.05449, over 956285.24 frames. ], batch size: 40, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:40:29,941 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-26 20:40:34,653 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94126.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:40:37,359 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-26 20:40:42,726 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-26 20:40:45,406 INFO [finetune.py:976] (4/7) Epoch 17, batch 2500, loss[loss=0.2052, simple_loss=0.2723, pruned_loss=0.06903, over 4809.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2494, pruned_loss=0.05611, over 957973.85 frames. ], batch size: 41, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:40:49,551 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.530e+01 1.683e+02 1.909e+02 2.220e+02 4.342e+02, threshold=3.819e+02, percent-clipped=2.0 +2023-03-26 20:41:09,045 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.9154, 4.3212, 4.5268, 4.7870, 4.6442, 4.3706, 5.0167, 1.5347], + device='cuda:4'), covar=tensor([0.0654, 0.0734, 0.0767, 0.0882, 0.1069, 0.1480, 0.0449, 0.5532], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0245, 0.0276, 0.0293, 0.0335, 0.0281, 0.0301, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:41:18,603 INFO [finetune.py:976] (4/7) Epoch 17, batch 2550, loss[loss=0.1778, simple_loss=0.2554, pruned_loss=0.05009, over 4907.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2536, pruned_loss=0.05726, over 956910.14 frames. ], batch size: 36, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:41:18,715 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6533, 1.3905, 2.0404, 1.3660, 1.7428, 1.8723, 1.3866, 1.9821], + device='cuda:4'), covar=tensor([0.1280, 0.2302, 0.0988, 0.1587, 0.0859, 0.1268, 0.2964, 0.0760], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0203, 0.0188, 0.0188, 0.0175, 0.0211, 0.0215, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:41:36,399 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3125, 1.4055, 1.1953, 1.4689, 1.6614, 1.5710, 1.3289, 1.2634], + device='cuda:4'), covar=tensor([0.0366, 0.0255, 0.0591, 0.0256, 0.0171, 0.0463, 0.0324, 0.0335], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0106, 0.0142, 0.0111, 0.0098, 0.0106, 0.0097, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.2644e-05, 8.2039e-05, 1.1184e-04, 8.5786e-05, 7.6515e-05, 7.8633e-05, + 7.2730e-05, 8.2582e-05], device='cuda:4') +2023-03-26 20:41:38,785 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=94223.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 20:41:48,222 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7806, 1.6060, 1.5290, 1.8634, 2.0955, 1.8978, 1.3918, 1.4951], + device='cuda:4'), covar=tensor([0.2238, 0.2043, 0.1921, 0.1713, 0.1680, 0.1222, 0.2355, 0.1933], + device='cuda:4'), in_proj_covar=tensor([0.0241, 0.0208, 0.0212, 0.0192, 0.0241, 0.0185, 0.0215, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:41:52,381 INFO [finetune.py:976] (4/7) Epoch 17, batch 2600, loss[loss=0.1381, simple_loss=0.1983, pruned_loss=0.03893, over 4494.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2548, pruned_loss=0.05778, over 956747.90 frames. ], batch size: 20, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:41:56,017 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.647e+02 1.955e+02 2.265e+02 3.573e+02, threshold=3.911e+02, percent-clipped=0.0 +2023-03-26 20:41:56,143 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6172, 1.8403, 1.5431, 1.5326, 2.1544, 2.1629, 1.8357, 1.8966], + device='cuda:4'), covar=tensor([0.0475, 0.0325, 0.0583, 0.0349, 0.0376, 0.0588, 0.0337, 0.0390], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0106, 0.0142, 0.0111, 0.0099, 0.0107, 0.0098, 0.0109], + device='cuda:4'), out_proj_covar=tensor([7.2950e-05, 8.2321e-05, 1.1228e-04, 8.6046e-05, 7.6890e-05, 7.8957e-05, + 7.3081e-05, 8.2857e-05], device='cuda:4') +2023-03-26 20:42:17,836 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7560, 1.3169, 0.8830, 1.6362, 2.0879, 1.5247, 1.4222, 1.6868], + device='cuda:4'), covar=tensor([0.1373, 0.1953, 0.1898, 0.1115, 0.1860, 0.2005, 0.1356, 0.1715], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0092, 0.0119, 0.0094, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 20:42:19,692 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=94284.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 20:42:22,267 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 20:42:25,432 INFO [finetune.py:976] (4/7) Epoch 17, batch 2650, loss[loss=0.148, simple_loss=0.2293, pruned_loss=0.03336, over 4746.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2566, pruned_loss=0.05824, over 957044.01 frames. ], batch size: 27, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:43:12,615 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94337.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:43:20,896 INFO [finetune.py:976] (4/7) Epoch 17, batch 2700, loss[loss=0.1677, simple_loss=0.2366, pruned_loss=0.0494, over 4756.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2551, pruned_loss=0.05733, over 956250.15 frames. ], batch size: 27, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:43:27,721 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=94348.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:43:28,199 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.025e+02 1.514e+02 1.766e+02 2.145e+02 4.618e+02, threshold=3.532e+02, percent-clipped=2.0 +2023-03-26 20:43:30,767 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3598, 1.4871, 1.5423, 1.6051, 1.6380, 3.0228, 1.4058, 1.6370], + device='cuda:4'), covar=tensor([0.1102, 0.1863, 0.1172, 0.1068, 0.1701, 0.0287, 0.1616, 0.1876], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0082, 0.0074, 0.0079, 0.0092, 0.0081, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 20:44:10,181 INFO [finetune.py:976] (4/7) Epoch 17, batch 2750, loss[loss=0.1411, simple_loss=0.2122, pruned_loss=0.03503, over 4892.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2532, pruned_loss=0.05709, over 956507.34 frames. ], batch size: 32, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:44:16,816 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 20:44:20,186 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=94409.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 20:44:28,821 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-03-26 20:44:31,927 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=94426.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:44:43,083 INFO [finetune.py:976] (4/7) Epoch 17, batch 2800, loss[loss=0.1968, simple_loss=0.2611, pruned_loss=0.06623, over 4827.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2495, pruned_loss=0.05623, over 957446.30 frames. ], batch size: 39, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:44:47,183 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.577e+01 1.587e+02 1.887e+02 2.313e+02 4.372e+02, threshold=3.775e+02, percent-clipped=5.0 +2023-03-26 20:45:03,348 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=94474.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:45:16,199 INFO [finetune.py:976] (4/7) Epoch 17, batch 2850, loss[loss=0.1767, simple_loss=0.2435, pruned_loss=0.05493, over 4819.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2477, pruned_loss=0.05565, over 957991.37 frames. ], batch size: 45, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:45:24,101 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6720, 1.5401, 2.0551, 3.2805, 2.2714, 2.3075, 1.1394, 2.7290], + device='cuda:4'), covar=tensor([0.1718, 0.1424, 0.1306, 0.0517, 0.0799, 0.1348, 0.1799, 0.0510], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0117, 0.0134, 0.0166, 0.0102, 0.0137, 0.0125, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 20:45:39,237 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.57 vs. limit=2.0 +2023-03-26 20:45:42,780 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2357, 2.2558, 1.7838, 2.2688, 2.1520, 1.9092, 2.5558, 2.2880], + device='cuda:4'), covar=tensor([0.1314, 0.2201, 0.2949, 0.2488, 0.2618, 0.1751, 0.3180, 0.1720], + device='cuda:4'), in_proj_covar=tensor([0.0184, 0.0189, 0.0234, 0.0254, 0.0245, 0.0203, 0.0214, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:45:49,604 INFO [finetune.py:976] (4/7) Epoch 17, batch 2900, loss[loss=0.1749, simple_loss=0.2557, pruned_loss=0.04703, over 4925.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2503, pruned_loss=0.05609, over 959865.10 frames. ], batch size: 33, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:45:53,201 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.550e+02 1.801e+02 2.117e+02 3.911e+02, threshold=3.601e+02, percent-clipped=1.0 +2023-03-26 20:46:12,853 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94579.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 20:46:18,797 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7683, 1.3798, 0.8014, 1.6901, 2.1565, 1.5425, 1.6213, 1.6784], + device='cuda:4'), covar=tensor([0.1504, 0.2079, 0.2046, 0.1231, 0.1927, 0.1968, 0.1387, 0.2021], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0092, 0.0119, 0.0094, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 20:46:22,372 INFO [finetune.py:976] (4/7) Epoch 17, batch 2950, loss[loss=0.2055, simple_loss=0.2844, pruned_loss=0.0633, over 4746.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2521, pruned_loss=0.05618, over 959266.21 frames. ], batch size: 54, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:46:30,221 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0500, 0.9191, 0.9672, 1.1584, 1.2009, 1.1653, 0.9808, 0.9540], + device='cuda:4'), covar=tensor([0.0345, 0.0312, 0.0643, 0.0295, 0.0264, 0.0421, 0.0338, 0.0422], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0106, 0.0141, 0.0111, 0.0098, 0.0106, 0.0097, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.2553e-05, 8.1854e-05, 1.1106e-04, 8.5454e-05, 7.6259e-05, 7.8521e-05, + 7.2703e-05, 8.2509e-05], device='cuda:4') +2023-03-26 20:46:46,065 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 20:46:52,134 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=94637.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:46:56,161 INFO [finetune.py:976] (4/7) Epoch 17, batch 3000, loss[loss=0.1598, simple_loss=0.2371, pruned_loss=0.0413, over 4770.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.253, pruned_loss=0.0563, over 957601.71 frames. ], batch size: 28, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:46:56,161 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 20:46:58,312 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4453, 1.3055, 1.3403, 1.5092, 1.7202, 1.6122, 1.3367, 1.2427], + device='cuda:4'), covar=tensor([0.0364, 0.0297, 0.0571, 0.0285, 0.0229, 0.0426, 0.0384, 0.0435], + device='cuda:4'), in_proj_covar=tensor([0.0093, 0.0106, 0.0140, 0.0111, 0.0098, 0.0106, 0.0097, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.2580e-05, 8.1741e-05, 1.1101e-04, 8.5369e-05, 7.6255e-05, 7.8509e-05, + 7.2713e-05, 8.2384e-05], device='cuda:4') +2023-03-26 20:47:06,772 INFO [finetune.py:1010] (4/7) Epoch 17, validation: loss=0.1562, simple_loss=0.2257, pruned_loss=0.04335, over 2265189.00 frames. +2023-03-26 20:47:06,773 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 20:47:10,421 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.006e+02 1.605e+02 1.916e+02 2.337e+02 3.800e+02, threshold=3.832e+02, percent-clipped=2.0 +2023-03-26 20:47:33,746 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=94685.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:47:38,997 INFO [finetune.py:976] (4/7) Epoch 17, batch 3050, loss[loss=0.1646, simple_loss=0.2488, pruned_loss=0.04015, over 4783.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2533, pruned_loss=0.0563, over 958683.17 frames. ], batch size: 29, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:47:47,343 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94704.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 20:48:19,138 INFO [finetune.py:976] (4/7) Epoch 17, batch 3100, loss[loss=0.1766, simple_loss=0.2423, pruned_loss=0.05544, over 4770.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2529, pruned_loss=0.05692, over 957724.67 frames. ], batch size: 28, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:48:27,684 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.533e+02 1.793e+02 2.269e+02 8.706e+02, threshold=3.585e+02, percent-clipped=3.0 +2023-03-26 20:48:35,728 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.9984, 3.4789, 3.6848, 3.7512, 3.8037, 3.6245, 4.0405, 1.8460], + device='cuda:4'), covar=tensor([0.0789, 0.0918, 0.0822, 0.1032, 0.1131, 0.1280, 0.0756, 0.4813], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0245, 0.0277, 0.0294, 0.0336, 0.0281, 0.0303, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:49:17,522 INFO [finetune.py:976] (4/7) Epoch 17, batch 3150, loss[loss=0.1767, simple_loss=0.2411, pruned_loss=0.05617, over 4732.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2512, pruned_loss=0.0565, over 955923.94 frames. ], batch size: 54, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:49:51,404 INFO [finetune.py:976] (4/7) Epoch 17, batch 3200, loss[loss=0.1768, simple_loss=0.2489, pruned_loss=0.05239, over 4839.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2476, pruned_loss=0.0552, over 957444.51 frames. ], batch size: 47, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:49:55,533 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.020e+02 1.463e+02 1.766e+02 2.027e+02 4.168e+02, threshold=3.532e+02, percent-clipped=1.0 +2023-03-26 20:50:16,304 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=94879.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 20:50:25,250 INFO [finetune.py:976] (4/7) Epoch 17, batch 3250, loss[loss=0.1383, simple_loss=0.2093, pruned_loss=0.03364, over 4765.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2476, pruned_loss=0.05525, over 954391.39 frames. ], batch size: 28, lr: 3.39e-03, grad_scale: 32.0 +2023-03-26 20:50:27,583 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-26 20:50:40,692 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0195, 1.7785, 2.2491, 3.9962, 2.8503, 2.6801, 0.7345, 3.3870], + device='cuda:4'), covar=tensor([0.1660, 0.1433, 0.1500, 0.0533, 0.0679, 0.1483, 0.2126, 0.0381], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0115, 0.0133, 0.0164, 0.0100, 0.0135, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 20:50:48,426 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=94927.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 20:50:58,674 INFO [finetune.py:976] (4/7) Epoch 17, batch 3300, loss[loss=0.1983, simple_loss=0.2476, pruned_loss=0.07448, over 4442.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2512, pruned_loss=0.05686, over 953123.94 frames. ], batch size: 19, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 20:51:02,382 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.215e+02 1.785e+02 2.188e+02 2.532e+02 5.228e+02, threshold=4.375e+02, percent-clipped=4.0 +2023-03-26 20:51:32,711 INFO [finetune.py:976] (4/7) Epoch 17, batch 3350, loss[loss=0.2325, simple_loss=0.3043, pruned_loss=0.08038, over 4932.00 frames. ], tot_loss[loss=0.184, simple_loss=0.253, pruned_loss=0.05746, over 953468.58 frames. ], batch size: 41, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 20:51:40,181 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95004.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:52:06,490 INFO [finetune.py:976] (4/7) Epoch 17, batch 3400, loss[loss=0.2167, simple_loss=0.2795, pruned_loss=0.07698, over 4862.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2542, pruned_loss=0.05799, over 954781.78 frames. ], batch size: 44, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:52:10,135 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.060e+02 1.553e+02 1.863e+02 2.086e+02 3.757e+02, threshold=3.727e+02, percent-clipped=0.0 +2023-03-26 20:52:12,047 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=95052.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:52:36,325 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9447, 1.8216, 1.5590, 1.4595, 1.9370, 1.6799, 1.8389, 1.9026], + device='cuda:4'), covar=tensor([0.1526, 0.2226, 0.3315, 0.2641, 0.2738, 0.1895, 0.2819, 0.2035], + device='cuda:4'), in_proj_covar=tensor([0.0183, 0.0188, 0.0234, 0.0253, 0.0245, 0.0202, 0.0212, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:52:40,293 INFO [finetune.py:976] (4/7) Epoch 17, batch 3450, loss[loss=0.2474, simple_loss=0.3061, pruned_loss=0.09435, over 4209.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2545, pruned_loss=0.05833, over 953988.04 frames. ], batch size: 65, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:52:47,695 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95105.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 20:53:06,954 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-26 20:53:13,098 INFO [finetune.py:976] (4/7) Epoch 17, batch 3500, loss[loss=0.1689, simple_loss=0.2433, pruned_loss=0.04723, over 4767.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2518, pruned_loss=0.05712, over 954010.75 frames. ], batch size: 28, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:53:17,182 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.619e+02 1.940e+02 2.279e+02 3.817e+02, threshold=3.880e+02, percent-clipped=1.0 +2023-03-26 20:53:35,200 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95166.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 20:53:54,797 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1202, 1.8251, 1.6273, 1.6278, 1.7239, 1.7223, 1.7560, 2.5157], + device='cuda:4'), covar=tensor([0.3407, 0.3949, 0.3192, 0.3693, 0.3965, 0.2444, 0.3560, 0.1583], + device='cuda:4'), in_proj_covar=tensor([0.0286, 0.0261, 0.0227, 0.0276, 0.0251, 0.0219, 0.0251, 0.0231], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:54:05,084 INFO [finetune.py:976] (4/7) Epoch 17, batch 3550, loss[loss=0.1478, simple_loss=0.2181, pruned_loss=0.03878, over 4830.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2491, pruned_loss=0.05653, over 954010.78 frames. ], batch size: 39, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:54:23,146 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 20:54:48,078 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1387, 1.9466, 1.3713, 0.5344, 1.6843, 1.7409, 1.5578, 1.8206], + device='cuda:4'), covar=tensor([0.0945, 0.0814, 0.1552, 0.2052, 0.1466, 0.2364, 0.2463, 0.0822], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0196, 0.0199, 0.0183, 0.0213, 0.0207, 0.0224, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:54:51,061 INFO [finetune.py:976] (4/7) Epoch 17, batch 3600, loss[loss=0.1707, simple_loss=0.2383, pruned_loss=0.05154, over 4255.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2467, pruned_loss=0.0554, over 954769.64 frames. ], batch size: 18, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:54:54,641 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.580e+02 1.871e+02 2.182e+02 4.206e+02, threshold=3.742e+02, percent-clipped=1.0 +2023-03-26 20:55:01,988 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95260.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:55:06,906 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95268.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:55:24,757 INFO [finetune.py:976] (4/7) Epoch 17, batch 3650, loss[loss=0.1994, simple_loss=0.2762, pruned_loss=0.06128, over 4817.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.249, pruned_loss=0.05594, over 955586.75 frames. ], batch size: 39, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:55:42,961 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95321.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:55:48,240 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95329.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:55:55,792 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7296, 1.7214, 2.1447, 3.2563, 2.3176, 2.3350, 1.0588, 2.7284], + device='cuda:4'), covar=tensor([0.1756, 0.1305, 0.1283, 0.0541, 0.0779, 0.1306, 0.1830, 0.0554], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0116, 0.0134, 0.0166, 0.0101, 0.0136, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 20:55:55,829 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95339.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:55:58,602 INFO [finetune.py:976] (4/7) Epoch 17, batch 3700, loss[loss=0.204, simple_loss=0.28, pruned_loss=0.06397, over 4817.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.252, pruned_loss=0.05629, over 956491.10 frames. ], batch size: 40, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:56:02,229 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.605e+02 1.907e+02 2.409e+02 3.957e+02, threshold=3.813e+02, percent-clipped=4.0 +2023-03-26 20:56:31,738 INFO [finetune.py:976] (4/7) Epoch 17, batch 3750, loss[loss=0.218, simple_loss=0.2745, pruned_loss=0.08069, over 4929.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2532, pruned_loss=0.05665, over 956141.44 frames. ], batch size: 33, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:56:36,729 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95400.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:57:04,529 INFO [finetune.py:976] (4/7) Epoch 17, batch 3800, loss[loss=0.1275, simple_loss=0.207, pruned_loss=0.024, over 4753.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2537, pruned_loss=0.05616, over 956557.71 frames. ], batch size: 27, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:57:09,546 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.108e+02 1.488e+02 1.737e+02 2.235e+02 4.648e+02, threshold=3.475e+02, percent-clipped=3.0 +2023-03-26 20:57:16,901 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95461.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 20:57:19,830 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1181, 0.9855, 1.0327, 0.4712, 0.9182, 1.1594, 1.1596, 1.0205], + device='cuda:4'), covar=tensor([0.0899, 0.0646, 0.0530, 0.0497, 0.0554, 0.0585, 0.0389, 0.0692], + device='cuda:4'), in_proj_covar=tensor([0.0126, 0.0152, 0.0124, 0.0128, 0.0132, 0.0130, 0.0144, 0.0149], + device='cuda:4'), out_proj_covar=tensor([9.2615e-05, 1.1056e-04, 8.9028e-05, 9.1085e-05, 9.2945e-05, 9.3265e-05, + 1.0405e-04, 1.0737e-04], device='cuda:4') +2023-03-26 20:57:37,562 INFO [finetune.py:976] (4/7) Epoch 17, batch 3850, loss[loss=0.1499, simple_loss=0.2258, pruned_loss=0.037, over 4773.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.2523, pruned_loss=0.05565, over 955847.62 frames. ], batch size: 26, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:58:10,766 INFO [finetune.py:976] (4/7) Epoch 17, batch 3900, loss[loss=0.1723, simple_loss=0.241, pruned_loss=0.05178, over 4810.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.251, pruned_loss=0.05567, over 957704.98 frames. ], batch size: 51, lr: 3.38e-03, grad_scale: 64.0 +2023-03-26 20:58:15,377 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.225e+02 1.550e+02 1.834e+02 2.229e+02 4.290e+02, threshold=3.669e+02, percent-clipped=3.0 +2023-03-26 20:58:15,476 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5590, 1.5590, 2.0078, 2.9161, 2.0816, 2.2323, 0.9687, 2.4745], + device='cuda:4'), covar=tensor([0.1588, 0.1289, 0.1054, 0.0599, 0.0756, 0.1277, 0.1612, 0.0543], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0165, 0.0101, 0.0135, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 20:58:23,745 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1522, 1.8687, 1.7686, 1.6705, 1.8433, 1.8866, 1.8300, 2.5893], + device='cuda:4'), covar=tensor([0.3626, 0.4147, 0.2994, 0.3833, 0.3947, 0.2335, 0.3574, 0.1663], + device='cuda:4'), in_proj_covar=tensor([0.0286, 0.0261, 0.0226, 0.0276, 0.0250, 0.0218, 0.0250, 0.0230], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:58:46,333 INFO [finetune.py:976] (4/7) Epoch 17, batch 3950, loss[loss=0.1986, simple_loss=0.2575, pruned_loss=0.06981, over 4821.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2485, pruned_loss=0.05556, over 956611.84 frames. ], batch size: 40, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 20:59:04,371 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95616.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:59:05,015 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95617.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:59:13,754 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95624.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 20:59:32,284 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1019, 2.0711, 1.7237, 2.1179, 2.0186, 1.7601, 2.4203, 2.1579], + device='cuda:4'), covar=tensor([0.1296, 0.2158, 0.2857, 0.2674, 0.2486, 0.1679, 0.3364, 0.1556], + device='cuda:4'), in_proj_covar=tensor([0.0185, 0.0189, 0.0236, 0.0255, 0.0247, 0.0204, 0.0214, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 20:59:34,015 INFO [finetune.py:976] (4/7) Epoch 17, batch 4000, loss[loss=0.1677, simple_loss=0.2441, pruned_loss=0.0457, over 4810.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2473, pruned_loss=0.05523, over 954370.97 frames. ], batch size: 51, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 20:59:36,548 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6343, 1.5535, 1.4957, 1.5659, 1.2981, 3.3171, 1.3955, 1.8472], + device='cuda:4'), covar=tensor([0.3051, 0.2357, 0.2020, 0.2185, 0.1528, 0.0218, 0.2442, 0.1217], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0114, 0.0119, 0.0121, 0.0112, 0.0095, 0.0095, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 20:59:42,363 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.075e+02 1.540e+02 1.979e+02 2.285e+02 3.877e+02, threshold=3.958e+02, percent-clipped=2.0 +2023-03-26 21:00:12,648 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95678.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:00:26,083 INFO [finetune.py:976] (4/7) Epoch 17, batch 4050, loss[loss=0.2057, simple_loss=0.2742, pruned_loss=0.06862, over 4784.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2495, pruned_loss=0.05551, over 955022.54 frames. ], batch size: 29, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 21:00:27,351 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95695.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:00:59,870 INFO [finetune.py:976] (4/7) Epoch 17, batch 4100, loss[loss=0.2098, simple_loss=0.2723, pruned_loss=0.07362, over 4061.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.2531, pruned_loss=0.05695, over 953336.36 frames. ], batch size: 65, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 21:01:04,070 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.600e+02 1.864e+02 2.304e+02 4.240e+02, threshold=3.729e+02, percent-clipped=2.0 +2023-03-26 21:01:12,324 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95761.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 21:01:33,031 INFO [finetune.py:976] (4/7) Epoch 17, batch 4150, loss[loss=0.1947, simple_loss=0.2595, pruned_loss=0.06497, over 4881.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2553, pruned_loss=0.05762, over 955505.19 frames. ], batch size: 32, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 21:01:44,397 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=95809.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 21:01:45,000 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8374, 3.5599, 3.3706, 1.7043, 3.7132, 2.8044, 0.9321, 2.5034], + device='cuda:4'), covar=tensor([0.2339, 0.1783, 0.1669, 0.3063, 0.1045, 0.0933, 0.4046, 0.1332], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0176, 0.0160, 0.0129, 0.0159, 0.0124, 0.0147, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 21:01:46,239 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95812.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 21:01:47,893 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2304, 2.8797, 2.7366, 1.2277, 2.9880, 2.1774, 0.8450, 1.8665], + device='cuda:4'), covar=tensor([0.2430, 0.1968, 0.1827, 0.3326, 0.1326, 0.1098, 0.3696, 0.1490], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0176, 0.0160, 0.0129, 0.0159, 0.0124, 0.0147, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 21:01:54,840 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4614, 2.3898, 1.8676, 2.5011, 2.3956, 2.0076, 2.9042, 2.4292], + device='cuda:4'), covar=tensor([0.1369, 0.2385, 0.3160, 0.2879, 0.2818, 0.1808, 0.3240, 0.1803], + device='cuda:4'), in_proj_covar=tensor([0.0187, 0.0191, 0.0238, 0.0257, 0.0248, 0.0205, 0.0216, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:02:06,759 INFO [finetune.py:976] (4/7) Epoch 17, batch 4200, loss[loss=0.1699, simple_loss=0.2382, pruned_loss=0.05078, over 4751.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2553, pruned_loss=0.05743, over 953075.19 frames. ], batch size: 26, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 21:02:09,311 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95847.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:02:11,508 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.143e+02 1.538e+02 1.932e+02 2.354e+02 8.206e+02, threshold=3.863e+02, percent-clipped=2.0 +2023-03-26 21:02:21,007 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5285, 1.1212, 0.6981, 1.3794, 1.9683, 0.7227, 1.2850, 1.4502], + device='cuda:4'), covar=tensor([0.1478, 0.2135, 0.1744, 0.1195, 0.1896, 0.1996, 0.1543, 0.1843], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0092, 0.0118, 0.0094, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 21:02:27,885 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95873.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 21:02:32,731 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2043, 1.5723, 0.8029, 2.0571, 2.5329, 1.9368, 1.9187, 2.0088], + device='cuda:4'), covar=tensor([0.1337, 0.1979, 0.1960, 0.1098, 0.1651, 0.1780, 0.1260, 0.1789], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0092, 0.0118, 0.0094, 0.0097, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 21:02:35,371 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-26 21:02:39,921 INFO [finetune.py:976] (4/7) Epoch 17, batch 4250, loss[loss=0.1788, simple_loss=0.2531, pruned_loss=0.05228, over 4928.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2526, pruned_loss=0.05651, over 954680.16 frames. ], batch size: 38, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 21:02:50,097 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95908.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 21:02:55,837 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95916.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:02:57,077 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4216, 1.7496, 1.3625, 1.5487, 1.9191, 1.9785, 1.6934, 1.7814], + device='cuda:4'), covar=tensor([0.0561, 0.0344, 0.0516, 0.0317, 0.0250, 0.0453, 0.0405, 0.0325], + device='cuda:4'), in_proj_covar=tensor([0.0096, 0.0108, 0.0143, 0.0113, 0.0100, 0.0109, 0.0098, 0.0110], + device='cuda:4'), out_proj_covar=tensor([7.4304e-05, 8.3489e-05, 1.1333e-04, 8.6834e-05, 7.7597e-05, 8.0329e-05, + 7.3404e-05, 8.4114e-05], device='cuda:4') +2023-03-26 21:03:02,163 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95924.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:03:11,231 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3758, 1.3474, 1.7385, 2.4705, 1.6548, 2.2476, 0.8581, 2.0973], + device='cuda:4'), covar=tensor([0.1685, 0.1374, 0.1045, 0.0719, 0.0908, 0.1249, 0.1548, 0.0578], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0115, 0.0132, 0.0164, 0.0100, 0.0134, 0.0123, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 21:03:13,538 INFO [finetune.py:976] (4/7) Epoch 17, batch 4300, loss[loss=0.1795, simple_loss=0.2456, pruned_loss=0.05669, over 4866.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2509, pruned_loss=0.05641, over 954713.64 frames. ], batch size: 31, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 21:03:18,258 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.024e+02 1.465e+02 1.654e+02 2.123e+02 3.225e+02, threshold=3.308e+02, percent-clipped=0.0 +2023-03-26 21:03:27,772 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=95964.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:03:33,599 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=95972.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:03:34,209 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95973.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:03:39,495 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95980.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:03:47,257 INFO [finetune.py:976] (4/7) Epoch 17, batch 4350, loss[loss=0.1915, simple_loss=0.2585, pruned_loss=0.06223, over 4860.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2479, pruned_loss=0.05545, over 955251.76 frames. ], batch size: 31, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 21:03:48,533 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95995.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:03:50,988 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9915, 1.8475, 2.4321, 3.9132, 2.7503, 2.6972, 0.9081, 3.2198], + device='cuda:4'), covar=tensor([0.1659, 0.1386, 0.1344, 0.0502, 0.0744, 0.1575, 0.1946, 0.0399], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0115, 0.0132, 0.0164, 0.0101, 0.0135, 0.0123, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 21:04:20,821 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96041.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:04:21,942 INFO [finetune.py:976] (4/7) Epoch 17, batch 4400, loss[loss=0.2143, simple_loss=0.2915, pruned_loss=0.06851, over 4870.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2486, pruned_loss=0.0556, over 954832.66 frames. ], batch size: 34, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 21:04:22,006 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96043.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:04:28,704 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.398e+01 1.490e+02 1.749e+02 2.200e+02 3.209e+02, threshold=3.497e+02, percent-clipped=0.0 +2023-03-26 21:04:59,739 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96080.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:05:13,437 INFO [finetune.py:976] (4/7) Epoch 17, batch 4450, loss[loss=0.1672, simple_loss=0.2486, pruned_loss=0.04289, over 4909.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2511, pruned_loss=0.05616, over 952315.72 frames. ], batch size: 37, lr: 3.38e-03, grad_scale: 32.0 +2023-03-26 21:05:42,826 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-03-26 21:05:43,028 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.62 vs. limit=5.0 +2023-03-26 21:05:58,470 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96141.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:05:59,553 INFO [finetune.py:976] (4/7) Epoch 17, batch 4500, loss[loss=0.2148, simple_loss=0.278, pruned_loss=0.07581, over 4826.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2542, pruned_loss=0.05771, over 952143.46 frames. ], batch size: 47, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:06:03,842 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.121e+02 1.724e+02 1.946e+02 2.358e+02 4.504e+02, threshold=3.891e+02, percent-clipped=3.0 +2023-03-26 21:06:10,591 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4096, 1.3434, 1.2441, 1.4770, 1.6830, 1.5814, 1.3566, 1.2368], + device='cuda:4'), covar=tensor([0.0302, 0.0320, 0.0575, 0.0307, 0.0184, 0.0526, 0.0271, 0.0403], + device='cuda:4'), in_proj_covar=tensor([0.0096, 0.0109, 0.0144, 0.0113, 0.0100, 0.0109, 0.0099, 0.0111], + device='cuda:4'), out_proj_covar=tensor([7.5064e-05, 8.4016e-05, 1.1385e-04, 8.7003e-05, 7.8139e-05, 8.0761e-05, + 7.3840e-05, 8.4445e-05], device='cuda:4') +2023-03-26 21:06:13,985 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96165.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:06:15,776 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96168.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 21:06:33,241 INFO [finetune.py:976] (4/7) Epoch 17, batch 4550, loss[loss=0.1553, simple_loss=0.2317, pruned_loss=0.03943, over 4768.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.2557, pruned_loss=0.05827, over 953105.51 frames. ], batch size: 28, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:06:34,516 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9386, 1.3238, 0.8780, 1.7530, 2.2420, 1.7256, 1.5811, 1.7805], + device='cuda:4'), covar=tensor([0.1557, 0.2175, 0.2046, 0.1293, 0.1952, 0.2029, 0.1503, 0.2045], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0096, 0.0111, 0.0093, 0.0119, 0.0094, 0.0098, 0.0090], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 21:06:39,487 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96203.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 21:06:45,461 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2208, 3.6769, 3.8576, 4.0399, 4.0003, 3.7461, 4.2829, 1.5261], + device='cuda:4'), covar=tensor([0.0750, 0.0828, 0.0791, 0.0887, 0.1124, 0.1431, 0.0684, 0.4969], + device='cuda:4'), in_proj_covar=tensor([0.0349, 0.0244, 0.0275, 0.0290, 0.0333, 0.0281, 0.0300, 0.0293], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:06:54,571 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96226.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:07:07,174 INFO [finetune.py:976] (4/7) Epoch 17, batch 4600, loss[loss=0.1719, simple_loss=0.2471, pruned_loss=0.04831, over 4864.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2544, pruned_loss=0.05703, over 953288.37 frames. ], batch size: 34, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:07:11,422 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.099e+02 1.530e+02 1.886e+02 2.340e+02 4.335e+02, threshold=3.772e+02, percent-clipped=2.0 +2023-03-26 21:07:13,750 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 21:07:17,068 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2710, 2.2205, 1.8363, 2.3055, 2.1471, 2.0556, 2.1158, 3.0677], + device='cuda:4'), covar=tensor([0.3776, 0.4891, 0.3459, 0.4443, 0.4429, 0.2450, 0.4454, 0.1641], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0260, 0.0226, 0.0276, 0.0251, 0.0219, 0.0251, 0.0230], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:07:23,743 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 21:07:26,488 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96273.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:07:40,046 INFO [finetune.py:976] (4/7) Epoch 17, batch 4650, loss[loss=0.1497, simple_loss=0.2111, pruned_loss=0.04411, over 4807.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2512, pruned_loss=0.05608, over 953456.91 frames. ], batch size: 25, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:07:57,985 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 21:07:58,451 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96321.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:08:00,254 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96324.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:08:08,004 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96336.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:08:12,363 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 21:08:13,199 INFO [finetune.py:976] (4/7) Epoch 17, batch 4700, loss[loss=0.2013, simple_loss=0.2508, pruned_loss=0.07593, over 4824.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2477, pruned_loss=0.05477, over 955318.13 frames. ], batch size: 30, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:08:18,315 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.127e+02 1.531e+02 1.882e+02 2.216e+02 4.319e+02, threshold=3.764e+02, percent-clipped=2.0 +2023-03-26 21:08:36,494 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 21:08:40,448 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96385.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:08:43,956 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96390.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:08:45,653 INFO [finetune.py:976] (4/7) Epoch 17, batch 4750, loss[loss=0.1709, simple_loss=0.2429, pruned_loss=0.0495, over 4912.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2477, pruned_loss=0.05585, over 954238.53 frames. ], batch size: 36, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:09:14,755 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96436.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:09:19,422 INFO [finetune.py:976] (4/7) Epoch 17, batch 4800, loss[loss=0.1728, simple_loss=0.2376, pruned_loss=0.05398, over 4699.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2508, pruned_loss=0.05745, over 953130.93 frames. ], batch size: 23, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:09:25,020 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.112e+02 1.609e+02 1.875e+02 2.422e+02 6.864e+02, threshold=3.750e+02, percent-clipped=2.0 +2023-03-26 21:09:25,774 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96451.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:09:36,646 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96468.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 21:09:54,759 INFO [finetune.py:976] (4/7) Epoch 17, batch 4850, loss[loss=0.1549, simple_loss=0.2162, pruned_loss=0.04684, over 4239.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2516, pruned_loss=0.05724, over 950618.64 frames. ], batch size: 18, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:09:56,546 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6719, 1.3643, 1.0138, 1.6081, 2.0469, 1.5285, 1.6439, 1.7068], + device='cuda:4'), covar=tensor([0.1647, 0.2113, 0.1867, 0.1281, 0.2035, 0.2019, 0.1427, 0.1986], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0095, 0.0110, 0.0092, 0.0119, 0.0094, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 21:10:02,928 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96503.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:10:07,291 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-03-26 21:10:15,706 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96516.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 21:10:18,750 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96521.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:10:38,686 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3411, 2.2534, 1.9951, 1.1964, 2.0921, 1.8916, 1.7924, 2.1185], + device='cuda:4'), covar=tensor([0.0964, 0.0665, 0.1406, 0.1738, 0.1346, 0.1828, 0.1817, 0.0846], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0197, 0.0201, 0.0185, 0.0215, 0.0209, 0.0225, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:10:45,280 INFO [finetune.py:976] (4/7) Epoch 17, batch 4900, loss[loss=0.2203, simple_loss=0.2844, pruned_loss=0.07812, over 4746.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2542, pruned_loss=0.05787, over 951847.68 frames. ], batch size: 27, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:10:54,622 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.592e+02 1.896e+02 2.164e+02 3.347e+02, threshold=3.792e+02, percent-clipped=0.0 +2023-03-26 21:10:55,805 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96551.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:11:26,123 INFO [finetune.py:976] (4/7) Epoch 17, batch 4950, loss[loss=0.1666, simple_loss=0.2523, pruned_loss=0.04048, over 4914.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2544, pruned_loss=0.05736, over 951181.69 frames. ], batch size: 36, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:11:39,658 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6772, 0.7309, 1.7188, 1.6094, 1.5306, 1.4544, 1.5340, 1.6142], + device='cuda:4'), covar=tensor([0.3472, 0.3645, 0.3161, 0.3117, 0.4223, 0.3440, 0.3691, 0.3002], + device='cuda:4'), in_proj_covar=tensor([0.0248, 0.0241, 0.0259, 0.0274, 0.0273, 0.0248, 0.0282, 0.0240], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:11:48,149 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6932, 0.7452, 1.7280, 1.6217, 1.5299, 1.4682, 1.5658, 1.6489], + device='cuda:4'), covar=tensor([0.3436, 0.3613, 0.3128, 0.3399, 0.4109, 0.3416, 0.3974, 0.3053], + device='cuda:4'), in_proj_covar=tensor([0.0248, 0.0240, 0.0259, 0.0273, 0.0272, 0.0247, 0.0282, 0.0240], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:11:55,635 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96636.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:11:59,764 INFO [finetune.py:976] (4/7) Epoch 17, batch 5000, loss[loss=0.1624, simple_loss=0.2298, pruned_loss=0.04746, over 4789.00 frames. ], tot_loss[loss=0.183, simple_loss=0.253, pruned_loss=0.05656, over 953090.91 frames. ], batch size: 25, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:12:04,410 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.531e+02 1.819e+02 2.156e+02 3.437e+02, threshold=3.638e+02, percent-clipped=0.0 +2023-03-26 21:12:24,602 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96680.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:12:26,962 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96684.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:12:33,216 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.58 vs. limit=5.0 +2023-03-26 21:12:33,328 INFO [finetune.py:976] (4/7) Epoch 17, batch 5050, loss[loss=0.176, simple_loss=0.2344, pruned_loss=0.05885, over 4865.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.251, pruned_loss=0.05627, over 953762.47 frames. ], batch size: 49, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:13:01,829 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96736.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:13:06,446 INFO [finetune.py:976] (4/7) Epoch 17, batch 5100, loss[loss=0.1594, simple_loss=0.2279, pruned_loss=0.04551, over 4922.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2483, pruned_loss=0.05574, over 954273.10 frames. ], batch size: 33, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:13:08,318 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96746.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:13:10,591 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.513e+02 1.848e+02 2.246e+02 3.685e+02, threshold=3.695e+02, percent-clipped=1.0 +2023-03-26 21:13:28,940 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96776.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:13:33,647 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96784.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:13:39,092 INFO [finetune.py:976] (4/7) Epoch 17, batch 5150, loss[loss=0.1525, simple_loss=0.2164, pruned_loss=0.04428, over 4780.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2479, pruned_loss=0.05577, over 954332.00 frames. ], batch size: 28, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:13:47,020 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8747, 1.8102, 1.6842, 1.7858, 1.3225, 4.4399, 1.5879, 2.0224], + device='cuda:4'), covar=tensor([0.3085, 0.2366, 0.2044, 0.2155, 0.1595, 0.0112, 0.2434, 0.1196], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0114, 0.0119, 0.0122, 0.0113, 0.0095, 0.0095, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 21:13:58,206 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96821.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:14:08,361 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96837.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:14:11,883 INFO [finetune.py:976] (4/7) Epoch 17, batch 5200, loss[loss=0.1844, simple_loss=0.2555, pruned_loss=0.05663, over 4863.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.251, pruned_loss=0.05684, over 951572.79 frames. ], batch size: 31, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:14:16,595 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.064e+02 1.635e+02 1.977e+02 2.300e+02 5.939e+02, threshold=3.955e+02, percent-clipped=5.0 +2023-03-26 21:14:17,082 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.35 vs. limit=5.0 +2023-03-26 21:14:29,996 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=96869.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:14:32,939 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1909, 2.1833, 1.8122, 2.3393, 2.1219, 2.0401, 2.0328, 2.9320], + device='cuda:4'), covar=tensor([0.3744, 0.4664, 0.3425, 0.4228, 0.4828, 0.2514, 0.4835, 0.1624], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0260, 0.0225, 0.0274, 0.0251, 0.0219, 0.0250, 0.0229], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:14:36,008 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 21:14:37,233 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 21:14:44,941 INFO [finetune.py:976] (4/7) Epoch 17, batch 5250, loss[loss=0.2277, simple_loss=0.2929, pruned_loss=0.08122, over 4816.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2539, pruned_loss=0.05737, over 952735.17 frames. ], batch size: 33, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:15:21,073 INFO [finetune.py:976] (4/7) Epoch 17, batch 5300, loss[loss=0.1645, simple_loss=0.2211, pruned_loss=0.05391, over 4068.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2559, pruned_loss=0.05764, over 953184.85 frames. ], batch size: 17, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:15:30,013 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.644e+02 1.959e+02 2.379e+02 3.599e+02, threshold=3.918e+02, percent-clipped=0.0 +2023-03-26 21:16:00,343 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96980.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:16:17,642 INFO [finetune.py:976] (4/7) Epoch 17, batch 5350, loss[loss=0.1439, simple_loss=0.2191, pruned_loss=0.03434, over 4723.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2549, pruned_loss=0.05694, over 952633.53 frames. ], batch size: 23, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:16:25,285 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97000.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:16:44,613 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97028.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:16:54,583 INFO [finetune.py:976] (4/7) Epoch 17, batch 5400, loss[loss=0.1276, simple_loss=0.2022, pruned_loss=0.02646, over 4795.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2517, pruned_loss=0.05575, over 953237.96 frames. ], batch size: 29, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:16:56,508 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97046.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:16:58,805 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.591e+02 1.860e+02 2.348e+02 4.043e+02, threshold=3.721e+02, percent-clipped=1.0 +2023-03-26 21:17:06,284 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97061.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:17:13,980 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.61 vs. limit=2.0 +2023-03-26 21:17:22,803 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.7166, 3.1641, 2.9586, 1.4966, 3.1276, 2.6351, 2.5092, 2.7787], + device='cuda:4'), covar=tensor([0.0805, 0.0829, 0.1730, 0.2022, 0.1272, 0.1777, 0.1714, 0.1109], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0196, 0.0201, 0.0184, 0.0215, 0.0209, 0.0223, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:17:27,398 INFO [finetune.py:976] (4/7) Epoch 17, batch 5450, loss[loss=0.1241, simple_loss=0.1986, pruned_loss=0.02484, over 4828.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2488, pruned_loss=0.05489, over 955054.59 frames. ], batch size: 25, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:17:28,070 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97094.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:17:44,503 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 21:17:44,787 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6645, 1.1907, 0.8339, 1.5124, 2.2340, 1.0908, 1.4110, 1.5621], + device='cuda:4'), covar=tensor([0.1530, 0.2208, 0.1809, 0.1257, 0.1793, 0.1965, 0.1559, 0.1954], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0095, 0.0110, 0.0092, 0.0119, 0.0094, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 21:17:52,427 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97132.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:18:00,549 INFO [finetune.py:976] (4/7) Epoch 17, batch 5500, loss[loss=0.1469, simple_loss=0.2149, pruned_loss=0.03946, over 4779.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.246, pruned_loss=0.0542, over 955638.20 frames. ], batch size: 26, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:18:04,741 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.221e+01 1.533e+02 1.769e+02 2.042e+02 3.202e+02, threshold=3.539e+02, percent-clipped=0.0 +2023-03-26 21:18:06,130 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3789, 1.4853, 1.9178, 1.5922, 1.5062, 3.2683, 1.2363, 1.4808], + device='cuda:4'), covar=tensor([0.0962, 0.1627, 0.1153, 0.0934, 0.1525, 0.0239, 0.1492, 0.1633], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0084, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 21:18:33,719 INFO [finetune.py:976] (4/7) Epoch 17, batch 5550, loss[loss=0.1679, simple_loss=0.2424, pruned_loss=0.04673, over 4800.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2481, pruned_loss=0.05552, over 956170.58 frames. ], batch size: 29, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:19:05,113 INFO [finetune.py:976] (4/7) Epoch 17, batch 5600, loss[loss=0.1844, simple_loss=0.2748, pruned_loss=0.04699, over 4746.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2508, pruned_loss=0.05611, over 954981.04 frames. ], batch size: 54, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:19:09,080 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.641e+02 1.914e+02 2.406e+02 4.422e+02, threshold=3.827e+02, percent-clipped=1.0 +2023-03-26 21:19:14,890 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97260.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:19:16,614 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4585, 1.5053, 1.8683, 1.1556, 1.6739, 1.7818, 1.3818, 1.9169], + device='cuda:4'), covar=tensor([0.1382, 0.2046, 0.1280, 0.1810, 0.0895, 0.1421, 0.2933, 0.0833], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0202, 0.0188, 0.0189, 0.0175, 0.0212, 0.0215, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:19:20,392 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.57 vs. limit=5.0 +2023-03-26 21:19:34,291 INFO [finetune.py:976] (4/7) Epoch 17, batch 5650, loss[loss=0.1688, simple_loss=0.2586, pruned_loss=0.03944, over 4819.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2551, pruned_loss=0.05711, over 954593.89 frames. ], batch size: 40, lr: 3.37e-03, grad_scale: 32.0 +2023-03-26 21:19:51,444 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97321.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:19:53,408 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.75 vs. limit=5.0 +2023-03-26 21:20:04,082 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97342.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:20:04,602 INFO [finetune.py:976] (4/7) Epoch 17, batch 5700, loss[loss=0.1736, simple_loss=0.233, pruned_loss=0.05706, over 4064.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2512, pruned_loss=0.05682, over 932555.38 frames. ], batch size: 17, lr: 3.36e-03, grad_scale: 32.0 +2023-03-26 21:20:07,046 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97347.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:20:08,736 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.528e+01 1.533e+02 1.739e+02 2.212e+02 3.283e+02, threshold=3.478e+02, percent-clipped=0.0 +2023-03-26 21:20:12,308 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97356.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:20:35,933 INFO [finetune.py:976] (4/7) Epoch 18, batch 0, loss[loss=0.2065, simple_loss=0.2825, pruned_loss=0.06525, over 4813.00 frames. ], tot_loss[loss=0.2065, simple_loss=0.2825, pruned_loss=0.06525, over 4813.00 frames. ], batch size: 39, lr: 3.36e-03, grad_scale: 32.0 +2023-03-26 21:20:35,933 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 21:20:46,783 INFO [finetune.py:1010] (4/7) Epoch 18, validation: loss=0.1584, simple_loss=0.2281, pruned_loss=0.0444, over 2265189.00 frames. +2023-03-26 21:20:46,784 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 21:20:49,145 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0725, 1.8318, 2.5501, 4.0394, 2.8053, 2.8061, 0.9986, 3.3940], + device='cuda:4'), covar=tensor([0.1677, 0.1452, 0.1292, 0.0433, 0.0758, 0.1295, 0.1952, 0.0407], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0134, 0.0164, 0.0100, 0.0136, 0.0123, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 21:20:49,174 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97374.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:21:26,091 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97403.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 21:21:34,110 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97408.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:21:39,449 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7268, 1.6000, 2.0796, 3.2625, 2.1818, 2.3711, 1.1754, 2.6582], + device='cuda:4'), covar=tensor([0.1543, 0.1316, 0.1257, 0.0508, 0.0747, 0.1560, 0.1581, 0.0492], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0134, 0.0164, 0.0100, 0.0136, 0.0123, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 21:21:47,681 INFO [finetune.py:976] (4/7) Epoch 18, batch 50, loss[loss=0.1737, simple_loss=0.235, pruned_loss=0.05618, over 4823.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.257, pruned_loss=0.05933, over 216732.90 frames. ], batch size: 33, lr: 3.36e-03, grad_scale: 32.0 +2023-03-26 21:21:58,962 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97432.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:22:00,769 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97435.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:22:09,806 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.163e+01 1.563e+02 1.902e+02 2.308e+02 3.615e+02, threshold=3.804e+02, percent-clipped=1.0 +2023-03-26 21:22:25,131 INFO [finetune.py:976] (4/7) Epoch 18, batch 100, loss[loss=0.1398, simple_loss=0.2182, pruned_loss=0.03074, over 4794.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2471, pruned_loss=0.05492, over 381435.46 frames. ], batch size: 29, lr: 3.36e-03, grad_scale: 32.0 +2023-03-26 21:22:31,665 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97480.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:22:54,254 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9601, 2.9864, 2.7694, 2.0983, 3.0205, 3.2725, 3.2902, 2.6855], + device='cuda:4'), covar=tensor([0.0561, 0.0549, 0.0679, 0.0842, 0.0568, 0.0567, 0.0541, 0.0881], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0134, 0.0139, 0.0122, 0.0123, 0.0138, 0.0140, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:22:58,723 INFO [finetune.py:976] (4/7) Epoch 18, batch 150, loss[loss=0.1529, simple_loss=0.2242, pruned_loss=0.0408, over 4759.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2448, pruned_loss=0.0551, over 509547.47 frames. ], batch size: 27, lr: 3.36e-03, grad_scale: 32.0 +2023-03-26 21:23:17,209 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.559e+02 1.792e+02 2.227e+02 6.409e+02, threshold=3.584e+02, percent-clipped=2.0 +2023-03-26 21:23:19,730 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4547, 1.6529, 0.7806, 2.3831, 2.7153, 1.9840, 2.0185, 2.3331], + device='cuda:4'), covar=tensor([0.1305, 0.1963, 0.2152, 0.1037, 0.1589, 0.1709, 0.1316, 0.1793], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0092, 0.0118, 0.0094, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 21:23:20,356 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97555.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:23:26,130 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7266, 0.9919, 1.6947, 1.6705, 1.4977, 1.4792, 1.5919, 1.6296], + device='cuda:4'), covar=tensor([0.4637, 0.4165, 0.3890, 0.3974, 0.5284, 0.4120, 0.4618, 0.3829], + device='cuda:4'), in_proj_covar=tensor([0.0249, 0.0240, 0.0259, 0.0274, 0.0273, 0.0248, 0.0282, 0.0240], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:23:32,346 INFO [finetune.py:976] (4/7) Epoch 18, batch 200, loss[loss=0.1579, simple_loss=0.2293, pruned_loss=0.0432, over 4852.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2441, pruned_loss=0.05479, over 610116.20 frames. ], batch size: 49, lr: 3.36e-03, grad_scale: 32.0 +2023-03-26 21:24:01,854 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97616.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:24:01,912 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97616.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:24:04,235 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9216, 1.3629, 1.9347, 1.9625, 1.6947, 1.6660, 1.8787, 1.7938], + device='cuda:4'), covar=tensor([0.3859, 0.4175, 0.3489, 0.3796, 0.5199, 0.3883, 0.4430, 0.3186], + device='cuda:4'), in_proj_covar=tensor([0.0248, 0.0239, 0.0259, 0.0273, 0.0272, 0.0247, 0.0282, 0.0240], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:24:05,309 INFO [finetune.py:976] (4/7) Epoch 18, batch 250, loss[loss=0.164, simple_loss=0.2341, pruned_loss=0.04693, over 4935.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2469, pruned_loss=0.05516, over 688327.47 frames. ], batch size: 33, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:24:24,122 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.114e+02 1.615e+02 1.960e+02 2.417e+02 4.168e+02, threshold=3.921e+02, percent-clipped=3.0 +2023-03-26 21:24:27,893 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97656.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:24:37,890 INFO [finetune.py:976] (4/7) Epoch 18, batch 300, loss[loss=0.1856, simple_loss=0.2775, pruned_loss=0.04678, over 4812.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2509, pruned_loss=0.05598, over 749908.40 frames. ], batch size: 45, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:24:40,301 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97674.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 21:24:56,233 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97698.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 21:24:59,263 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97703.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:24:59,859 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97704.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:25:10,594 INFO [finetune.py:976] (4/7) Epoch 18, batch 350, loss[loss=0.1654, simple_loss=0.2471, pruned_loss=0.04182, over 4318.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2525, pruned_loss=0.05643, over 796483.01 frames. ], batch size: 65, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:25:17,553 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97730.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:25:21,078 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97735.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 21:25:28,927 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2616, 2.0665, 1.6600, 2.0429, 2.0106, 1.8138, 2.3978, 2.1975], + device='cuda:4'), covar=tensor([0.1367, 0.2330, 0.3281, 0.2965, 0.2837, 0.1828, 0.3406, 0.1822], + device='cuda:4'), in_proj_covar=tensor([0.0184, 0.0188, 0.0235, 0.0254, 0.0246, 0.0203, 0.0214, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:25:30,598 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.611e+02 1.882e+02 2.387e+02 3.928e+02, threshold=3.763e+02, percent-clipped=1.0 +2023-03-26 21:25:43,300 INFO [finetune.py:976] (4/7) Epoch 18, batch 400, loss[loss=0.1976, simple_loss=0.2692, pruned_loss=0.06306, over 4912.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2528, pruned_loss=0.05638, over 829192.43 frames. ], batch size: 46, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:25:46,125 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5978, 1.1601, 0.8690, 1.5017, 2.0380, 1.0638, 1.3605, 1.5335], + device='cuda:4'), covar=tensor([0.1487, 0.2095, 0.1841, 0.1145, 0.1836, 0.2017, 0.1495, 0.1897], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0095, 0.0110, 0.0092, 0.0119, 0.0094, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 21:25:52,697 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.54 vs. limit=5.0 +2023-03-26 21:26:22,834 INFO [finetune.py:976] (4/7) Epoch 18, batch 450, loss[loss=0.2393, simple_loss=0.2959, pruned_loss=0.09136, over 4797.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2527, pruned_loss=0.05667, over 858247.29 frames. ], batch size: 29, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:26:24,380 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 21:26:42,231 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4422, 2.4340, 2.4082, 1.8808, 2.3498, 2.7258, 2.6424, 2.1213], + device='cuda:4'), covar=tensor([0.0511, 0.0515, 0.0628, 0.0847, 0.0947, 0.0527, 0.0526, 0.0974], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0134, 0.0140, 0.0122, 0.0124, 0.0139, 0.0140, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:27:01,043 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.519e+02 1.772e+02 2.128e+02 3.513e+02, threshold=3.544e+02, percent-clipped=0.0 +2023-03-26 21:27:16,934 INFO [finetune.py:976] (4/7) Epoch 18, batch 500, loss[loss=0.1664, simple_loss=0.2378, pruned_loss=0.0475, over 4809.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2494, pruned_loss=0.05547, over 880356.01 frames. ], batch size: 51, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:27:24,749 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 21:27:29,768 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97888.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:27:32,149 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9892, 1.3093, 1.9954, 1.9927, 1.7981, 1.7936, 1.8388, 1.8836], + device='cuda:4'), covar=tensor([0.3719, 0.4222, 0.3455, 0.3614, 0.4793, 0.3428, 0.4350, 0.3078], + device='cuda:4'), in_proj_covar=tensor([0.0250, 0.0241, 0.0261, 0.0276, 0.0275, 0.0249, 0.0284, 0.0242], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:27:44,610 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97911.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:27:47,698 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97916.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:27:50,635 INFO [finetune.py:976] (4/7) Epoch 18, batch 550, loss[loss=0.1748, simple_loss=0.2481, pruned_loss=0.0507, over 4855.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2475, pruned_loss=0.05505, over 896642.86 frames. ], batch size: 44, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:28:15,835 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 21:28:19,278 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97949.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:28:19,731 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.294e+01 1.530e+02 1.816e+02 2.060e+02 3.951e+02, threshold=3.633e+02, percent-clipped=3.0 +2023-03-26 21:28:28,264 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=97964.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:28:32,534 INFO [finetune.py:976] (4/7) Epoch 18, batch 600, loss[loss=0.2449, simple_loss=0.3069, pruned_loss=0.09148, over 4795.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2497, pruned_loss=0.05643, over 911120.86 frames. ], batch size: 45, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:28:36,363 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 21:28:51,972 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97998.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:28:56,767 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98003.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:29:07,660 INFO [finetune.py:976] (4/7) Epoch 18, batch 650, loss[loss=0.1833, simple_loss=0.2603, pruned_loss=0.05312, over 4852.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2523, pruned_loss=0.05698, over 922494.23 frames. ], batch size: 44, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:29:13,280 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98030.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 21:29:13,314 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98030.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:29:25,507 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98046.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:29:28,301 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.027e+02 1.543e+02 1.878e+02 2.128e+02 3.672e+02, threshold=3.757e+02, percent-clipped=1.0 +2023-03-26 21:29:28,988 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98051.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:29:38,332 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 21:29:41,537 INFO [finetune.py:976] (4/7) Epoch 18, batch 700, loss[loss=0.1474, simple_loss=0.2281, pruned_loss=0.03338, over 4823.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2534, pruned_loss=0.05698, over 932055.04 frames. ], batch size: 38, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:29:45,868 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98078.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:30:03,079 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 21:30:05,704 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98106.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:30:05,885 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-26 21:30:11,029 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7312, 1.2690, 0.7575, 1.5608, 2.0892, 1.0726, 1.4730, 1.5405], + device='cuda:4'), covar=tensor([0.1533, 0.2142, 0.2052, 0.1207, 0.1929, 0.2078, 0.1466, 0.2085], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0096, 0.0111, 0.0093, 0.0120, 0.0095, 0.0099, 0.0090], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 21:30:15,240 INFO [finetune.py:976] (4/7) Epoch 18, batch 750, loss[loss=0.1732, simple_loss=0.2429, pruned_loss=0.05179, over 4805.00 frames. ], tot_loss[loss=0.1831, simple_loss=0.2533, pruned_loss=0.0564, over 937073.26 frames. ], batch size: 40, lr: 3.36e-03, grad_scale: 64.0 +2023-03-26 21:30:15,960 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4329, 1.5274, 1.2230, 1.4635, 1.8104, 1.7440, 1.5432, 1.3796], + device='cuda:4'), covar=tensor([0.0301, 0.0268, 0.0573, 0.0298, 0.0185, 0.0390, 0.0273, 0.0322], + device='cuda:4'), in_proj_covar=tensor([0.0096, 0.0107, 0.0144, 0.0113, 0.0100, 0.0109, 0.0099, 0.0110], + device='cuda:4'), out_proj_covar=tensor([7.4859e-05, 8.2978e-05, 1.1386e-04, 8.6788e-05, 7.8189e-05, 8.0725e-05, + 7.3726e-05, 8.4062e-05], device='cuda:4') +2023-03-26 21:30:34,724 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.501e+02 1.785e+02 2.309e+02 4.193e+02, threshold=3.569e+02, percent-clipped=2.0 +2023-03-26 21:30:46,070 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98167.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:30:48,362 INFO [finetune.py:976] (4/7) Epoch 18, batch 800, loss[loss=0.2057, simple_loss=0.2679, pruned_loss=0.0718, over 4817.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2528, pruned_loss=0.05629, over 939001.37 frames. ], batch size: 33, lr: 3.36e-03, grad_scale: 32.0 +2023-03-26 21:30:48,493 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2457, 1.3706, 1.1710, 1.3951, 1.5971, 1.5564, 1.4312, 1.2797], + device='cuda:4'), covar=tensor([0.0412, 0.0261, 0.0514, 0.0286, 0.0186, 0.0436, 0.0261, 0.0327], + device='cuda:4'), in_proj_covar=tensor([0.0096, 0.0107, 0.0144, 0.0112, 0.0100, 0.0108, 0.0098, 0.0110], + device='cuda:4'), out_proj_covar=tensor([7.4368e-05, 8.2601e-05, 1.1327e-04, 8.6393e-05, 7.7817e-05, 8.0255e-05, + 7.3430e-05, 8.3629e-05], device='cuda:4') +2023-03-26 21:31:15,633 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98211.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:31:22,189 INFO [finetune.py:976] (4/7) Epoch 18, batch 850, loss[loss=0.1649, simple_loss=0.2266, pruned_loss=0.05163, over 4915.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2515, pruned_loss=0.05636, over 942615.42 frames. ], batch size: 37, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:31:22,899 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6418, 1.2289, 0.8642, 1.5748, 2.0325, 1.1066, 1.3575, 1.4707], + device='cuda:4'), covar=tensor([0.1422, 0.1959, 0.1811, 0.1098, 0.1909, 0.1937, 0.1469, 0.1913], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0096, 0.0111, 0.0093, 0.0120, 0.0095, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 21:31:45,837 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98244.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:31:55,531 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.155e+02 1.511e+02 1.794e+02 2.111e+02 3.360e+02, threshold=3.589e+02, percent-clipped=0.0 +2023-03-26 21:32:07,094 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98259.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:32:25,143 INFO [finetune.py:976] (4/7) Epoch 18, batch 900, loss[loss=0.1659, simple_loss=0.2255, pruned_loss=0.05312, over 4718.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2491, pruned_loss=0.05545, over 944880.38 frames. ], batch size: 23, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:32:48,936 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9869, 1.8320, 1.9135, 1.2861, 1.9891, 1.9475, 1.9726, 1.5552], + device='cuda:4'), covar=tensor([0.0527, 0.0622, 0.0658, 0.0905, 0.0714, 0.0664, 0.0613, 0.1206], + device='cuda:4'), in_proj_covar=tensor([0.0136, 0.0137, 0.0143, 0.0124, 0.0126, 0.0142, 0.0143, 0.0165], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:32:50,438 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.99 vs. limit=5.0 +2023-03-26 21:33:02,858 INFO [finetune.py:976] (4/7) Epoch 18, batch 950, loss[loss=0.1503, simple_loss=0.2188, pruned_loss=0.04084, over 4765.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2473, pruned_loss=0.05483, over 947616.55 frames. ], batch size: 28, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:33:08,458 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98330.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 21:33:23,086 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.511e+02 1.760e+02 2.160e+02 3.441e+02, threshold=3.521e+02, percent-clipped=0.0 +2023-03-26 21:33:37,315 INFO [finetune.py:976] (4/7) Epoch 18, batch 1000, loss[loss=0.1356, simple_loss=0.2051, pruned_loss=0.03306, over 4058.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2486, pruned_loss=0.05476, over 950598.01 frames. ], batch size: 17, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:33:42,165 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98378.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 21:34:02,777 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98410.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:34:14,801 INFO [finetune.py:976] (4/7) Epoch 18, batch 1050, loss[loss=0.149, simple_loss=0.221, pruned_loss=0.03853, over 4775.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2508, pruned_loss=0.05494, over 949185.61 frames. ], batch size: 26, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:34:36,631 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.558e+02 1.856e+02 2.287e+02 5.753e+02, threshold=3.713e+02, percent-clipped=4.0 +2023-03-26 21:34:44,672 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98462.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:34:51,515 INFO [finetune.py:976] (4/7) Epoch 18, batch 1100, loss[loss=0.2136, simple_loss=0.2891, pruned_loss=0.06904, over 4923.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2515, pruned_loss=0.05503, over 951038.22 frames. ], batch size: 42, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:34:51,644 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98471.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:35:08,488 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1926, 1.3971, 1.1918, 1.4402, 1.6706, 1.5378, 1.3755, 1.2598], + device='cuda:4'), covar=tensor([0.0393, 0.0283, 0.0605, 0.0269, 0.0212, 0.0441, 0.0325, 0.0388], + device='cuda:4'), in_proj_covar=tensor([0.0097, 0.0108, 0.0144, 0.0113, 0.0101, 0.0109, 0.0099, 0.0111], + device='cuda:4'), out_proj_covar=tensor([7.5148e-05, 8.3451e-05, 1.1378e-04, 8.6666e-05, 7.8786e-05, 8.0849e-05, + 7.4099e-05, 8.4446e-05], device='cuda:4') +2023-03-26 21:35:14,990 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8851, 1.4603, 2.0241, 1.9339, 1.7080, 1.6596, 1.8425, 1.8074], + device='cuda:4'), covar=tensor([0.3962, 0.3945, 0.3071, 0.3597, 0.4516, 0.3628, 0.4300, 0.3038], + device='cuda:4'), in_proj_covar=tensor([0.0248, 0.0240, 0.0259, 0.0274, 0.0273, 0.0248, 0.0283, 0.0241], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:35:24,231 INFO [finetune.py:976] (4/7) Epoch 18, batch 1150, loss[loss=0.1819, simple_loss=0.26, pruned_loss=0.05189, over 4861.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2529, pruned_loss=0.05584, over 950643.13 frames. ], batch size: 34, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:35:39,053 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98543.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:35:39,638 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98544.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:35:43,755 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.197e+02 1.665e+02 1.912e+02 2.323e+02 4.830e+02, threshold=3.825e+02, percent-clipped=3.0 +2023-03-26 21:35:57,263 INFO [finetune.py:976] (4/7) Epoch 18, batch 1200, loss[loss=0.1647, simple_loss=0.2366, pruned_loss=0.04644, over 4797.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.2522, pruned_loss=0.05573, over 950786.38 frames. ], batch size: 29, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:35:59,149 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1421, 1.9685, 1.6628, 1.6691, 1.8473, 1.8703, 1.8635, 2.5310], + device='cuda:4'), covar=tensor([0.3880, 0.4079, 0.3443, 0.3874, 0.3956, 0.2364, 0.3742, 0.1791], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0263, 0.0228, 0.0276, 0.0253, 0.0220, 0.0252, 0.0233], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:36:12,053 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98592.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:36:19,566 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98604.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:36:31,243 INFO [finetune.py:976] (4/7) Epoch 18, batch 1250, loss[loss=0.1553, simple_loss=0.2121, pruned_loss=0.0493, over 4731.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2497, pruned_loss=0.05504, over 951340.16 frames. ], batch size: 59, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:37:01,603 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.074e+02 1.465e+02 1.816e+02 2.333e+02 4.053e+02, threshold=3.632e+02, percent-clipped=1.0 +2023-03-26 21:37:29,819 INFO [finetune.py:976] (4/7) Epoch 18, batch 1300, loss[loss=0.1771, simple_loss=0.2487, pruned_loss=0.05273, over 4916.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2472, pruned_loss=0.05399, over 953751.36 frames. ], batch size: 37, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:37:54,849 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2285, 1.9349, 1.4002, 0.5525, 1.6805, 1.8659, 1.7314, 1.8452], + device='cuda:4'), covar=tensor([0.0953, 0.0849, 0.1662, 0.2025, 0.1471, 0.2103, 0.2120, 0.0953], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0196, 0.0201, 0.0183, 0.0214, 0.0208, 0.0223, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:38:07,458 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0414, 1.9933, 2.0168, 1.5802, 2.0602, 2.1367, 2.1517, 1.7431], + device='cuda:4'), covar=tensor([0.0560, 0.0617, 0.0676, 0.0792, 0.0636, 0.0593, 0.0558, 0.1085], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0135, 0.0141, 0.0122, 0.0125, 0.0140, 0.0141, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:38:11,399 INFO [finetune.py:976] (4/7) Epoch 18, batch 1350, loss[loss=0.2303, simple_loss=0.2916, pruned_loss=0.08449, over 4913.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2475, pruned_loss=0.05441, over 953416.48 frames. ], batch size: 37, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:38:16,738 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5722, 1.4728, 1.4401, 1.4971, 0.9730, 2.9759, 1.0252, 1.5006], + device='cuda:4'), covar=tensor([0.3450, 0.2546, 0.2270, 0.2561, 0.2017, 0.0259, 0.2805, 0.1440], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0115, 0.0121, 0.0122, 0.0113, 0.0096, 0.0096, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 21:38:31,462 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.112e+02 1.676e+02 1.899e+02 2.271e+02 3.691e+02, threshold=3.798e+02, percent-clipped=1.0 +2023-03-26 21:38:38,224 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98762.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:38:41,111 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98766.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:38:41,139 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3792, 1.3765, 1.4701, 1.5821, 1.4830, 2.9175, 1.3181, 1.4713], + device='cuda:4'), covar=tensor([0.0969, 0.1916, 0.1208, 0.0988, 0.1676, 0.0263, 0.1582, 0.1851], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 21:38:44,606 INFO [finetune.py:976] (4/7) Epoch 18, batch 1400, loss[loss=0.2031, simple_loss=0.2745, pruned_loss=0.06586, over 4805.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2512, pruned_loss=0.05563, over 954350.93 frames. ], batch size: 45, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:38:46,493 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-26 21:39:04,084 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7731, 1.5753, 1.9994, 1.3822, 1.9236, 2.0353, 1.5037, 2.1775], + device='cuda:4'), covar=tensor([0.1097, 0.1818, 0.1480, 0.1919, 0.0831, 0.1295, 0.2526, 0.0725], + device='cuda:4'), in_proj_covar=tensor([0.0194, 0.0203, 0.0192, 0.0191, 0.0176, 0.0213, 0.0216, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:39:04,840 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3072, 1.3808, 1.1895, 1.4050, 1.7104, 1.5864, 1.4247, 1.2044], + device='cuda:4'), covar=tensor([0.0391, 0.0294, 0.0668, 0.0317, 0.0209, 0.0414, 0.0344, 0.0367], + device='cuda:4'), in_proj_covar=tensor([0.0096, 0.0107, 0.0143, 0.0111, 0.0100, 0.0108, 0.0098, 0.0110], + device='cuda:4'), out_proj_covar=tensor([7.4502e-05, 8.2511e-05, 1.1313e-04, 8.5763e-05, 7.7652e-05, 8.0067e-05, + 7.3422e-05, 8.3809e-05], device='cuda:4') +2023-03-26 21:39:10,805 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=98810.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:39:12,639 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0202, 0.9945, 0.9453, 1.1424, 1.2201, 1.1730, 1.0053, 0.9560], + device='cuda:4'), covar=tensor([0.0383, 0.0327, 0.0623, 0.0287, 0.0257, 0.0425, 0.0328, 0.0383], + device='cuda:4'), in_proj_covar=tensor([0.0096, 0.0107, 0.0144, 0.0112, 0.0100, 0.0108, 0.0098, 0.0110], + device='cuda:4'), out_proj_covar=tensor([7.4592e-05, 8.2613e-05, 1.1332e-04, 8.5820e-05, 7.7717e-05, 8.0128e-05, + 7.3544e-05, 8.3886e-05], device='cuda:4') +2023-03-26 21:39:17,933 INFO [finetune.py:976] (4/7) Epoch 18, batch 1450, loss[loss=0.1949, simple_loss=0.2764, pruned_loss=0.05666, over 4905.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2529, pruned_loss=0.05555, over 955845.70 frames. ], batch size: 37, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:39:19,754 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2071, 2.0529, 1.6723, 2.0625, 2.0496, 1.8372, 2.3284, 2.2088], + device='cuda:4'), covar=tensor([0.1177, 0.1880, 0.2840, 0.2597, 0.2530, 0.1612, 0.3971, 0.1524], + device='cuda:4'), in_proj_covar=tensor([0.0183, 0.0187, 0.0234, 0.0254, 0.0245, 0.0202, 0.0214, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:39:40,444 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.618e+02 1.947e+02 2.343e+02 3.750e+02, threshold=3.895e+02, percent-clipped=0.0 +2023-03-26 21:39:52,508 INFO [finetune.py:976] (4/7) Epoch 18, batch 1500, loss[loss=0.1821, simple_loss=0.2533, pruned_loss=0.0554, over 4749.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2552, pruned_loss=0.05677, over 957049.47 frames. ], batch size: 54, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:39:53,290 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7109, 1.6911, 1.4789, 1.8402, 2.2138, 1.8436, 1.5362, 1.4269], + device='cuda:4'), covar=tensor([0.2171, 0.2045, 0.1952, 0.1574, 0.1707, 0.1263, 0.2445, 0.1900], + device='cuda:4'), in_proj_covar=tensor([0.0242, 0.0209, 0.0212, 0.0191, 0.0241, 0.0186, 0.0215, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:39:59,593 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.2953, 1.3994, 1.5609, 0.7697, 1.5229, 1.7669, 1.7336, 1.3365], + device='cuda:4'), covar=tensor([0.0962, 0.0674, 0.0521, 0.0568, 0.0469, 0.0591, 0.0365, 0.0826], + device='cuda:4'), in_proj_covar=tensor([0.0124, 0.0150, 0.0123, 0.0126, 0.0130, 0.0128, 0.0142, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.0693e-05, 1.0893e-04, 8.8221e-05, 8.9495e-05, 9.1716e-05, 9.1744e-05, + 1.0202e-04, 1.0564e-04], device='cuda:4') +2023-03-26 21:40:06,462 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98890.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:40:10,033 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98895.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:40:12,871 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98899.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:40:26,175 INFO [finetune.py:976] (4/7) Epoch 18, batch 1550, loss[loss=0.248, simple_loss=0.3118, pruned_loss=0.09214, over 4299.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2547, pruned_loss=0.05619, over 957298.85 frames. ], batch size: 66, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:40:47,933 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.111e+02 1.516e+02 1.782e+02 2.221e+02 4.511e+02, threshold=3.564e+02, percent-clipped=1.0 +2023-03-26 21:40:48,055 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98951.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:40:51,132 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98956.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:41:00,138 INFO [finetune.py:976] (4/7) Epoch 18, batch 1600, loss[loss=0.1966, simple_loss=0.2608, pruned_loss=0.06619, over 4852.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2526, pruned_loss=0.05595, over 954360.48 frames. ], batch size: 49, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:41:00,249 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98971.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:41:33,949 INFO [finetune.py:976] (4/7) Epoch 18, batch 1650, loss[loss=0.1448, simple_loss=0.1985, pruned_loss=0.04555, over 4203.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2486, pruned_loss=0.05461, over 953754.48 frames. ], batch size: 18, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:41:41,743 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99032.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:42:02,530 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.819e+01 1.633e+02 2.015e+02 2.343e+02 3.841e+02, threshold=4.030e+02, percent-clipped=3.0 +2023-03-26 21:42:03,107 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99051.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:42:22,693 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99066.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:42:25,678 INFO [finetune.py:976] (4/7) Epoch 18, batch 1700, loss[loss=0.1493, simple_loss=0.2262, pruned_loss=0.03619, over 4822.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2476, pruned_loss=0.05496, over 953726.99 frames. ], batch size: 41, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:43:01,006 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99112.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:43:02,162 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99114.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:43:10,653 INFO [finetune.py:976] (4/7) Epoch 18, batch 1750, loss[loss=0.2044, simple_loss=0.2772, pruned_loss=0.06583, over 4836.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2505, pruned_loss=0.05599, over 954486.21 frames. ], batch size: 47, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:43:38,844 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.106e+01 1.643e+02 1.917e+02 2.422e+02 4.876e+02, threshold=3.835e+02, percent-clipped=2.0 +2023-03-26 21:43:41,312 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1549, 1.4490, 2.1516, 2.0670, 1.9071, 1.7842, 1.9570, 1.9874], + device='cuda:4'), covar=tensor([0.3209, 0.3351, 0.3182, 0.3163, 0.4418, 0.3321, 0.3930, 0.2844], + device='cuda:4'), in_proj_covar=tensor([0.0248, 0.0240, 0.0259, 0.0274, 0.0273, 0.0248, 0.0284, 0.0241], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:43:51,914 INFO [finetune.py:976] (4/7) Epoch 18, batch 1800, loss[loss=0.1926, simple_loss=0.2716, pruned_loss=0.05682, over 4818.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.252, pruned_loss=0.05628, over 954165.00 frames. ], batch size: 40, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:43:58,495 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-26 21:44:10,875 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99199.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:44:13,978 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4829, 2.3665, 2.0630, 2.4174, 2.3942, 2.2429, 2.7758, 2.4348], + device='cuda:4'), covar=tensor([0.1366, 0.1776, 0.3075, 0.2183, 0.2316, 0.1543, 0.2284, 0.1613], + device='cuda:4'), in_proj_covar=tensor([0.0183, 0.0188, 0.0234, 0.0253, 0.0245, 0.0202, 0.0214, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:44:25,712 INFO [finetune.py:976] (4/7) Epoch 18, batch 1850, loss[loss=0.1707, simple_loss=0.2594, pruned_loss=0.04099, over 4902.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2521, pruned_loss=0.05609, over 955307.98 frames. ], batch size: 43, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:44:30,696 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99229.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:44:42,400 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99246.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:44:43,017 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99247.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:44:45,842 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.618e+02 1.939e+02 2.302e+02 3.831e+02, threshold=3.878e+02, percent-clipped=0.0 +2023-03-26 21:44:45,939 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99251.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:44:47,175 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2717, 2.2214, 1.8923, 1.1250, 2.0088, 1.8339, 1.8112, 2.0215], + device='cuda:4'), covar=tensor([0.0871, 0.0608, 0.1289, 0.1729, 0.1282, 0.1716, 0.1542, 0.0794], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0194, 0.0201, 0.0183, 0.0214, 0.0209, 0.0222, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:44:57,630 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99268.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:44:59,346 INFO [finetune.py:976] (4/7) Epoch 18, batch 1900, loss[loss=0.1722, simple_loss=0.2488, pruned_loss=0.04782, over 4764.00 frames. ], tot_loss[loss=0.1834, simple_loss=0.2539, pruned_loss=0.05646, over 954735.69 frames. ], batch size: 28, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:45:11,536 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99290.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:45:33,104 INFO [finetune.py:976] (4/7) Epoch 18, batch 1950, loss[loss=0.1615, simple_loss=0.232, pruned_loss=0.04553, over 4749.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2512, pruned_loss=0.05503, over 955586.34 frames. ], batch size: 59, lr: 3.35e-03, grad_scale: 32.0 +2023-03-26 21:45:35,068 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99324.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:45:36,833 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99327.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:45:38,111 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99329.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:45:52,710 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.452e+02 1.660e+02 1.987e+02 4.820e+02, threshold=3.320e+02, percent-clipped=1.0 +2023-03-26 21:46:06,276 INFO [finetune.py:976] (4/7) Epoch 18, batch 2000, loss[loss=0.1648, simple_loss=0.2269, pruned_loss=0.05129, over 4905.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.249, pruned_loss=0.05488, over 954921.16 frames. ], batch size: 32, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:46:15,409 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99385.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:46:24,828 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4685, 1.4448, 1.9851, 1.7430, 1.4699, 3.4483, 1.3080, 1.5976], + device='cuda:4'), covar=tensor([0.0960, 0.1838, 0.1148, 0.0984, 0.1656, 0.0203, 0.1581, 0.1715], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 21:46:28,779 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.53 vs. limit=5.0 +2023-03-26 21:46:30,179 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99407.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:46:39,579 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99420.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:46:40,078 INFO [finetune.py:976] (4/7) Epoch 18, batch 2050, loss[loss=0.1775, simple_loss=0.2379, pruned_loss=0.05853, over 4731.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2456, pruned_loss=0.05386, over 955584.94 frames. ], batch size: 54, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:46:54,374 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.63 vs. limit=5.0 +2023-03-26 21:46:59,827 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.109e+02 1.434e+02 1.742e+02 2.145e+02 5.049e+02, threshold=3.484e+02, percent-clipped=5.0 +2023-03-26 21:47:17,711 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99468.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 21:47:19,434 INFO [finetune.py:976] (4/7) Epoch 18, batch 2100, loss[loss=0.1939, simple_loss=0.2693, pruned_loss=0.05923, over 4855.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2461, pruned_loss=0.05433, over 954707.67 frames. ], batch size: 44, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:47:31,494 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99481.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:47:39,513 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7483, 1.8324, 1.5929, 1.9962, 2.1967, 1.9984, 1.7066, 1.4715], + device='cuda:4'), covar=tensor([0.2159, 0.1807, 0.1792, 0.1517, 0.1755, 0.1171, 0.2200, 0.1859], + device='cuda:4'), in_proj_covar=tensor([0.0241, 0.0208, 0.0211, 0.0190, 0.0239, 0.0185, 0.0214, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:47:54,064 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8108, 1.7061, 1.6527, 1.7677, 1.4486, 4.4864, 1.5838, 1.9484], + device='cuda:4'), covar=tensor([0.3329, 0.2552, 0.2239, 0.2356, 0.1679, 0.0124, 0.2655, 0.1351], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0115, 0.0120, 0.0122, 0.0113, 0.0095, 0.0095, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 21:47:58,125 INFO [finetune.py:976] (4/7) Epoch 18, batch 2150, loss[loss=0.2318, simple_loss=0.3031, pruned_loss=0.08029, over 4892.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2503, pruned_loss=0.05632, over 954437.84 frames. ], batch size: 43, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:48:08,437 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99529.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 21:48:16,785 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3768, 2.1637, 1.7887, 0.8055, 1.9976, 1.8022, 1.6960, 2.0130], + device='cuda:4'), covar=tensor([0.0860, 0.0932, 0.1635, 0.2213, 0.1408, 0.2479, 0.2260, 0.0964], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0196, 0.0201, 0.0184, 0.0215, 0.0210, 0.0224, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:48:28,288 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99546.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:48:35,647 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.054e+02 1.546e+02 1.925e+02 2.366e+02 3.688e+02, threshold=3.850e+02, percent-clipped=2.0 +2023-03-26 21:48:35,745 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99551.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:48:52,015 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99568.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:48:53,719 INFO [finetune.py:976] (4/7) Epoch 18, batch 2200, loss[loss=0.2008, simple_loss=0.2699, pruned_loss=0.06585, over 4867.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2543, pruned_loss=0.05811, over 954301.79 frames. ], batch size: 31, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:49:03,269 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99585.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:49:05,756 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99589.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:49:08,743 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99594.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:49:11,758 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99599.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:49:11,827 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.6537, 3.1893, 2.8986, 1.6322, 3.1087, 2.6093, 2.4755, 2.8343], + device='cuda:4'), covar=tensor([0.0616, 0.0910, 0.1450, 0.2089, 0.1403, 0.1896, 0.1789, 0.1013], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0195, 0.0200, 0.0184, 0.0214, 0.0209, 0.0224, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:49:27,187 INFO [finetune.py:976] (4/7) Epoch 18, batch 2250, loss[loss=0.1353, simple_loss=0.1998, pruned_loss=0.03545, over 3909.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2548, pruned_loss=0.05779, over 953669.25 frames. ], batch size: 17, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:49:29,582 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99624.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:49:31,397 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99627.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:49:33,076 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99629.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:49:46,336 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99650.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:49:46,822 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.914e+01 1.516e+02 1.824e+02 2.092e+02 3.162e+02, threshold=3.647e+02, percent-clipped=0.0 +2023-03-26 21:49:55,241 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9679, 2.7401, 2.5075, 3.0652, 2.5893, 2.6916, 2.6808, 3.6356], + device='cuda:4'), covar=tensor([0.3083, 0.4024, 0.2856, 0.3059, 0.3858, 0.2303, 0.3716, 0.1419], + device='cuda:4'), in_proj_covar=tensor([0.0288, 0.0261, 0.0227, 0.0275, 0.0251, 0.0219, 0.0251, 0.0231], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:50:00,834 INFO [finetune.py:976] (4/7) Epoch 18, batch 2300, loss[loss=0.1516, simple_loss=0.2176, pruned_loss=0.04278, over 4932.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2554, pruned_loss=0.05771, over 954814.62 frames. ], batch size: 33, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:50:03,810 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99675.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:50:06,803 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99680.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:50:06,830 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99680.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:50:24,137 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99707.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:50:34,083 INFO [finetune.py:976] (4/7) Epoch 18, batch 2350, loss[loss=0.1462, simple_loss=0.2261, pruned_loss=0.03314, over 4927.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2518, pruned_loss=0.05624, over 955591.10 frames. ], batch size: 38, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:50:47,861 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99741.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:50:54,365 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.551e+02 1.845e+02 2.144e+02 4.060e+02, threshold=3.690e+02, percent-clipped=1.0 +2023-03-26 21:50:56,921 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99755.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:51:08,064 INFO [finetune.py:976] (4/7) Epoch 18, batch 2400, loss[loss=0.151, simple_loss=0.2143, pruned_loss=0.04382, over 4214.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2477, pruned_loss=0.05465, over 954467.72 frames. ], batch size: 18, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:51:11,643 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99776.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:51:41,398 INFO [finetune.py:976] (4/7) Epoch 18, batch 2450, loss[loss=0.1442, simple_loss=0.2111, pruned_loss=0.03862, over 3998.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2455, pruned_loss=0.05419, over 955141.42 frames. ], batch size: 17, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:51:43,696 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99824.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 21:51:45,153 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-26 21:51:58,444 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.5556, 3.9277, 4.1799, 4.3996, 4.3161, 3.9944, 4.6544, 1.4161], + device='cuda:4'), covar=tensor([0.0826, 0.0901, 0.0840, 0.1134, 0.1278, 0.1724, 0.0637, 0.5830], + device='cuda:4'), in_proj_covar=tensor([0.0348, 0.0244, 0.0278, 0.0292, 0.0334, 0.0282, 0.0300, 0.0295], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:52:01,731 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.091e+01 1.409e+02 1.705e+02 2.080e+02 4.896e+02, threshold=3.409e+02, percent-clipped=2.0 +2023-03-26 21:52:11,146 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.41 vs. limit=5.0 +2023-03-26 21:52:14,325 INFO [finetune.py:976] (4/7) Epoch 18, batch 2500, loss[loss=0.1874, simple_loss=0.2788, pruned_loss=0.04795, over 4800.00 frames. ], tot_loss[loss=0.179, simple_loss=0.248, pruned_loss=0.05502, over 955864.30 frames. ], batch size: 45, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:52:26,293 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99885.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:52:50,111 INFO [finetune.py:976] (4/7) Epoch 18, batch 2550, loss[loss=0.2114, simple_loss=0.2973, pruned_loss=0.06271, over 4844.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2522, pruned_loss=0.05625, over 957738.78 frames. ], batch size: 49, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:52:52,502 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99924.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:52:52,520 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99924.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:52:58,898 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99933.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:53:06,654 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99945.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:53:10,621 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.138e+02 1.647e+02 1.933e+02 2.438e+02 4.501e+02, threshold=3.867e+02, percent-clipped=7.0 +2023-03-26 21:53:29,760 INFO [finetune.py:976] (4/7) Epoch 18, batch 2600, loss[loss=0.2149, simple_loss=0.278, pruned_loss=0.07587, over 4865.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.2545, pruned_loss=0.05699, over 957470.49 frames. ], batch size: 34, lr: 3.34e-03, grad_scale: 32.0 +2023-03-26 21:53:30,434 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=99972.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:53:40,421 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99980.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:54:12,874 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100007.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 21:54:24,616 INFO [finetune.py:976] (4/7) Epoch 18, batch 2650, loss[loss=0.1705, simple_loss=0.2404, pruned_loss=0.05031, over 4896.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2554, pruned_loss=0.05734, over 957883.24 frames. ], batch size: 32, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 21:54:29,389 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100028.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:54:35,199 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100036.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:54:46,209 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.050e+02 1.530e+02 1.912e+02 2.295e+02 4.144e+02, threshold=3.823e+02, percent-clipped=1.0 +2023-03-26 21:54:56,610 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100068.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 21:54:58,305 INFO [finetune.py:976] (4/7) Epoch 18, batch 2700, loss[loss=0.1648, simple_loss=0.2348, pruned_loss=0.04742, over 4787.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2544, pruned_loss=0.05653, over 958118.39 frames. ], batch size: 51, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 21:55:01,442 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100076.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:55:31,862 INFO [finetune.py:976] (4/7) Epoch 18, batch 2750, loss[loss=0.1712, simple_loss=0.2382, pruned_loss=0.05213, over 4936.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2519, pruned_loss=0.05642, over 956921.21 frames. ], batch size: 33, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 21:55:33,702 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100124.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:55:33,746 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100124.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 21:55:52,598 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 21:55:52,998 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.121e+02 1.510e+02 1.766e+02 2.096e+02 4.575e+02, threshold=3.532e+02, percent-clipped=1.0 +2023-03-26 21:56:05,347 INFO [finetune.py:976] (4/7) Epoch 18, batch 2800, loss[loss=0.1757, simple_loss=0.2449, pruned_loss=0.05321, over 4826.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2481, pruned_loss=0.05481, over 957655.23 frames. ], batch size: 30, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 21:56:06,012 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100172.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 21:56:38,930 INFO [finetune.py:976] (4/7) Epoch 18, batch 2850, loss[loss=0.1756, simple_loss=0.2437, pruned_loss=0.05369, over 4800.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2473, pruned_loss=0.05483, over 957853.59 frames. ], batch size: 45, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 21:56:40,869 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100224.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:56:49,830 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100238.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:56:54,562 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100245.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:56:59,192 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.104e+02 1.635e+02 1.899e+02 2.309e+02 4.393e+02, threshold=3.799e+02, percent-clipped=4.0 +2023-03-26 21:57:11,715 INFO [finetune.py:976] (4/7) Epoch 18, batch 2900, loss[loss=0.1392, simple_loss=0.206, pruned_loss=0.03621, over 4762.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2503, pruned_loss=0.05635, over 957108.20 frames. ], batch size: 26, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 21:57:12,868 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100272.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:57:18,420 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3657, 1.7596, 1.4238, 1.4484, 1.8516, 1.7454, 1.7172, 1.7346], + device='cuda:4'), covar=tensor([0.0530, 0.0311, 0.0532, 0.0351, 0.0342, 0.0697, 0.0322, 0.0353], + device='cuda:4'), in_proj_covar=tensor([0.0096, 0.0107, 0.0143, 0.0111, 0.0101, 0.0109, 0.0099, 0.0110], + device='cuda:4'), out_proj_covar=tensor([7.4682e-05, 8.2502e-05, 1.1302e-04, 8.4919e-05, 7.8905e-05, 8.0496e-05, + 7.3786e-05, 8.3890e-05], device='cuda:4') +2023-03-26 21:57:26,156 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100293.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:57:30,861 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100299.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:57:45,540 INFO [finetune.py:976] (4/7) Epoch 18, batch 2950, loss[loss=0.1678, simple_loss=0.2413, pruned_loss=0.04718, over 4197.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2533, pruned_loss=0.05657, over 955386.46 frames. ], batch size: 65, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 21:57:52,855 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5955, 1.3902, 1.6055, 0.8518, 1.5524, 1.5323, 1.5053, 1.3267], + device='cuda:4'), covar=tensor([0.0573, 0.0848, 0.0623, 0.1015, 0.0805, 0.0687, 0.0679, 0.1417], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0134, 0.0140, 0.0121, 0.0123, 0.0138, 0.0140, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 21:57:55,279 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100336.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:58:06,321 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.490e+02 1.822e+02 2.174e+02 4.072e+02, threshold=3.643e+02, percent-clipped=2.0 +2023-03-26 21:58:13,588 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100363.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 21:58:18,818 INFO [finetune.py:976] (4/7) Epoch 18, batch 3000, loss[loss=0.2124, simple_loss=0.2737, pruned_loss=0.07552, over 4892.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2555, pruned_loss=0.05776, over 956151.16 frames. ], batch size: 37, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 21:58:18,818 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 21:58:24,620 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6750, 1.5890, 1.5197, 1.5688, 0.9848, 3.0399, 1.1260, 1.5986], + device='cuda:4'), covar=tensor([0.3415, 0.2423, 0.2188, 0.2374, 0.1865, 0.0295, 0.2585, 0.1275], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0114, 0.0120, 0.0122, 0.0113, 0.0095, 0.0095, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 21:58:31,198 INFO [finetune.py:1010] (4/7) Epoch 18, validation: loss=0.1568, simple_loss=0.2261, pruned_loss=0.04375, over 2265189.00 frames. +2023-03-26 21:58:31,198 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 21:58:44,386 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100384.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:59:06,993 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100402.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 21:59:29,743 INFO [finetune.py:976] (4/7) Epoch 18, batch 3050, loss[loss=0.1783, simple_loss=0.2512, pruned_loss=0.05268, over 4925.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2553, pruned_loss=0.05708, over 954553.98 frames. ], batch size: 33, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 21:59:53,774 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.049e+02 1.572e+02 1.917e+02 2.276e+02 3.597e+02, threshold=3.833e+02, percent-clipped=0.0 +2023-03-26 22:00:01,163 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100463.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 22:00:07,402 INFO [finetune.py:976] (4/7) Epoch 18, batch 3100, loss[loss=0.1682, simple_loss=0.2492, pruned_loss=0.04367, over 4914.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2531, pruned_loss=0.05617, over 952805.42 frames. ], batch size: 35, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 22:00:18,796 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5958, 1.5293, 1.3939, 1.7306, 1.5960, 1.7485, 0.9966, 1.3888], + device='cuda:4'), covar=tensor([0.2071, 0.1888, 0.1811, 0.1450, 0.1528, 0.1191, 0.2503, 0.1816], + device='cuda:4'), in_proj_covar=tensor([0.0240, 0.0207, 0.0210, 0.0190, 0.0239, 0.0185, 0.0214, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:00:40,542 INFO [finetune.py:976] (4/7) Epoch 18, batch 3150, loss[loss=0.174, simple_loss=0.2384, pruned_loss=0.05478, over 4741.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.25, pruned_loss=0.05543, over 954793.59 frames. ], batch size: 54, lr: 3.34e-03, grad_scale: 16.0 +2023-03-26 22:00:54,128 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8711, 1.5783, 2.2513, 1.4755, 1.9828, 2.1960, 1.5359, 2.3033], + device='cuda:4'), covar=tensor([0.1387, 0.2391, 0.1273, 0.2006, 0.0987, 0.1405, 0.3119, 0.0864], + device='cuda:4'), in_proj_covar=tensor([0.0190, 0.0202, 0.0189, 0.0186, 0.0174, 0.0210, 0.0214, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:01:00,946 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.466e+01 1.523e+02 1.835e+02 2.195e+02 4.344e+02, threshold=3.670e+02, percent-clipped=3.0 +2023-03-26 22:01:12,895 INFO [finetune.py:976] (4/7) Epoch 18, batch 3200, loss[loss=0.2329, simple_loss=0.2851, pruned_loss=0.09037, over 4752.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2461, pruned_loss=0.05416, over 954665.92 frames. ], batch size: 54, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:01:28,940 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100594.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:01:46,316 INFO [finetune.py:976] (4/7) Epoch 18, batch 3250, loss[loss=0.1911, simple_loss=0.2686, pruned_loss=0.05682, over 4872.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2483, pruned_loss=0.05549, over 954222.08 frames. ], batch size: 31, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:02:08,114 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.084e+02 1.502e+02 1.862e+02 2.235e+02 4.464e+02, threshold=3.723e+02, percent-clipped=3.0 +2023-03-26 22:02:14,360 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3946, 2.2716, 1.7426, 2.3303, 2.3195, 2.0282, 2.6326, 2.3854], + device='cuda:4'), covar=tensor([0.1308, 0.2136, 0.3153, 0.2566, 0.2508, 0.1611, 0.2848, 0.1733], + device='cuda:4'), in_proj_covar=tensor([0.0182, 0.0187, 0.0233, 0.0251, 0.0243, 0.0201, 0.0212, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:02:14,917 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100663.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 22:02:20,156 INFO [finetune.py:976] (4/7) Epoch 18, batch 3300, loss[loss=0.1959, simple_loss=0.253, pruned_loss=0.06938, over 4827.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2507, pruned_loss=0.05616, over 953375.46 frames. ], batch size: 25, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:02:22,651 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1303, 1.7948, 1.8378, 0.8300, 2.1735, 2.3176, 2.0648, 1.6777], + device='cuda:4'), covar=tensor([0.0861, 0.0758, 0.0459, 0.0657, 0.0402, 0.0614, 0.0462, 0.0747], + device='cuda:4'), in_proj_covar=tensor([0.0125, 0.0152, 0.0124, 0.0126, 0.0131, 0.0129, 0.0142, 0.0148], + device='cuda:4'), out_proj_covar=tensor([9.1446e-05, 1.1011e-04, 8.8804e-05, 8.9821e-05, 9.2159e-05, 9.2648e-05, + 1.0210e-04, 1.0617e-04], device='cuda:4') +2023-03-26 22:02:47,027 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100711.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 22:02:53,405 INFO [finetune.py:976] (4/7) Epoch 18, batch 3350, loss[loss=0.1926, simple_loss=0.2643, pruned_loss=0.06046, over 4895.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2511, pruned_loss=0.05573, over 951791.98 frames. ], batch size: 37, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:02:59,525 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100730.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 22:03:00,610 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3520, 1.2942, 1.2068, 1.3169, 1.5996, 1.4882, 1.2982, 1.1983], + device='cuda:4'), covar=tensor([0.0317, 0.0285, 0.0579, 0.0298, 0.0211, 0.0433, 0.0341, 0.0377], + device='cuda:4'), in_proj_covar=tensor([0.0095, 0.0107, 0.0143, 0.0110, 0.0101, 0.0108, 0.0098, 0.0109], + device='cuda:4'), out_proj_covar=tensor([7.4040e-05, 8.2311e-05, 1.1255e-04, 8.4739e-05, 7.8322e-05, 7.9751e-05, + 7.3354e-05, 8.3219e-05], device='cuda:4') +2023-03-26 22:03:14,086 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.901e+01 1.577e+02 1.865e+02 2.249e+02 4.268e+02, threshold=3.731e+02, percent-clipped=3.0 +2023-03-26 22:03:18,264 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100758.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 22:03:26,649 INFO [finetune.py:976] (4/7) Epoch 18, batch 3400, loss[loss=0.1743, simple_loss=0.2434, pruned_loss=0.0526, over 4834.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2519, pruned_loss=0.05569, over 953007.84 frames. ], batch size: 49, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:03:40,308 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100791.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 22:03:41,597 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.61 vs. limit=5.0 +2023-03-26 22:04:13,290 INFO [finetune.py:976] (4/7) Epoch 18, batch 3450, loss[loss=0.1775, simple_loss=0.238, pruned_loss=0.0585, over 4723.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2522, pruned_loss=0.05608, over 952687.15 frames. ], batch size: 23, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:04:49,164 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.551e+02 1.762e+02 2.136e+02 3.810e+02, threshold=3.524e+02, percent-clipped=1.0 +2023-03-26 22:05:04,971 INFO [finetune.py:976] (4/7) Epoch 18, batch 3500, loss[loss=0.1294, simple_loss=0.1967, pruned_loss=0.03106, over 4783.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.25, pruned_loss=0.0557, over 953049.96 frames. ], batch size: 26, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:05:21,052 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100894.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:05:38,764 INFO [finetune.py:976] (4/7) Epoch 18, batch 3550, loss[loss=0.1599, simple_loss=0.2309, pruned_loss=0.04447, over 4748.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2464, pruned_loss=0.05408, over 952147.91 frames. ], batch size: 28, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:05:52,402 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=100942.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:05:59,322 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.557e+01 1.553e+02 1.835e+02 2.348e+02 4.609e+02, threshold=3.670e+02, percent-clipped=4.0 +2023-03-26 22:06:12,133 INFO [finetune.py:976] (4/7) Epoch 18, batch 3600, loss[loss=0.1361, simple_loss=0.2101, pruned_loss=0.03106, over 4828.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2443, pruned_loss=0.05314, over 952841.15 frames. ], batch size: 30, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:06:46,075 INFO [finetune.py:976] (4/7) Epoch 18, batch 3650, loss[loss=0.1978, simple_loss=0.2739, pruned_loss=0.06084, over 4868.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2487, pruned_loss=0.05486, over 952456.79 frames. ], batch size: 34, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:06:58,257 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5766, 0.7953, 1.6143, 1.4870, 1.3970, 1.2908, 1.4715, 1.4763], + device='cuda:4'), covar=tensor([0.2587, 0.3010, 0.2536, 0.2737, 0.3426, 0.2992, 0.2896, 0.2406], + device='cuda:4'), in_proj_covar=tensor([0.0248, 0.0238, 0.0258, 0.0274, 0.0273, 0.0247, 0.0282, 0.0239], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:07:06,797 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.559e+02 1.860e+02 2.177e+02 4.070e+02, threshold=3.719e+02, percent-clipped=1.0 +2023-03-26 22:07:11,005 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=101058.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:07:18,907 INFO [finetune.py:976] (4/7) Epoch 18, batch 3700, loss[loss=0.2338, simple_loss=0.3006, pruned_loss=0.08352, over 4809.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2523, pruned_loss=0.05541, over 953028.18 frames. ], batch size: 38, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:07:23,709 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1018, 1.9114, 2.0343, 1.2829, 1.9315, 2.1404, 2.0477, 1.5739], + device='cuda:4'), covar=tensor([0.0568, 0.0690, 0.0686, 0.0950, 0.0698, 0.0601, 0.0624, 0.1248], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0134, 0.0139, 0.0120, 0.0123, 0.0138, 0.0139, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:07:28,506 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=101086.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 22:07:29,151 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6239, 1.5198, 1.4582, 1.5588, 1.0532, 3.5754, 1.3517, 1.7575], + device='cuda:4'), covar=tensor([0.3186, 0.2598, 0.2248, 0.2346, 0.1816, 0.0190, 0.2672, 0.1320], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0115, 0.0120, 0.0122, 0.0113, 0.0095, 0.0095, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 22:07:42,151 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=101106.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:07:43,259 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101107.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:07:52,640 INFO [finetune.py:976] (4/7) Epoch 18, batch 3750, loss[loss=0.1939, simple_loss=0.274, pruned_loss=0.05693, over 4912.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2541, pruned_loss=0.05693, over 952118.31 frames. ], batch size: 38, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:08:12,837 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.641e+02 1.814e+02 2.131e+02 4.110e+02, threshold=3.627e+02, percent-clipped=2.0 +2023-03-26 22:08:15,828 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0961, 1.9476, 1.6709, 1.8590, 1.8548, 1.8075, 1.8646, 2.5706], + device='cuda:4'), covar=tensor([0.3751, 0.4066, 0.3205, 0.3473, 0.3925, 0.2399, 0.3717, 0.1634], + device='cuda:4'), in_proj_covar=tensor([0.0286, 0.0260, 0.0228, 0.0273, 0.0251, 0.0219, 0.0250, 0.0231], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:08:24,451 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=101168.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:08:26,160 INFO [finetune.py:976] (4/7) Epoch 18, batch 3800, loss[loss=0.1618, simple_loss=0.2324, pruned_loss=0.04562, over 4191.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2553, pruned_loss=0.05716, over 952717.83 frames. ], batch size: 66, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:08:48,846 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7980, 1.1701, 0.8108, 1.6193, 2.0873, 1.4808, 1.5174, 1.6898], + device='cuda:4'), covar=tensor([0.1425, 0.2162, 0.1963, 0.1118, 0.1985, 0.1949, 0.1439, 0.1917], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0096, 0.0110, 0.0092, 0.0121, 0.0094, 0.0099, 0.0090], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 22:08:49,565 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.04 vs. limit=5.0 +2023-03-26 22:08:59,889 INFO [finetune.py:976] (4/7) Epoch 18, batch 3850, loss[loss=0.2062, simple_loss=0.273, pruned_loss=0.06976, over 4811.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2521, pruned_loss=0.05614, over 952230.56 frames. ], batch size: 51, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:09:06,894 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=2.00 vs. limit=2.0 +2023-03-26 22:09:11,882 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7590, 2.5647, 2.0812, 2.7230, 2.6612, 2.3038, 3.1433, 2.6846], + device='cuda:4'), covar=tensor([0.1250, 0.2087, 0.3010, 0.2548, 0.2418, 0.1622, 0.2802, 0.1732], + device='cuda:4'), in_proj_covar=tensor([0.0185, 0.0189, 0.0236, 0.0254, 0.0246, 0.0204, 0.0215, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:09:30,994 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.034e+02 1.525e+02 1.862e+02 2.180e+02 4.556e+02, threshold=3.724e+02, percent-clipped=2.0 +2023-03-26 22:09:57,717 INFO [finetune.py:976] (4/7) Epoch 18, batch 3900, loss[loss=0.1543, simple_loss=0.2268, pruned_loss=0.04091, over 4823.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.25, pruned_loss=0.0558, over 954162.01 frames. ], batch size: 38, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:10:41,972 INFO [finetune.py:976] (4/7) Epoch 18, batch 3950, loss[loss=0.1521, simple_loss=0.2306, pruned_loss=0.03685, over 4809.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2463, pruned_loss=0.05391, over 955910.33 frames. ], batch size: 25, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:10:52,496 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-26 22:10:56,151 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.89 vs. limit=5.0 +2023-03-26 22:11:02,252 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.024e+02 1.470e+02 1.754e+02 2.083e+02 3.090e+02, threshold=3.508e+02, percent-clipped=0.0 +2023-03-26 22:11:15,361 INFO [finetune.py:976] (4/7) Epoch 18, batch 4000, loss[loss=0.1836, simple_loss=0.2736, pruned_loss=0.04681, over 4906.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2458, pruned_loss=0.05389, over 953934.88 frames. ], batch size: 37, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:11:26,026 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=101386.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 22:11:49,435 INFO [finetune.py:976] (4/7) Epoch 18, batch 4050, loss[loss=0.1575, simple_loss=0.2339, pruned_loss=0.04057, over 4898.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2481, pruned_loss=0.05527, over 952392.70 frames. ], batch size: 35, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:11:58,804 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=101434.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 22:12:10,025 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.148e+02 1.670e+02 2.040e+02 2.363e+02 9.256e+02, threshold=4.080e+02, percent-clipped=2.0 +2023-03-26 22:12:17,270 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=101463.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:12:22,998 INFO [finetune.py:976] (4/7) Epoch 18, batch 4100, loss[loss=0.1573, simple_loss=0.2275, pruned_loss=0.04356, over 4793.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2514, pruned_loss=0.05581, over 953892.62 frames. ], batch size: 25, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:12:39,632 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.1342, 4.6359, 4.4051, 2.4619, 4.8761, 3.8471, 0.7299, 3.3897], + device='cuda:4'), covar=tensor([0.2330, 0.1599, 0.1487, 0.3009, 0.0649, 0.0705, 0.4841, 0.1233], + device='cuda:4'), in_proj_covar=tensor([0.0154, 0.0178, 0.0162, 0.0131, 0.0162, 0.0126, 0.0150, 0.0125], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 22:12:56,255 INFO [finetune.py:976] (4/7) Epoch 18, batch 4150, loss[loss=0.1483, simple_loss=0.2295, pruned_loss=0.03357, over 4772.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2524, pruned_loss=0.0558, over 953188.82 frames. ], batch size: 26, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:13:16,874 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.507e+01 1.508e+02 1.851e+02 2.208e+02 3.984e+02, threshold=3.702e+02, percent-clipped=0.0 +2023-03-26 22:13:26,548 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4054, 1.3233, 1.3259, 1.3348, 0.8193, 2.3881, 0.7708, 1.2299], + device='cuda:4'), covar=tensor([0.3426, 0.2520, 0.2238, 0.2492, 0.1970, 0.0326, 0.2732, 0.1300], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0115, 0.0119, 0.0122, 0.0113, 0.0095, 0.0095, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 22:13:29,469 INFO [finetune.py:976] (4/7) Epoch 18, batch 4200, loss[loss=0.1361, simple_loss=0.1993, pruned_loss=0.03647, over 4174.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2522, pruned_loss=0.05533, over 954167.91 frames. ], batch size: 18, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:14:03,053 INFO [finetune.py:976] (4/7) Epoch 18, batch 4250, loss[loss=0.1948, simple_loss=0.2591, pruned_loss=0.06525, over 4863.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2498, pruned_loss=0.05423, over 954683.03 frames. ], batch size: 44, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:14:11,893 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0658, 1.9046, 1.6591, 1.6944, 1.8395, 1.8295, 1.8850, 2.4870], + device='cuda:4'), covar=tensor([0.3749, 0.4102, 0.3201, 0.3573, 0.3712, 0.2382, 0.3399, 0.1682], + device='cuda:4'), in_proj_covar=tensor([0.0288, 0.0262, 0.0229, 0.0274, 0.0251, 0.0220, 0.0251, 0.0233], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:14:24,253 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.150e+02 1.580e+02 1.774e+02 2.219e+02 3.425e+02, threshold=3.547e+02, percent-clipped=0.0 +2023-03-26 22:14:38,483 INFO [finetune.py:976] (4/7) Epoch 18, batch 4300, loss[loss=0.188, simple_loss=0.2528, pruned_loss=0.06163, over 4861.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2471, pruned_loss=0.05356, over 956161.17 frames. ], batch size: 31, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:14:54,320 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101684.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:15:33,151 INFO [finetune.py:976] (4/7) Epoch 18, batch 4350, loss[loss=0.1876, simple_loss=0.2656, pruned_loss=0.05481, over 4842.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2449, pruned_loss=0.05289, over 957827.27 frames. ], batch size: 47, lr: 3.33e-03, grad_scale: 16.0 +2023-03-26 22:15:43,181 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 22:16:05,166 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=101745.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:16:10,243 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.272e+01 1.483e+02 1.686e+02 2.087e+02 3.591e+02, threshold=3.373e+02, percent-clipped=1.0 +2023-03-26 22:16:16,946 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=101763.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:16:21,724 INFO [finetune.py:976] (4/7) Epoch 18, batch 4400, loss[loss=0.1436, simple_loss=0.2149, pruned_loss=0.03619, over 4807.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2462, pruned_loss=0.05377, over 956739.30 frames. ], batch size: 25, lr: 3.32e-03, grad_scale: 16.0 +2023-03-26 22:16:49,626 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=101811.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:16:55,630 INFO [finetune.py:976] (4/7) Epoch 18, batch 4450, loss[loss=0.1584, simple_loss=0.239, pruned_loss=0.03896, over 4109.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2508, pruned_loss=0.05546, over 954444.55 frames. ], batch size: 65, lr: 3.32e-03, grad_scale: 16.0 +2023-03-26 22:17:00,976 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2864, 3.6926, 3.9421, 4.1590, 4.0947, 3.8230, 4.3772, 1.4516], + device='cuda:4'), covar=tensor([0.0857, 0.0910, 0.0776, 0.1073, 0.1137, 0.1535, 0.0628, 0.5374], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0242, 0.0276, 0.0290, 0.0331, 0.0280, 0.0300, 0.0294], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:17:08,157 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.83 vs. limit=5.0 +2023-03-26 22:17:16,370 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-26 22:17:16,756 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.606e+02 1.893e+02 2.313e+02 4.401e+02, threshold=3.785e+02, percent-clipped=7.0 +2023-03-26 22:17:24,652 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-26 22:17:29,384 INFO [finetune.py:976] (4/7) Epoch 18, batch 4500, loss[loss=0.1922, simple_loss=0.2649, pruned_loss=0.05973, over 4898.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2522, pruned_loss=0.05551, over 954844.83 frames. ], batch size: 35, lr: 3.32e-03, grad_scale: 16.0 +2023-03-26 22:18:00,215 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.7329, 3.1973, 3.4294, 3.5990, 3.5305, 3.2996, 3.8033, 1.3718], + device='cuda:4'), covar=tensor([0.0855, 0.0951, 0.0810, 0.1026, 0.1187, 0.1386, 0.0886, 0.4917], + device='cuda:4'), in_proj_covar=tensor([0.0346, 0.0241, 0.0275, 0.0289, 0.0330, 0.0279, 0.0300, 0.0293], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:18:02,610 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.9196, 3.4967, 3.7004, 3.6269, 3.4683, 3.4233, 4.0674, 1.4444], + device='cuda:4'), covar=tensor([0.1264, 0.1723, 0.1390, 0.1939, 0.2027, 0.2203, 0.1315, 0.6632], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0242, 0.0275, 0.0289, 0.0330, 0.0279, 0.0300, 0.0293], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:18:02,645 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101920.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 22:18:03,132 INFO [finetune.py:976] (4/7) Epoch 18, batch 4550, loss[loss=0.206, simple_loss=0.2723, pruned_loss=0.06983, over 4805.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2529, pruned_loss=0.05589, over 954301.00 frames. ], batch size: 40, lr: 3.32e-03, grad_scale: 16.0 +2023-03-26 22:18:14,478 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4692, 2.3619, 2.0236, 2.2122, 2.2422, 2.2110, 2.2748, 3.0330], + device='cuda:4'), covar=tensor([0.3372, 0.3663, 0.2911, 0.3338, 0.3392, 0.2378, 0.3296, 0.1450], + device='cuda:4'), in_proj_covar=tensor([0.0285, 0.0260, 0.0227, 0.0272, 0.0250, 0.0218, 0.0250, 0.0231], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:18:18,538 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3499, 1.5355, 1.7128, 0.8837, 1.6172, 1.8688, 1.9016, 1.4856], + device='cuda:4'), covar=tensor([0.0852, 0.0683, 0.0439, 0.0495, 0.0466, 0.0564, 0.0294, 0.0703], + device='cuda:4'), in_proj_covar=tensor([0.0124, 0.0151, 0.0123, 0.0126, 0.0131, 0.0129, 0.0141, 0.0148], + device='cuda:4'), out_proj_covar=tensor([9.1202e-05, 1.0962e-04, 8.8156e-05, 8.9866e-05, 9.2308e-05, 9.2621e-05, + 1.0169e-04, 1.0628e-04], device='cuda:4') +2023-03-26 22:18:24,137 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.522e+02 1.794e+02 2.336e+02 4.256e+02, threshold=3.587e+02, percent-clipped=2.0 +2023-03-26 22:18:36,748 INFO [finetune.py:976] (4/7) Epoch 18, batch 4600, loss[loss=0.1996, simple_loss=0.278, pruned_loss=0.06066, over 4905.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2511, pruned_loss=0.05523, over 952865.69 frames. ], batch size: 37, lr: 3.32e-03, grad_scale: 16.0 +2023-03-26 22:18:42,886 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=101981.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 22:18:47,486 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101987.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:19:11,219 INFO [finetune.py:976] (4/7) Epoch 18, batch 4650, loss[loss=0.1507, simple_loss=0.2213, pruned_loss=0.04009, over 4749.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2491, pruned_loss=0.05518, over 953032.92 frames. ], batch size: 27, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:19:19,636 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-03-26 22:19:23,878 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=102040.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:19:29,251 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=102048.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:19:31,548 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.060e+02 1.564e+02 1.867e+02 2.217e+02 4.281e+02, threshold=3.734e+02, percent-clipped=4.0 +2023-03-26 22:19:32,851 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6037, 1.5632, 2.1483, 3.5275, 2.3969, 2.4394, 0.9489, 2.8952], + device='cuda:4'), covar=tensor([0.1787, 0.1528, 0.1383, 0.0573, 0.0795, 0.1350, 0.1985, 0.0489], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0116, 0.0133, 0.0164, 0.0100, 0.0135, 0.0124, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 22:19:45,057 INFO [finetune.py:976] (4/7) Epoch 18, batch 4700, loss[loss=0.2078, simple_loss=0.2602, pruned_loss=0.07776, over 4898.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2469, pruned_loss=0.05473, over 953693.58 frames. ], batch size: 32, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:19:48,158 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7805, 1.2280, 0.8578, 1.7295, 2.2211, 1.4369, 1.5615, 1.5258], + device='cuda:4'), covar=tensor([0.1519, 0.2118, 0.1898, 0.1182, 0.1927, 0.1862, 0.1420, 0.2078], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0110, 0.0092, 0.0120, 0.0093, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 22:19:57,270 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.4046, 4.6179, 4.9434, 5.2686, 5.1465, 4.7885, 5.5524, 1.7013], + device='cuda:4'), covar=tensor([0.0703, 0.0809, 0.0706, 0.0879, 0.1128, 0.1494, 0.0475, 0.5821], + device='cuda:4'), in_proj_covar=tensor([0.0349, 0.0243, 0.0278, 0.0290, 0.0331, 0.0281, 0.0301, 0.0295], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:20:28,670 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7624, 1.8519, 1.5702, 1.9275, 2.3895, 1.9936, 1.6817, 1.4250], + device='cuda:4'), covar=tensor([0.2215, 0.1848, 0.1898, 0.1590, 0.1629, 0.1209, 0.2306, 0.2012], + device='cuda:4'), in_proj_covar=tensor([0.0242, 0.0209, 0.0213, 0.0193, 0.0241, 0.0186, 0.0214, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:20:31,472 INFO [finetune.py:976] (4/7) Epoch 18, batch 4750, loss[loss=0.1598, simple_loss=0.2311, pruned_loss=0.04427, over 4741.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2452, pruned_loss=0.05441, over 954906.15 frames. ], batch size: 23, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:20:56,162 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.538e+02 1.904e+02 2.326e+02 4.380e+02, threshold=3.807e+02, percent-clipped=2.0 +2023-03-26 22:21:23,319 INFO [finetune.py:976] (4/7) Epoch 18, batch 4800, loss[loss=0.2155, simple_loss=0.2833, pruned_loss=0.07386, over 4815.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2488, pruned_loss=0.05584, over 955560.97 frames. ], batch size: 51, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:21:56,996 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4760, 1.3252, 1.7254, 2.4607, 1.6901, 2.2150, 0.9315, 2.1507], + device='cuda:4'), covar=tensor([0.1738, 0.1457, 0.1111, 0.0775, 0.0985, 0.1036, 0.1606, 0.0597], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0115, 0.0132, 0.0164, 0.0100, 0.0134, 0.0123, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 22:22:00,409 INFO [finetune.py:976] (4/7) Epoch 18, batch 4850, loss[loss=0.1632, simple_loss=0.2282, pruned_loss=0.04912, over 4789.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2501, pruned_loss=0.0555, over 954964.13 frames. ], batch size: 25, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:22:20,219 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.897e+01 1.478e+02 1.803e+02 2.234e+02 3.533e+02, threshold=3.606e+02, percent-clipped=0.0 +2023-03-26 22:22:33,608 INFO [finetune.py:976] (4/7) Epoch 18, batch 4900, loss[loss=0.2001, simple_loss=0.2735, pruned_loss=0.06336, over 4844.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2517, pruned_loss=0.05612, over 953535.26 frames. ], batch size: 44, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:22:37,395 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-26 22:22:37,663 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=102276.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 22:23:06,931 INFO [finetune.py:976] (4/7) Epoch 18, batch 4950, loss[loss=0.1784, simple_loss=0.2588, pruned_loss=0.04904, over 4893.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2522, pruned_loss=0.05607, over 954237.81 frames. ], batch size: 43, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:23:19,538 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=102340.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:23:21,352 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=102343.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:23:26,706 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.970e+01 1.449e+02 1.847e+02 2.188e+02 4.191e+02, threshold=3.694e+02, percent-clipped=1.0 +2023-03-26 22:23:40,084 INFO [finetune.py:976] (4/7) Epoch 18, batch 5000, loss[loss=0.1351, simple_loss=0.2018, pruned_loss=0.03418, over 4932.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.251, pruned_loss=0.05563, over 954076.40 frames. ], batch size: 38, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:23:44,774 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0366, 1.0000, 1.0284, 0.4603, 0.8786, 1.2046, 1.2253, 1.0278], + device='cuda:4'), covar=tensor([0.0864, 0.0591, 0.0492, 0.0556, 0.0538, 0.0540, 0.0404, 0.0639], + device='cuda:4'), in_proj_covar=tensor([0.0124, 0.0151, 0.0124, 0.0126, 0.0131, 0.0129, 0.0142, 0.0148], + device='cuda:4'), out_proj_covar=tensor([9.0980e-05, 1.0946e-04, 8.8428e-05, 8.9577e-05, 9.2391e-05, 9.2233e-05, + 1.0205e-04, 1.0617e-04], device='cuda:4') +2023-03-26 22:23:51,807 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=102388.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:23:58,516 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.9850, 4.3103, 4.4668, 4.7745, 4.7094, 4.3318, 5.1316, 1.5410], + device='cuda:4'), covar=tensor([0.0817, 0.0931, 0.0741, 0.1031, 0.1332, 0.1807, 0.0561, 0.6270], + device='cuda:4'), in_proj_covar=tensor([0.0348, 0.0242, 0.0277, 0.0289, 0.0331, 0.0281, 0.0300, 0.0295], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:24:13,296 INFO [finetune.py:976] (4/7) Epoch 18, batch 5050, loss[loss=0.1966, simple_loss=0.2515, pruned_loss=0.07086, over 4907.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2481, pruned_loss=0.05441, over 954519.06 frames. ], batch size: 36, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:24:33,970 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.223e+01 1.582e+02 1.966e+02 2.354e+02 3.513e+02, threshold=3.932e+02, percent-clipped=0.0 +2023-03-26 22:24:46,896 INFO [finetune.py:976] (4/7) Epoch 18, batch 5100, loss[loss=0.1658, simple_loss=0.2335, pruned_loss=0.04904, over 4902.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2446, pruned_loss=0.05326, over 955101.84 frames. ], batch size: 43, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:25:20,567 INFO [finetune.py:976] (4/7) Epoch 18, batch 5150, loss[loss=0.1631, simple_loss=0.2354, pruned_loss=0.04538, over 4848.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2459, pruned_loss=0.0542, over 952193.48 frames. ], batch size: 49, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:25:53,893 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.631e+02 1.989e+02 2.441e+02 4.766e+02, threshold=3.977e+02, percent-clipped=3.0 +2023-03-26 22:26:14,502 INFO [finetune.py:976] (4/7) Epoch 18, batch 5200, loss[loss=0.1797, simple_loss=0.2527, pruned_loss=0.05335, over 4855.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2489, pruned_loss=0.05486, over 952338.30 frames. ], batch size: 44, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:26:22,140 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=102576.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 22:26:56,624 INFO [finetune.py:976] (4/7) Epoch 18, batch 5250, loss[loss=0.2612, simple_loss=0.3201, pruned_loss=0.1011, over 4921.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2506, pruned_loss=0.05547, over 949928.78 frames. ], batch size: 42, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:26:58,549 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=102624.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 22:27:11,989 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=102643.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:27:17,763 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.182e+02 1.670e+02 1.981e+02 2.217e+02 4.217e+02, threshold=3.962e+02, percent-clipped=1.0 +2023-03-26 22:27:27,034 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3811, 1.6107, 1.2873, 1.5199, 1.7474, 1.6326, 1.4978, 1.4136], + device='cuda:4'), covar=tensor([0.0433, 0.0292, 0.0583, 0.0276, 0.0244, 0.0577, 0.0329, 0.0358], + device='cuda:4'), in_proj_covar=tensor([0.0095, 0.0107, 0.0143, 0.0110, 0.0099, 0.0108, 0.0098, 0.0110], + device='cuda:4'), out_proj_covar=tensor([7.3927e-05, 8.2310e-05, 1.1271e-04, 8.4591e-05, 7.7309e-05, 7.9578e-05, + 7.3443e-05, 8.3837e-05], device='cuda:4') +2023-03-26 22:27:29,393 INFO [finetune.py:976] (4/7) Epoch 18, batch 5300, loss[loss=0.159, simple_loss=0.2384, pruned_loss=0.03975, over 4771.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2525, pruned_loss=0.05593, over 949715.33 frames. ], batch size: 29, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:27:44,024 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=102691.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:27:44,076 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8544, 1.7999, 1.7518, 1.8180, 1.6727, 4.6231, 1.8611, 2.3798], + device='cuda:4'), covar=tensor([0.3292, 0.2415, 0.2081, 0.2370, 0.1533, 0.0103, 0.2302, 0.1148], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0115, 0.0120, 0.0123, 0.0113, 0.0096, 0.0096, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 22:27:50,675 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-03-26 22:28:03,084 INFO [finetune.py:976] (4/7) Epoch 18, batch 5350, loss[loss=0.1626, simple_loss=0.2389, pruned_loss=0.04313, over 4820.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2523, pruned_loss=0.05538, over 950497.41 frames. ], batch size: 30, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:28:06,217 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9742, 1.6942, 2.2890, 3.7439, 2.4977, 2.7227, 0.8274, 3.1364], + device='cuda:4'), covar=tensor([0.1565, 0.1415, 0.1336, 0.0573, 0.0741, 0.1482, 0.1867, 0.0412], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0115, 0.0133, 0.0163, 0.0100, 0.0134, 0.0123, 0.0098], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 22:28:25,304 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.195e+02 1.536e+02 1.872e+02 2.228e+02 4.473e+02, threshold=3.745e+02, percent-clipped=1.0 +2023-03-26 22:28:36,912 INFO [finetune.py:976] (4/7) Epoch 18, batch 5400, loss[loss=0.1938, simple_loss=0.2604, pruned_loss=0.06362, over 4823.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2484, pruned_loss=0.0538, over 950908.55 frames. ], batch size: 38, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:29:01,234 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.7762, 3.2791, 3.4656, 3.6135, 3.5294, 3.3654, 3.8482, 1.2577], + device='cuda:4'), covar=tensor([0.0876, 0.0884, 0.0843, 0.1135, 0.1342, 0.1410, 0.0843, 0.5312], + device='cuda:4'), in_proj_covar=tensor([0.0355, 0.0246, 0.0281, 0.0294, 0.0336, 0.0285, 0.0306, 0.0299], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:29:10,819 INFO [finetune.py:976] (4/7) Epoch 18, batch 5450, loss[loss=0.1728, simple_loss=0.2435, pruned_loss=0.05108, over 4864.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2455, pruned_loss=0.05307, over 949473.07 frames. ], batch size: 44, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:29:15,173 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2019, 2.2466, 1.4679, 2.4453, 2.2231, 1.8294, 3.0978, 2.2114], + device='cuda:4'), covar=tensor([0.1886, 0.2405, 0.4186, 0.3263, 0.3355, 0.2321, 0.2425, 0.2412], + device='cuda:4'), in_proj_covar=tensor([0.0183, 0.0187, 0.0233, 0.0251, 0.0244, 0.0201, 0.0212, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:29:30,962 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=102851.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:29:31,448 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.551e+01 1.511e+02 1.793e+02 2.102e+02 5.113e+02, threshold=3.586e+02, percent-clipped=1.0 +2023-03-26 22:29:40,204 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9184, 1.2968, 0.8225, 1.6932, 2.1753, 1.5098, 1.5817, 1.6155], + device='cuda:4'), covar=tensor([0.1287, 0.1974, 0.1957, 0.1112, 0.1717, 0.1926, 0.1351, 0.1888], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0092, 0.0120, 0.0093, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 22:29:44,409 INFO [finetune.py:976] (4/7) Epoch 18, batch 5500, loss[loss=0.1678, simple_loss=0.2392, pruned_loss=0.04816, over 4839.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2431, pruned_loss=0.05254, over 948631.48 frames. ], batch size: 47, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:29:54,068 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=102887.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:29:54,754 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-26 22:30:12,699 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=102912.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:30:18,101 INFO [finetune.py:976] (4/7) Epoch 18, batch 5550, loss[loss=0.204, simple_loss=0.271, pruned_loss=0.06848, over 4801.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2449, pruned_loss=0.05372, over 949056.17 frames. ], batch size: 45, lr: 3.32e-03, grad_scale: 32.0 +2023-03-26 22:30:36,211 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=102948.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:30:39,503 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.093e+02 1.559e+02 1.902e+02 2.213e+02 4.520e+02, threshold=3.805e+02, percent-clipped=3.0 +2023-03-26 22:30:50,054 INFO [finetune.py:976] (4/7) Epoch 18, batch 5600, loss[loss=0.2076, simple_loss=0.2854, pruned_loss=0.06485, over 4899.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2491, pruned_loss=0.05473, over 950214.74 frames. ], batch size: 36, lr: 3.32e-03, grad_scale: 16.0 +2023-03-26 22:31:42,257 INFO [finetune.py:976] (4/7) Epoch 18, batch 5650, loss[loss=0.1705, simple_loss=0.2367, pruned_loss=0.05217, over 4154.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2522, pruned_loss=0.05553, over 951548.66 frames. ], batch size: 18, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:31:44,116 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8541, 3.8492, 3.6644, 2.0276, 3.9698, 3.0364, 0.8789, 2.7049], + device='cuda:4'), covar=tensor([0.2240, 0.2056, 0.1448, 0.2989, 0.0980, 0.0839, 0.4203, 0.1435], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0175, 0.0158, 0.0128, 0.0159, 0.0122, 0.0147, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 22:31:45,377 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8178, 3.9127, 3.6775, 2.2007, 4.0577, 3.1494, 1.0158, 2.7308], + device='cuda:4'), covar=tensor([0.2353, 0.2140, 0.1722, 0.3224, 0.0973, 0.0892, 0.4648, 0.1471], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0175, 0.0159, 0.0128, 0.0159, 0.0122, 0.0147, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 22:32:09,152 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.800e+01 1.506e+02 1.835e+02 2.191e+02 3.638e+02, threshold=3.670e+02, percent-clipped=0.0 +2023-03-26 22:32:19,851 INFO [finetune.py:976] (4/7) Epoch 18, batch 5700, loss[loss=0.1373, simple_loss=0.1917, pruned_loss=0.04151, over 4003.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2473, pruned_loss=0.05427, over 929849.15 frames. ], batch size: 17, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:32:48,079 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103098.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:32:48,576 INFO [finetune.py:976] (4/7) Epoch 19, batch 0, loss[loss=0.2038, simple_loss=0.2604, pruned_loss=0.0736, over 4930.00 frames. ], tot_loss[loss=0.2038, simple_loss=0.2604, pruned_loss=0.0736, over 4930.00 frames. ], batch size: 41, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:32:48,576 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 22:33:04,745 INFO [finetune.py:1010] (4/7) Epoch 19, validation: loss=0.1586, simple_loss=0.2282, pruned_loss=0.04454, over 2265189.00 frames. +2023-03-26 22:33:04,745 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 22:33:25,842 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103132.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:33:33,040 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 22:33:38,062 INFO [finetune.py:976] (4/7) Epoch 19, batch 50, loss[loss=0.1913, simple_loss=0.2646, pruned_loss=0.05903, over 4903.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2462, pruned_loss=0.05161, over 217179.23 frames. ], batch size: 46, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:33:40,496 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.779e+01 1.456e+02 1.782e+02 2.150e+02 3.860e+02, threshold=3.565e+02, percent-clipped=1.0 +2023-03-26 22:33:41,937 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-26 22:33:44,742 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103159.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:33:50,939 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0942, 0.8728, 0.9154, 0.3506, 0.7670, 1.0346, 1.0144, 0.8709], + device='cuda:4'), covar=tensor([0.0797, 0.0521, 0.0481, 0.0581, 0.0675, 0.0483, 0.0353, 0.0554], + device='cuda:4'), in_proj_covar=tensor([0.0125, 0.0151, 0.0124, 0.0127, 0.0131, 0.0129, 0.0142, 0.0148], + device='cuda:4'), out_proj_covar=tensor([9.1448e-05, 1.0924e-04, 8.8782e-05, 9.0096e-05, 9.2173e-05, 9.2449e-05, + 1.0181e-04, 1.0649e-04], device='cuda:4') +2023-03-26 22:34:07,715 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103193.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:34:11,664 INFO [finetune.py:976] (4/7) Epoch 19, batch 100, loss[loss=0.1385, simple_loss=0.2001, pruned_loss=0.03847, over 4161.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2458, pruned_loss=0.05414, over 378799.59 frames. ], batch size: 65, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:34:17,537 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103207.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:34:41,204 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103243.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:34:45,884 INFO [finetune.py:976] (4/7) Epoch 19, batch 150, loss[loss=0.1887, simple_loss=0.2441, pruned_loss=0.06665, over 4704.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2417, pruned_loss=0.05283, over 505850.53 frames. ], batch size: 23, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:34:48,708 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.463e+02 1.787e+02 2.269e+02 3.542e+02, threshold=3.573e+02, percent-clipped=0.0 +2023-03-26 22:35:08,292 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7159, 1.7214, 2.0599, 1.9710, 1.8893, 3.1443, 1.5876, 1.7636], + device='cuda:4'), covar=tensor([0.0816, 0.1393, 0.1161, 0.0780, 0.1208, 0.0305, 0.1243, 0.1366], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 22:35:19,729 INFO [finetune.py:976] (4/7) Epoch 19, batch 200, loss[loss=0.1479, simple_loss=0.2194, pruned_loss=0.03817, over 4747.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.239, pruned_loss=0.05199, over 604155.57 frames. ], batch size: 23, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:35:22,737 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1581, 2.1653, 1.5871, 2.2365, 2.0835, 1.8423, 2.3972, 2.2103], + device='cuda:4'), covar=tensor([0.1347, 0.2144, 0.2959, 0.2422, 0.2637, 0.1726, 0.3291, 0.1761], + device='cuda:4'), in_proj_covar=tensor([0.0185, 0.0188, 0.0235, 0.0254, 0.0246, 0.0203, 0.0214, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:35:42,749 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7598, 1.2014, 0.9151, 1.6238, 2.0439, 1.3375, 1.4201, 1.6178], + device='cuda:4'), covar=tensor([0.1352, 0.2012, 0.1875, 0.1112, 0.1908, 0.1973, 0.1408, 0.1783], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0092, 0.0120, 0.0094, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 22:35:53,176 INFO [finetune.py:976] (4/7) Epoch 19, batch 250, loss[loss=0.1543, simple_loss=0.2426, pruned_loss=0.03302, over 4862.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2438, pruned_loss=0.05337, over 680392.98 frames. ], batch size: 44, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:35:56,518 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.389e+01 1.572e+02 1.886e+02 2.263e+02 4.128e+02, threshold=3.772e+02, percent-clipped=1.0 +2023-03-26 22:36:05,017 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103366.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 22:36:06,884 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5292, 1.4537, 1.6181, 1.7497, 1.7114, 3.0662, 1.2879, 1.5376], + device='cuda:4'), covar=tensor([0.1121, 0.2347, 0.1139, 0.1031, 0.1765, 0.0343, 0.2122, 0.2378], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0079, 0.0084, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 22:36:17,214 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.71 vs. limit=5.0 +2023-03-26 22:36:25,723 INFO [finetune.py:976] (4/7) Epoch 19, batch 300, loss[loss=0.1641, simple_loss=0.2376, pruned_loss=0.0453, over 4845.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2471, pruned_loss=0.05417, over 739958.39 frames. ], batch size: 30, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:37:02,202 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103427.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 22:37:21,218 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2904, 1.2389, 1.2394, 1.3512, 1.6037, 1.4887, 1.3180, 1.1935], + device='cuda:4'), covar=tensor([0.0338, 0.0306, 0.0571, 0.0266, 0.0226, 0.0396, 0.0322, 0.0393], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0105, 0.0139, 0.0108, 0.0097, 0.0106, 0.0097, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.2818e-05, 8.1198e-05, 1.0983e-04, 8.3082e-05, 7.5469e-05, 7.8029e-05, + 7.2140e-05, 8.2289e-05], device='cuda:4') +2023-03-26 22:37:21,696 INFO [finetune.py:976] (4/7) Epoch 19, batch 350, loss[loss=0.1967, simple_loss=0.2714, pruned_loss=0.06098, over 4895.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.251, pruned_loss=0.05568, over 787595.05 frames. ], batch size: 43, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:37:28,081 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.109e+02 1.554e+02 1.898e+02 2.403e+02 5.343e+02, threshold=3.796e+02, percent-clipped=4.0 +2023-03-26 22:37:29,796 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103454.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:38:03,042 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103488.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:38:04,301 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3799, 1.4489, 1.5389, 0.8780, 1.4501, 1.7679, 1.8523, 1.3331], + device='cuda:4'), covar=tensor([0.1061, 0.0768, 0.0582, 0.0660, 0.0526, 0.0674, 0.0304, 0.0876], + device='cuda:4'), in_proj_covar=tensor([0.0125, 0.0151, 0.0124, 0.0127, 0.0131, 0.0129, 0.0142, 0.0149], + device='cuda:4'), out_proj_covar=tensor([9.1611e-05, 1.0959e-04, 8.8976e-05, 9.0237e-05, 9.2332e-05, 9.2663e-05, + 1.0223e-04, 1.0671e-04], device='cuda:4') +2023-03-26 22:38:05,515 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0238, 2.0745, 1.8904, 1.8549, 2.5887, 2.6515, 2.0717, 2.0793], + device='cuda:4'), covar=tensor([0.0326, 0.0410, 0.0463, 0.0324, 0.0189, 0.0404, 0.0340, 0.0358], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0105, 0.0139, 0.0108, 0.0097, 0.0106, 0.0097, 0.0108], + device='cuda:4'), out_proj_covar=tensor([7.2698e-05, 8.1159e-05, 1.0978e-04, 8.3015e-05, 7.5424e-05, 7.7888e-05, + 7.2049e-05, 8.2335e-05], device='cuda:4') +2023-03-26 22:38:10,178 INFO [finetune.py:976] (4/7) Epoch 19, batch 400, loss[loss=0.1869, simple_loss=0.2545, pruned_loss=0.05966, over 4854.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.253, pruned_loss=0.05671, over 824048.53 frames. ], batch size: 31, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:38:16,128 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103507.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:38:39,537 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103543.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:38:43,097 INFO [finetune.py:976] (4/7) Epoch 19, batch 450, loss[loss=0.1821, simple_loss=0.2619, pruned_loss=0.05115, over 4883.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.252, pruned_loss=0.05626, over 853877.01 frames. ], batch size: 32, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:38:45,994 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.312e+01 1.502e+02 1.696e+02 2.061e+02 2.854e+02, threshold=3.392e+02, percent-clipped=0.0 +2023-03-26 22:38:47,239 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=103555.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:39:07,970 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-26 22:39:19,609 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=103591.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:39:24,374 INFO [finetune.py:976] (4/7) Epoch 19, batch 500, loss[loss=0.178, simple_loss=0.2445, pruned_loss=0.05576, over 4906.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2485, pruned_loss=0.05469, over 875598.83 frames. ], batch size: 35, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:39:50,573 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5058, 1.4777, 1.3824, 1.6317, 1.8475, 1.8397, 1.5523, 1.3903], + device='cuda:4'), covar=tensor([0.0365, 0.0320, 0.0563, 0.0284, 0.0195, 0.0372, 0.0316, 0.0371], + device='cuda:4'), in_proj_covar=tensor([0.0094, 0.0106, 0.0141, 0.0109, 0.0098, 0.0106, 0.0097, 0.0109], + device='cuda:4'), out_proj_covar=tensor([7.3120e-05, 8.1792e-05, 1.1067e-04, 8.3370e-05, 7.6084e-05, 7.8252e-05, + 7.2689e-05, 8.2959e-05], device='cuda:4') +2023-03-26 22:39:54,165 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9864, 1.3185, 0.7410, 1.9674, 2.4559, 1.7075, 1.6847, 1.9083], + device='cuda:4'), covar=tensor([0.1488, 0.2054, 0.2084, 0.1123, 0.1868, 0.2059, 0.1403, 0.1932], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0092, 0.0120, 0.0093, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 22:39:57,693 INFO [finetune.py:976] (4/7) Epoch 19, batch 550, loss[loss=0.1684, simple_loss=0.2294, pruned_loss=0.05366, over 4820.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2456, pruned_loss=0.05401, over 891492.23 frames. ], batch size: 30, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:40:00,581 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.093e+02 1.543e+02 1.834e+02 2.179e+02 4.966e+02, threshold=3.668e+02, percent-clipped=2.0 +2023-03-26 22:40:01,336 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1965, 2.3214, 2.1556, 2.4807, 2.8681, 2.3092, 2.3019, 1.8063], + device='cuda:4'), covar=tensor([0.2420, 0.1962, 0.1985, 0.1627, 0.1813, 0.1217, 0.2214, 0.2088], + device='cuda:4'), in_proj_covar=tensor([0.0242, 0.0208, 0.0213, 0.0192, 0.0241, 0.0186, 0.0215, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:40:31,342 INFO [finetune.py:976] (4/7) Epoch 19, batch 600, loss[loss=0.1819, simple_loss=0.2508, pruned_loss=0.05653, over 4831.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2449, pruned_loss=0.05329, over 907315.65 frames. ], batch size: 30, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:40:47,266 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2800, 3.7719, 3.9348, 4.1157, 4.0503, 3.8111, 4.3668, 1.4572], + device='cuda:4'), covar=tensor([0.0734, 0.0794, 0.0844, 0.0900, 0.1133, 0.1650, 0.0694, 0.5439], + device='cuda:4'), in_proj_covar=tensor([0.0345, 0.0241, 0.0277, 0.0288, 0.0330, 0.0280, 0.0298, 0.0293], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:40:47,269 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103722.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 22:40:58,705 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103740.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:41:04,045 INFO [finetune.py:976] (4/7) Epoch 19, batch 650, loss[loss=0.1789, simple_loss=0.2523, pruned_loss=0.0527, over 4928.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2505, pruned_loss=0.05521, over 917431.66 frames. ], batch size: 33, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:41:06,467 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.131e+02 1.539e+02 1.795e+02 2.169e+02 3.837e+02, threshold=3.591e+02, percent-clipped=1.0 +2023-03-26 22:41:07,210 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103754.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:41:15,353 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103765.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:41:23,069 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1418, 1.9272, 1.4615, 0.5403, 1.6518, 1.7537, 1.6785, 1.7364], + device='cuda:4'), covar=tensor([0.1036, 0.0931, 0.1655, 0.2130, 0.1471, 0.2484, 0.2253, 0.0927], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0193, 0.0199, 0.0183, 0.0211, 0.0207, 0.0222, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:41:30,849 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103788.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:41:37,417 INFO [finetune.py:976] (4/7) Epoch 19, batch 700, loss[loss=0.1635, simple_loss=0.2371, pruned_loss=0.04498, over 4756.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2519, pruned_loss=0.05571, over 921877.53 frames. ], batch size: 27, lr: 3.31e-03, grad_scale: 16.0 +2023-03-26 22:41:38,876 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103801.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:41:39,437 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=103802.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:42:02,036 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103826.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:42:13,879 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=103836.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:42:14,956 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.73 vs. limit=5.0 +2023-03-26 22:42:26,323 INFO [finetune.py:976] (4/7) Epoch 19, batch 750, loss[loss=0.1558, simple_loss=0.2464, pruned_loss=0.03261, over 4918.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2524, pruned_loss=0.05612, over 930165.98 frames. ], batch size: 38, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:42:33,271 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.573e+02 1.871e+02 2.192e+02 5.260e+02, threshold=3.742e+02, percent-clipped=2.0 +2023-03-26 22:43:03,183 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103876.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:43:22,083 INFO [finetune.py:976] (4/7) Epoch 19, batch 800, loss[loss=0.1823, simple_loss=0.2566, pruned_loss=0.05396, over 4892.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.2523, pruned_loss=0.05567, over 936521.25 frames. ], batch size: 43, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:43:34,182 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0538, 1.9754, 1.6454, 1.7813, 2.1050, 1.8073, 2.2067, 2.1107], + device='cuda:4'), covar=tensor([0.1361, 0.1935, 0.2933, 0.2508, 0.2396, 0.1684, 0.3119, 0.1780], + device='cuda:4'), in_proj_covar=tensor([0.0186, 0.0189, 0.0236, 0.0255, 0.0247, 0.0204, 0.0215, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:43:48,645 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103937.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 22:43:55,833 INFO [finetune.py:976] (4/7) Epoch 19, batch 850, loss[loss=0.1656, simple_loss=0.2324, pruned_loss=0.0494, over 4906.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2507, pruned_loss=0.05577, over 939130.22 frames. ], batch size: 35, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:43:58,238 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.499e+02 1.798e+02 2.103e+02 3.961e+02, threshold=3.597e+02, percent-clipped=1.0 +2023-03-26 22:44:23,829 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-26 22:44:25,966 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103990.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:44:30,231 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103997.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:44:31,323 INFO [finetune.py:976] (4/7) Epoch 19, batch 900, loss[loss=0.1966, simple_loss=0.2577, pruned_loss=0.06777, over 4825.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2487, pruned_loss=0.05535, over 943744.85 frames. ], batch size: 38, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:44:45,511 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.1959, 4.7899, 4.5095, 2.8002, 4.7964, 3.6998, 0.9885, 3.5228], + device='cuda:4'), covar=tensor([0.2119, 0.1516, 0.1379, 0.2812, 0.0806, 0.0896, 0.4787, 0.1308], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0179, 0.0162, 0.0130, 0.0162, 0.0125, 0.0150, 0.0125], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 22:44:46,732 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104022.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 22:44:47,543 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.94 vs. limit=5.0 +2023-03-26 22:44:50,162 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0993, 2.0591, 2.0598, 1.3733, 2.0292, 2.0255, 2.1588, 1.6938], + device='cuda:4'), covar=tensor([0.0515, 0.0523, 0.0605, 0.0862, 0.0612, 0.0694, 0.0544, 0.1118], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0133, 0.0139, 0.0120, 0.0124, 0.0137, 0.0139, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:45:05,988 INFO [finetune.py:976] (4/7) Epoch 19, batch 950, loss[loss=0.154, simple_loss=0.2322, pruned_loss=0.03788, over 4791.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2452, pruned_loss=0.05365, over 945881.61 frames. ], batch size: 29, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:45:06,091 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=104049.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:45:07,316 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=104051.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:45:08,390 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.542e+01 1.532e+02 1.748e+02 2.079e+02 4.067e+02, threshold=3.497e+02, percent-clipped=1.0 +2023-03-26 22:45:11,543 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=104058.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:45:18,740 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104070.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 22:45:36,948 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104096.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:45:38,700 INFO [finetune.py:976] (4/7) Epoch 19, batch 1000, loss[loss=0.1759, simple_loss=0.2524, pruned_loss=0.04973, over 4740.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2477, pruned_loss=0.05426, over 948480.84 frames. ], batch size: 54, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:45:45,434 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=104110.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:45:52,086 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104121.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:46:12,345 INFO [finetune.py:976] (4/7) Epoch 19, batch 1050, loss[loss=0.2151, simple_loss=0.2882, pruned_loss=0.07097, over 4899.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2494, pruned_loss=0.05424, over 949656.45 frames. ], batch size: 37, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:46:14,761 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.111e+02 1.591e+02 1.940e+02 2.273e+02 3.456e+02, threshold=3.881e+02, percent-clipped=0.0 +2023-03-26 22:46:53,935 INFO [finetune.py:976] (4/7) Epoch 19, batch 1100, loss[loss=0.1636, simple_loss=0.2394, pruned_loss=0.04395, over 4839.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2513, pruned_loss=0.05513, over 951386.28 frames. ], batch size: 49, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:47:16,924 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104232.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 22:47:29,946 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3272, 2.2228, 2.1922, 1.4935, 2.1706, 2.2058, 2.3862, 1.7260], + device='cuda:4'), covar=tensor([0.0591, 0.0607, 0.0697, 0.0915, 0.0687, 0.0757, 0.0584, 0.1188], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0133, 0.0138, 0.0119, 0.0123, 0.0136, 0.0139, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:47:36,262 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5893, 1.5557, 1.5045, 1.5788, 0.9915, 3.0022, 1.1079, 1.4821], + device='cuda:4'), covar=tensor([0.3485, 0.2609, 0.2316, 0.2540, 0.1893, 0.0251, 0.2867, 0.1389], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0116, 0.0120, 0.0123, 0.0114, 0.0096, 0.0096, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 22:47:39,157 INFO [finetune.py:976] (4/7) Epoch 19, batch 1150, loss[loss=0.1658, simple_loss=0.2255, pruned_loss=0.05303, over 4472.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2529, pruned_loss=0.05647, over 950651.02 frames. ], batch size: 19, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:47:40,959 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1515, 2.0056, 2.9151, 4.3321, 2.9528, 2.7877, 0.7770, 3.5997], + device='cuda:4'), covar=tensor([0.1644, 0.1461, 0.1192, 0.0433, 0.0702, 0.1321, 0.2164, 0.0386], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0117, 0.0133, 0.0164, 0.0099, 0.0136, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 22:47:47,026 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.710e+02 1.992e+02 2.366e+02 4.129e+02, threshold=3.984e+02, percent-clipped=1.0 +2023-03-26 22:48:25,526 INFO [finetune.py:976] (4/7) Epoch 19, batch 1200, loss[loss=0.1499, simple_loss=0.2207, pruned_loss=0.03955, over 4885.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2503, pruned_loss=0.05488, over 951159.55 frames. ], batch size: 32, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:48:42,443 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5027, 1.0228, 0.7995, 1.4239, 2.0148, 0.7436, 1.3080, 1.3947], + device='cuda:4'), covar=tensor([0.1596, 0.2210, 0.1697, 0.1170, 0.1882, 0.1899, 0.1504, 0.1942], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0096, 0.0111, 0.0092, 0.0120, 0.0094, 0.0100, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 22:48:50,431 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2483, 2.3210, 1.8533, 2.3745, 2.2604, 2.1283, 2.1966, 2.9956], + device='cuda:4'), covar=tensor([0.3950, 0.4362, 0.3539, 0.4312, 0.4229, 0.2700, 0.4183, 0.1894], + device='cuda:4'), in_proj_covar=tensor([0.0287, 0.0261, 0.0230, 0.0276, 0.0251, 0.0220, 0.0252, 0.0232], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:49:05,635 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104346.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:49:06,886 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4124, 2.2795, 2.8063, 1.7403, 2.5212, 2.7741, 2.0390, 2.9154], + device='cuda:4'), covar=tensor([0.1157, 0.1616, 0.1370, 0.2016, 0.0842, 0.1159, 0.2404, 0.0731], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0204, 0.0190, 0.0188, 0.0175, 0.0213, 0.0217, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:49:07,374 INFO [finetune.py:976] (4/7) Epoch 19, batch 1250, loss[loss=0.1698, simple_loss=0.241, pruned_loss=0.0493, over 4917.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2484, pruned_loss=0.05414, over 951353.95 frames. ], batch size: 37, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:49:10,322 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.979e+01 1.472e+02 1.754e+02 2.218e+02 4.171e+02, threshold=3.509e+02, percent-clipped=1.0 +2023-03-26 22:49:10,411 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104353.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:49:29,610 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 22:49:39,161 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104396.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:49:41,418 INFO [finetune.py:976] (4/7) Epoch 19, batch 1300, loss[loss=0.1969, simple_loss=0.249, pruned_loss=0.07243, over 4893.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2459, pruned_loss=0.05352, over 953965.67 frames. ], batch size: 35, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:49:45,745 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104405.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:49:56,435 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104421.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:50:06,064 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6668, 3.7726, 3.5984, 1.9744, 3.9050, 2.9552, 1.0344, 2.6546], + device='cuda:4'), covar=tensor([0.2545, 0.2057, 0.1490, 0.3100, 0.0899, 0.0999, 0.4168, 0.1463], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0177, 0.0161, 0.0129, 0.0161, 0.0124, 0.0149, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 22:50:10,346 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104444.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:50:14,762 INFO [finetune.py:976] (4/7) Epoch 19, batch 1350, loss[loss=0.1947, simple_loss=0.2776, pruned_loss=0.05591, over 4916.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2464, pruned_loss=0.05391, over 954285.51 frames. ], batch size: 37, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:50:17,642 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.617e+02 1.931e+02 2.310e+02 3.973e+02, threshold=3.863e+02, percent-clipped=4.0 +2023-03-26 22:50:29,051 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104469.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:50:48,510 INFO [finetune.py:976] (4/7) Epoch 19, batch 1400, loss[loss=0.1669, simple_loss=0.2417, pruned_loss=0.0461, over 4897.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.25, pruned_loss=0.05524, over 954581.96 frames. ], batch size: 35, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:50:57,984 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1947, 2.0863, 2.1619, 1.5041, 2.1388, 2.1422, 2.2490, 1.7208], + device='cuda:4'), covar=tensor([0.0557, 0.0643, 0.0653, 0.0883, 0.0662, 0.0695, 0.0561, 0.1112], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0133, 0.0138, 0.0118, 0.0123, 0.0136, 0.0138, 0.0159], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:51:10,558 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104532.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 22:51:12,565 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-26 22:51:21,268 INFO [finetune.py:976] (4/7) Epoch 19, batch 1450, loss[loss=0.1847, simple_loss=0.2523, pruned_loss=0.05858, over 4825.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2522, pruned_loss=0.05626, over 954941.85 frames. ], batch size: 30, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:51:24,644 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.650e+02 1.913e+02 2.290e+02 4.485e+02, threshold=3.826e+02, percent-clipped=3.0 +2023-03-26 22:51:42,313 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.0558, 3.4935, 3.7028, 3.9314, 3.8205, 3.6021, 4.1100, 1.3052], + device='cuda:4'), covar=tensor([0.0744, 0.0844, 0.0803, 0.0864, 0.1144, 0.1443, 0.0736, 0.5346], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0241, 0.0279, 0.0289, 0.0331, 0.0280, 0.0301, 0.0294], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:51:42,883 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104580.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:52:02,826 INFO [finetune.py:976] (4/7) Epoch 19, batch 1500, loss[loss=0.1602, simple_loss=0.232, pruned_loss=0.04418, over 4917.00 frames. ], tot_loss[loss=0.184, simple_loss=0.254, pruned_loss=0.05704, over 956543.31 frames. ], batch size: 42, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:52:09,497 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8228, 1.7011, 2.1632, 3.6216, 2.4406, 2.5233, 0.9658, 2.9541], + device='cuda:4'), covar=tensor([0.1628, 0.1381, 0.1386, 0.0512, 0.0722, 0.1303, 0.1896, 0.0467], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0115, 0.0132, 0.0162, 0.0098, 0.0135, 0.0122, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 22:52:22,438 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4498, 1.4011, 1.6665, 2.5338, 1.6987, 2.1377, 1.0161, 2.1813], + device='cuda:4'), covar=tensor([0.1694, 0.1328, 0.1072, 0.0674, 0.0856, 0.1203, 0.1453, 0.0562], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0115, 0.0131, 0.0162, 0.0098, 0.0134, 0.0122, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 22:52:33,826 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104646.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:52:35,538 INFO [finetune.py:976] (4/7) Epoch 19, batch 1550, loss[loss=0.166, simple_loss=0.2344, pruned_loss=0.04883, over 4789.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2536, pruned_loss=0.05684, over 957219.19 frames. ], batch size: 25, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:52:40,199 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.878e+01 1.494e+02 1.864e+02 2.283e+02 3.386e+02, threshold=3.728e+02, percent-clipped=0.0 +2023-03-26 22:52:40,302 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104653.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:53:16,272 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0369, 1.9295, 1.6534, 1.9093, 1.9784, 1.7242, 2.1655, 1.9849], + device='cuda:4'), covar=tensor([0.1235, 0.1860, 0.2699, 0.2301, 0.2497, 0.1593, 0.2820, 0.1677], + device='cuda:4'), in_proj_covar=tensor([0.0185, 0.0187, 0.0234, 0.0253, 0.0245, 0.0203, 0.0213, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:53:26,173 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104694.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:53:29,185 INFO [finetune.py:976] (4/7) Epoch 19, batch 1600, loss[loss=0.1552, simple_loss=0.2295, pruned_loss=0.04042, over 4227.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2502, pruned_loss=0.05535, over 956094.63 frames. ], batch size: 18, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:53:35,259 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104701.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:53:38,276 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104705.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:54:11,339 INFO [finetune.py:976] (4/7) Epoch 19, batch 1650, loss[loss=0.1602, simple_loss=0.2285, pruned_loss=0.04599, over 4827.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.248, pruned_loss=0.05454, over 956680.89 frames. ], batch size: 38, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:54:13,774 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.513e+02 1.754e+02 2.121e+02 3.523e+02, threshold=3.508e+02, percent-clipped=0.0 +2023-03-26 22:54:13,853 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=104753.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 22:54:44,696 INFO [finetune.py:976] (4/7) Epoch 19, batch 1700, loss[loss=0.2123, simple_loss=0.2719, pruned_loss=0.07634, over 4901.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2455, pruned_loss=0.05352, over 958319.15 frames. ], batch size: 37, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:54:50,352 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7927, 1.1894, 0.7957, 1.5995, 2.0090, 1.4579, 1.4629, 1.5927], + device='cuda:4'), covar=tensor([0.1477, 0.2182, 0.2040, 0.1162, 0.2053, 0.2021, 0.1482, 0.2038], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0096, 0.0110, 0.0092, 0.0120, 0.0094, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 22:54:55,341 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.2377, 2.9274, 3.0318, 3.1703, 3.0223, 2.8447, 3.3022, 1.0093], + device='cuda:4'), covar=tensor([0.1189, 0.0991, 0.1145, 0.1407, 0.1701, 0.1857, 0.1126, 0.5390], + device='cuda:4'), in_proj_covar=tensor([0.0348, 0.0241, 0.0279, 0.0290, 0.0331, 0.0282, 0.0301, 0.0295], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:55:00,573 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9251, 1.4860, 2.3727, 1.4269, 2.0614, 2.2354, 1.5248, 2.3354], + device='cuda:4'), covar=tensor([0.1483, 0.2600, 0.1423, 0.2149, 0.1011, 0.1473, 0.3296, 0.0906], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0204, 0.0190, 0.0189, 0.0174, 0.0213, 0.0217, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 22:55:17,899 INFO [finetune.py:976] (4/7) Epoch 19, batch 1750, loss[loss=0.223, simple_loss=0.2915, pruned_loss=0.07724, over 4905.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2481, pruned_loss=0.05479, over 957700.00 frames. ], batch size: 43, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:55:20,305 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.630e+02 1.888e+02 2.368e+02 5.925e+02, threshold=3.776e+02, percent-clipped=5.0 +2023-03-26 22:55:51,605 INFO [finetune.py:976] (4/7) Epoch 19, batch 1800, loss[loss=0.1928, simple_loss=0.2691, pruned_loss=0.05823, over 4902.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2483, pruned_loss=0.05429, over 956792.24 frames. ], batch size: 36, lr: 3.30e-03, grad_scale: 16.0 +2023-03-26 22:56:25,142 INFO [finetune.py:976] (4/7) Epoch 19, batch 1850, loss[loss=0.1708, simple_loss=0.2354, pruned_loss=0.05313, over 4734.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2507, pruned_loss=0.05542, over 955235.09 frames. ], batch size: 54, lr: 3.30e-03, grad_scale: 32.0 +2023-03-26 22:56:27,539 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.864e+01 1.560e+02 1.785e+02 2.312e+02 4.235e+02, threshold=3.569e+02, percent-clipped=1.0 +2023-03-26 22:57:00,527 INFO [finetune.py:976] (4/7) Epoch 19, batch 1900, loss[loss=0.1725, simple_loss=0.2477, pruned_loss=0.04866, over 4784.00 frames. ], tot_loss[loss=0.1826, simple_loss=0.2529, pruned_loss=0.05619, over 957296.42 frames. ], batch size: 25, lr: 3.30e-03, grad_scale: 32.0 +2023-03-26 22:57:42,256 INFO [finetune.py:976] (4/7) Epoch 19, batch 1950, loss[loss=0.1569, simple_loss=0.2318, pruned_loss=0.04096, over 4848.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2506, pruned_loss=0.05478, over 958576.60 frames. ], batch size: 49, lr: 3.30e-03, grad_scale: 32.0 +2023-03-26 22:57:44,664 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.372e+01 1.430e+02 1.759e+02 2.099e+02 5.293e+02, threshold=3.517e+02, percent-clipped=3.0 +2023-03-26 22:58:31,191 INFO [finetune.py:976] (4/7) Epoch 19, batch 2000, loss[loss=0.2096, simple_loss=0.2735, pruned_loss=0.07284, over 4760.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2474, pruned_loss=0.0535, over 958172.70 frames. ], batch size: 54, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 22:58:57,801 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-26 22:59:17,044 INFO [finetune.py:976] (4/7) Epoch 19, batch 2050, loss[loss=0.2045, simple_loss=0.252, pruned_loss=0.07848, over 4045.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2454, pruned_loss=0.05326, over 956009.51 frames. ], batch size: 65, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 22:59:19,899 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.050e+02 1.557e+02 1.803e+02 2.317e+02 4.729e+02, threshold=3.605e+02, percent-clipped=3.0 +2023-03-26 22:59:50,002 INFO [finetune.py:976] (4/7) Epoch 19, batch 2100, loss[loss=0.2082, simple_loss=0.2762, pruned_loss=0.07005, over 4923.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2462, pruned_loss=0.05409, over 955022.60 frames. ], batch size: 42, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:00:23,802 INFO [finetune.py:976] (4/7) Epoch 19, batch 2150, loss[loss=0.3073, simple_loss=0.3617, pruned_loss=0.1264, over 4859.00 frames. ], tot_loss[loss=0.181, simple_loss=0.25, pruned_loss=0.05601, over 951705.37 frames. ], batch size: 44, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:00:26,653 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.573e+02 1.937e+02 2.406e+02 5.182e+02, threshold=3.875e+02, percent-clipped=4.0 +2023-03-26 23:00:57,385 INFO [finetune.py:976] (4/7) Epoch 19, batch 2200, loss[loss=0.1942, simple_loss=0.2675, pruned_loss=0.06043, over 4787.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2524, pruned_loss=0.05669, over 952631.11 frames. ], batch size: 51, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:01:08,749 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4962, 1.1019, 0.8159, 1.3576, 1.8967, 0.7112, 1.2835, 1.3509], + device='cuda:4'), covar=tensor([0.1564, 0.2141, 0.1735, 0.1170, 0.2003, 0.1938, 0.1501, 0.1957], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0091, 0.0119, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 23:01:10,597 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6887, 1.5532, 1.1482, 0.3063, 1.2277, 1.4985, 1.4430, 1.3723], + device='cuda:4'), covar=tensor([0.0949, 0.0855, 0.1369, 0.2015, 0.1533, 0.2391, 0.2291, 0.0934], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0193, 0.0200, 0.0182, 0.0210, 0.0207, 0.0222, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:01:30,612 INFO [finetune.py:976] (4/7) Epoch 19, batch 2250, loss[loss=0.1532, simple_loss=0.2386, pruned_loss=0.03389, over 4766.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2533, pruned_loss=0.05691, over 953039.63 frames. ], batch size: 28, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:01:33,460 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.177e+02 1.629e+02 1.918e+02 2.372e+02 6.301e+02, threshold=3.835e+02, percent-clipped=3.0 +2023-03-26 23:01:33,686 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-26 23:01:56,202 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105388.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:02:03,217 INFO [finetune.py:976] (4/7) Epoch 19, batch 2300, loss[loss=0.2189, simple_loss=0.2781, pruned_loss=0.07985, over 4809.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.2534, pruned_loss=0.05663, over 954454.17 frames. ], batch size: 40, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:02:45,812 INFO [finetune.py:976] (4/7) Epoch 19, batch 2350, loss[loss=0.1707, simple_loss=0.2478, pruned_loss=0.0468, over 4804.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2517, pruned_loss=0.05624, over 953345.41 frames. ], batch size: 41, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:02:45,952 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105449.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 23:02:48,224 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.493e+02 1.723e+02 2.054e+02 4.367e+02, threshold=3.447e+02, percent-clipped=1.0 +2023-03-26 23:03:00,628 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3581, 2.3290, 1.8493, 2.2629, 2.2514, 1.9818, 2.6422, 2.3260], + device='cuda:4'), covar=tensor([0.1332, 0.1835, 0.2984, 0.2459, 0.2476, 0.1669, 0.2806, 0.1890], + device='cuda:4'), in_proj_covar=tensor([0.0185, 0.0188, 0.0236, 0.0253, 0.0247, 0.0203, 0.0215, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:03:10,223 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105486.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:03:19,411 INFO [finetune.py:976] (4/7) Epoch 19, batch 2400, loss[loss=0.1676, simple_loss=0.2254, pruned_loss=0.05488, over 4766.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2481, pruned_loss=0.05485, over 954763.51 frames. ], batch size: 26, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:03:23,369 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105502.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:03:46,226 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 23:04:14,887 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105547.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:04:15,985 INFO [finetune.py:976] (4/7) Epoch 19, batch 2450, loss[loss=0.1885, simple_loss=0.2461, pruned_loss=0.06548, over 4199.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.246, pruned_loss=0.05433, over 955548.22 frames. ], batch size: 65, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:04:18,404 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.415e+01 1.457e+02 1.742e+02 2.171e+02 6.143e+02, threshold=3.484e+02, percent-clipped=3.0 +2023-03-26 23:04:25,635 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105563.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:04:42,066 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.68 vs. limit=5.0 +2023-03-26 23:04:49,908 INFO [finetune.py:976] (4/7) Epoch 19, batch 2500, loss[loss=0.2559, simple_loss=0.319, pruned_loss=0.09642, over 4741.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2464, pruned_loss=0.05466, over 954795.62 frames. ], batch size: 59, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:05:06,293 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 23:05:09,112 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6827, 1.6301, 2.2592, 3.4776, 2.2708, 2.5460, 0.9307, 2.8760], + device='cuda:4'), covar=tensor([0.1701, 0.1437, 0.1294, 0.0599, 0.0828, 0.1232, 0.1945, 0.0445], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0162, 0.0099, 0.0134, 0.0123, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 23:05:22,453 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8151, 1.5898, 2.1228, 1.3683, 2.0194, 2.1013, 1.5010, 2.1702], + device='cuda:4'), covar=tensor([0.1461, 0.2429, 0.1339, 0.2101, 0.0915, 0.1442, 0.3152, 0.0886], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0204, 0.0190, 0.0187, 0.0173, 0.0212, 0.0216, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:05:23,461 INFO [finetune.py:976] (4/7) Epoch 19, batch 2550, loss[loss=0.1755, simple_loss=0.2547, pruned_loss=0.04817, over 4909.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2512, pruned_loss=0.05577, over 956113.31 frames. ], batch size: 36, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:05:26,383 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.162e+02 1.575e+02 1.924e+02 2.412e+02 4.379e+02, threshold=3.848e+02, percent-clipped=4.0 +2023-03-26 23:05:26,531 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2383, 2.1050, 2.3322, 0.8861, 2.5072, 2.8505, 2.3089, 1.9983], + device='cuda:4'), covar=tensor([0.0971, 0.0683, 0.0439, 0.0793, 0.0469, 0.0486, 0.0459, 0.0685], + device='cuda:4'), in_proj_covar=tensor([0.0125, 0.0152, 0.0126, 0.0127, 0.0132, 0.0130, 0.0142, 0.0149], + device='cuda:4'), out_proj_covar=tensor([9.1696e-05, 1.1021e-04, 8.9824e-05, 9.0168e-05, 9.3282e-05, 9.3078e-05, + 1.0230e-04, 1.0701e-04], device='cuda:4') +2023-03-26 23:05:56,905 INFO [finetune.py:976] (4/7) Epoch 19, batch 2600, loss[loss=0.212, simple_loss=0.2854, pruned_loss=0.06928, over 4837.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2518, pruned_loss=0.05536, over 953471.54 frames. ], batch size: 47, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:06:12,848 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8981, 1.9430, 1.7908, 1.9139, 1.8104, 4.7110, 1.9112, 2.2675], + device='cuda:4'), covar=tensor([0.3343, 0.2523, 0.2131, 0.2413, 0.1550, 0.0137, 0.2383, 0.1224], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0114, 0.0119, 0.0122, 0.0113, 0.0095, 0.0095, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 23:06:27,125 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=105744.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 23:06:30,093 INFO [finetune.py:976] (4/7) Epoch 19, batch 2650, loss[loss=0.1839, simple_loss=0.2592, pruned_loss=0.05429, over 4864.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2531, pruned_loss=0.05561, over 954555.61 frames. ], batch size: 31, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:06:32,907 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.372e+01 1.545e+02 1.903e+02 2.181e+02 3.189e+02, threshold=3.806e+02, percent-clipped=0.0 +2023-03-26 23:06:59,521 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105792.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:07:03,678 INFO [finetune.py:976] (4/7) Epoch 19, batch 2700, loss[loss=0.1549, simple_loss=0.2326, pruned_loss=0.03862, over 4754.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2507, pruned_loss=0.05417, over 953732.99 frames. ], batch size: 27, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:07:03,782 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105799.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:07:21,495 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5083, 1.5840, 2.0960, 3.0904, 2.0485, 2.2315, 1.0331, 2.5536], + device='cuda:4'), covar=tensor([0.1654, 0.1287, 0.1102, 0.0530, 0.0767, 0.1466, 0.1678, 0.0510], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0116, 0.0133, 0.0163, 0.0099, 0.0135, 0.0123, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 23:07:33,452 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=105842.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:07:33,511 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.7463, 1.6317, 1.8394, 1.0059, 1.8831, 2.0402, 1.9200, 1.5904], + device='cuda:4'), covar=tensor([0.0856, 0.0758, 0.0486, 0.0572, 0.0458, 0.0531, 0.0377, 0.0656], + device='cuda:4'), in_proj_covar=tensor([0.0125, 0.0153, 0.0126, 0.0127, 0.0133, 0.0130, 0.0143, 0.0149], + device='cuda:4'), out_proj_covar=tensor([9.1745e-05, 1.1062e-04, 9.0066e-05, 9.0111e-05, 9.3699e-05, 9.3193e-05, + 1.0271e-04, 1.0696e-04], device='cuda:4') +2023-03-26 23:07:37,642 INFO [finetune.py:976] (4/7) Epoch 19, batch 2750, loss[loss=0.1969, simple_loss=0.2588, pruned_loss=0.06753, over 4759.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2476, pruned_loss=0.05345, over 951875.58 frames. ], batch size: 28, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:07:40,096 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.803e+01 1.395e+02 1.671e+02 1.966e+02 3.086e+02, threshold=3.343e+02, percent-clipped=0.0 +2023-03-26 23:07:40,245 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105853.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:07:43,733 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=105858.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:07:45,017 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105860.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:08:22,329 INFO [finetune.py:976] (4/7) Epoch 19, batch 2800, loss[loss=0.1664, simple_loss=0.2375, pruned_loss=0.04766, over 4836.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2454, pruned_loss=0.05287, over 953154.63 frames. ], batch size: 33, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:09:03,615 INFO [finetune.py:976] (4/7) Epoch 19, batch 2850, loss[loss=0.181, simple_loss=0.2502, pruned_loss=0.05587, over 4706.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2451, pruned_loss=0.05328, over 952071.32 frames. ], batch size: 23, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:09:10,695 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.025e+01 1.444e+02 1.775e+02 2.176e+02 4.047e+02, threshold=3.549e+02, percent-clipped=5.0 +2023-03-26 23:09:49,338 INFO [finetune.py:976] (4/7) Epoch 19, batch 2900, loss[loss=0.1913, simple_loss=0.2805, pruned_loss=0.05108, over 4906.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2475, pruned_loss=0.0538, over 952685.92 frames. ], batch size: 43, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:10:20,992 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106044.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:10:24,380 INFO [finetune.py:976] (4/7) Epoch 19, batch 2950, loss[loss=0.1936, simple_loss=0.2621, pruned_loss=0.06258, over 4883.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2507, pruned_loss=0.05437, over 953675.74 frames. ], batch size: 32, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:10:27,319 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.599e+02 1.865e+02 2.251e+02 4.962e+02, threshold=3.729e+02, percent-clipped=1.0 +2023-03-26 23:10:43,300 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=106077.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:10:53,298 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106092.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:10:57,087 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.68 vs. limit=2.0 +2023-03-26 23:10:57,520 INFO [finetune.py:976] (4/7) Epoch 19, batch 3000, loss[loss=0.1761, simple_loss=0.2503, pruned_loss=0.0509, over 4891.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2514, pruned_loss=0.05401, over 953565.94 frames. ], batch size: 35, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:10:57,520 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 23:11:08,360 INFO [finetune.py:1010] (4/7) Epoch 19, validation: loss=0.1576, simple_loss=0.2259, pruned_loss=0.04462, over 2265189.00 frames. +2023-03-26 23:11:08,361 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 23:11:35,493 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-26 23:11:43,857 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106138.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:11:46,232 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106142.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:11:50,366 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106148.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:11:50,874 INFO [finetune.py:976] (4/7) Epoch 19, batch 3050, loss[loss=0.1782, simple_loss=0.2484, pruned_loss=0.05405, over 4856.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2516, pruned_loss=0.05371, over 954966.21 frames. ], batch size: 44, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:11:52,731 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5496, 1.4507, 1.1630, 1.2710, 1.7707, 1.7928, 1.5271, 1.2802], + device='cuda:4'), covar=tensor([0.0343, 0.0352, 0.0794, 0.0399, 0.0264, 0.0430, 0.0385, 0.0437], + device='cuda:4'), in_proj_covar=tensor([0.0096, 0.0107, 0.0142, 0.0110, 0.0099, 0.0110, 0.0100, 0.0110], + device='cuda:4'), out_proj_covar=tensor([7.4468e-05, 8.2351e-05, 1.1200e-04, 8.4478e-05, 7.7048e-05, 8.0984e-05, + 7.4438e-05, 8.4336e-05], device='cuda:4') +2023-03-26 23:11:53,807 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.036e+02 1.577e+02 1.927e+02 2.196e+02 3.458e+02, threshold=3.854e+02, percent-clipped=0.0 +2023-03-26 23:11:55,079 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106155.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:11:57,415 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106158.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:12:18,598 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106190.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:12:22,869 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.2156, 4.5366, 4.7669, 5.0345, 4.9319, 4.6610, 5.3787, 1.6129], + device='cuda:4'), covar=tensor([0.0691, 0.0842, 0.0697, 0.0870, 0.1127, 0.1382, 0.0509, 0.5672], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0242, 0.0278, 0.0290, 0.0330, 0.0281, 0.0301, 0.0293], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:12:24,009 INFO [finetune.py:976] (4/7) Epoch 19, batch 3100, loss[loss=0.1764, simple_loss=0.2455, pruned_loss=0.05371, over 4927.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2496, pruned_loss=0.05298, over 957594.61 frames. ], batch size: 33, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:12:28,995 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.40 vs. limit=5.0 +2023-03-26 23:12:29,245 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106206.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:12:37,246 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-26 23:12:51,437 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-26 23:12:54,203 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8982, 1.0846, 1.8743, 1.8296, 1.6501, 1.6288, 1.6944, 1.7951], + device='cuda:4'), covar=tensor([0.3953, 0.3765, 0.3195, 0.3541, 0.4632, 0.3723, 0.4163, 0.3073], + device='cuda:4'), in_proj_covar=tensor([0.0251, 0.0240, 0.0260, 0.0278, 0.0277, 0.0251, 0.0285, 0.0242], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:12:57,681 INFO [finetune.py:976] (4/7) Epoch 19, batch 3150, loss[loss=0.1645, simple_loss=0.2329, pruned_loss=0.04808, over 4831.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2467, pruned_loss=0.05242, over 955889.58 frames. ], batch size: 30, lr: 3.29e-03, grad_scale: 32.0 +2023-03-26 23:12:58,391 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7881, 1.3055, 1.0463, 1.6397, 2.0130, 1.4387, 1.4916, 1.6978], + device='cuda:4'), covar=tensor([0.1375, 0.1959, 0.1854, 0.1111, 0.1969, 0.1877, 0.1379, 0.1845], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0110, 0.0091, 0.0119, 0.0092, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 23:13:00,116 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.828e+01 1.675e+02 1.879e+02 2.192e+02 3.916e+02, threshold=3.758e+02, percent-clipped=1.0 +2023-03-26 23:13:20,396 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=106268.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:13:23,670 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 23:13:41,053 INFO [finetune.py:976] (4/7) Epoch 19, batch 3200, loss[loss=0.1858, simple_loss=0.2557, pruned_loss=0.05797, over 4862.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2433, pruned_loss=0.05149, over 953898.75 frames. ], batch size: 44, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:14:01,430 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106329.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:14:05,079 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3735, 1.4523, 1.8389, 1.7622, 1.5814, 3.2698, 1.3799, 1.5456], + device='cuda:4'), covar=tensor([0.0916, 0.1697, 0.1067, 0.0868, 0.1472, 0.0236, 0.1420, 0.1624], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0080, 0.0084, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 23:14:05,719 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=106336.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:14:16,902 INFO [finetune.py:976] (4/7) Epoch 19, batch 3250, loss[loss=0.1722, simple_loss=0.2481, pruned_loss=0.04817, over 4897.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2445, pruned_loss=0.05224, over 955649.05 frames. ], batch size: 35, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:14:24,769 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.520e+02 1.839e+02 2.222e+02 4.428e+02, threshold=3.677e+02, percent-clipped=2.0 +2023-03-26 23:15:06,017 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-26 23:15:08,379 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106397.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:15:10,071 INFO [finetune.py:976] (4/7) Epoch 19, batch 3300, loss[loss=0.1604, simple_loss=0.2435, pruned_loss=0.03866, over 4792.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.249, pruned_loss=0.05344, over 955622.54 frames. ], batch size: 29, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:15:20,387 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-26 23:15:39,031 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-03-26 23:15:41,462 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106433.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:15:51,025 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106448.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:15:51,548 INFO [finetune.py:976] (4/7) Epoch 19, batch 3350, loss[loss=0.1446, simple_loss=0.2159, pruned_loss=0.03667, over 4619.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2499, pruned_loss=0.05438, over 952761.09 frames. ], batch size: 20, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:15:54,457 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.049e+02 1.616e+02 1.883e+02 2.222e+02 4.657e+02, threshold=3.766e+02, percent-clipped=2.0 +2023-03-26 23:15:55,187 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=106454.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:15:56,222 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106455.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:16:31,244 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106496.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:16:33,552 INFO [finetune.py:976] (4/7) Epoch 19, batch 3400, loss[loss=0.1737, simple_loss=0.2521, pruned_loss=0.04764, over 4791.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.252, pruned_loss=0.05516, over 954316.59 frames. ], batch size: 29, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:16:36,075 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106503.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:16:44,436 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106515.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:16:47,520 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0752, 1.9934, 1.6518, 1.8349, 2.0574, 1.7926, 2.1843, 2.0592], + device='cuda:4'), covar=tensor([0.1346, 0.1890, 0.2837, 0.2353, 0.2523, 0.1743, 0.2645, 0.1642], + device='cuda:4'), in_proj_covar=tensor([0.0184, 0.0187, 0.0234, 0.0252, 0.0246, 0.0202, 0.0215, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:16:51,258 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-26 23:17:04,559 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-26 23:17:06,753 INFO [finetune.py:976] (4/7) Epoch 19, batch 3450, loss[loss=0.1523, simple_loss=0.2273, pruned_loss=0.03864, over 4755.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2515, pruned_loss=0.05534, over 953489.01 frames. ], batch size: 27, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:17:09,624 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.792e+01 1.525e+02 1.781e+02 2.060e+02 3.433e+02, threshold=3.562e+02, percent-clipped=0.0 +2023-03-26 23:17:12,643 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2176, 2.8852, 2.7717, 1.2167, 3.0653, 2.2494, 0.7869, 1.9550], + device='cuda:4'), covar=tensor([0.2474, 0.2007, 0.1923, 0.3466, 0.1407, 0.1180, 0.4056, 0.1552], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0177, 0.0160, 0.0129, 0.0160, 0.0123, 0.0146, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 23:17:12,724 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9727, 1.8058, 1.6518, 1.3971, 1.7893, 1.8169, 1.8208, 2.2935], + device='cuda:4'), covar=tensor([0.3419, 0.3608, 0.2914, 0.3327, 0.3249, 0.2193, 0.3084, 0.1627], + device='cuda:4'), in_proj_covar=tensor([0.0287, 0.0260, 0.0230, 0.0274, 0.0251, 0.0221, 0.0251, 0.0232], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:17:24,602 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-26 23:17:40,390 INFO [finetune.py:976] (4/7) Epoch 19, batch 3500, loss[loss=0.219, simple_loss=0.2736, pruned_loss=0.08226, over 4908.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2501, pruned_loss=0.05542, over 956629.98 frames. ], batch size: 43, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:17:57,676 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106624.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:18:14,144 INFO [finetune.py:976] (4/7) Epoch 19, batch 3550, loss[loss=0.1563, simple_loss=0.2169, pruned_loss=0.04789, over 4832.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2471, pruned_loss=0.05433, over 958488.21 frames. ], batch size: 30, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:18:16,541 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 1.567e+02 1.861e+02 2.307e+02 3.604e+02, threshold=3.722e+02, percent-clipped=2.0 +2023-03-26 23:18:23,837 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.65 vs. limit=5.0 +2023-03-26 23:18:31,082 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3507, 2.2971, 2.0864, 2.4537, 2.9911, 2.4784, 2.2686, 2.0212], + device='cuda:4'), covar=tensor([0.2070, 0.1766, 0.1785, 0.1612, 0.1390, 0.1021, 0.1920, 0.1832], + device='cuda:4'), in_proj_covar=tensor([0.0243, 0.0210, 0.0212, 0.0193, 0.0242, 0.0188, 0.0215, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:18:50,821 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0591, 0.9750, 0.9675, 0.4031, 0.9499, 1.1410, 1.2312, 0.9753], + device='cuda:4'), covar=tensor([0.0888, 0.0541, 0.0520, 0.0564, 0.0569, 0.0602, 0.0374, 0.0662], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0150, 0.0124, 0.0124, 0.0130, 0.0128, 0.0140, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.0505e-05, 1.0859e-04, 8.8717e-05, 8.8009e-05, 9.1508e-05, 9.1716e-05, + 1.0065e-04, 1.0551e-04], device='cuda:4') +2023-03-26 23:18:51,976 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106692.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:18:56,156 INFO [finetune.py:976] (4/7) Epoch 19, batch 3600, loss[loss=0.1495, simple_loss=0.2204, pruned_loss=0.03927, over 4836.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2446, pruned_loss=0.05382, over 957583.21 frames. ], batch size: 33, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:19:19,203 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106733.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:19:29,858 INFO [finetune.py:976] (4/7) Epoch 19, batch 3650, loss[loss=0.1734, simple_loss=0.2319, pruned_loss=0.05744, over 4245.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2475, pruned_loss=0.05498, over 955450.62 frames. ], batch size: 18, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:19:34,984 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.178e+02 1.609e+02 2.013e+02 2.438e+02 4.457e+02, threshold=4.025e+02, percent-clipped=1.0 +2023-03-26 23:19:47,792 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7049, 1.4638, 1.8692, 1.1405, 1.8320, 1.9038, 1.4142, 1.9879], + device='cuda:4'), covar=tensor([0.1119, 0.2065, 0.1408, 0.1846, 0.0822, 0.1259, 0.2779, 0.0780], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0204, 0.0190, 0.0188, 0.0174, 0.0212, 0.0217, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:20:03,525 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106781.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:20:23,369 INFO [finetune.py:976] (4/7) Epoch 19, batch 3700, loss[loss=0.1742, simple_loss=0.2546, pruned_loss=0.04687, over 4772.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2499, pruned_loss=0.05488, over 954955.12 frames. ], batch size: 27, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:20:32,839 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106810.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:20:59,511 INFO [finetune.py:976] (4/7) Epoch 19, batch 3750, loss[loss=0.1375, simple_loss=0.2005, pruned_loss=0.03727, over 4729.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2505, pruned_loss=0.05538, over 956080.34 frames. ], batch size: 23, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:21:06,524 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.604e+02 1.833e+02 2.350e+02 4.465e+02, threshold=3.666e+02, percent-clipped=2.0 +2023-03-26 23:21:10,848 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4553, 1.2939, 1.0763, 1.2456, 1.6828, 1.5729, 1.3690, 1.1978], + device='cuda:4'), covar=tensor([0.0320, 0.0314, 0.0877, 0.0367, 0.0254, 0.0470, 0.0355, 0.0450], + device='cuda:4'), in_proj_covar=tensor([0.0097, 0.0108, 0.0145, 0.0112, 0.0100, 0.0111, 0.0100, 0.0112], + device='cuda:4'), out_proj_covar=tensor([7.5636e-05, 8.3124e-05, 1.1395e-04, 8.5939e-05, 7.8099e-05, 8.2224e-05, + 7.4869e-05, 8.5458e-05], device='cuda:4') +2023-03-26 23:21:30,170 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-26 23:21:48,088 INFO [finetune.py:976] (4/7) Epoch 19, batch 3800, loss[loss=0.1491, simple_loss=0.2174, pruned_loss=0.04037, over 4721.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2522, pruned_loss=0.05598, over 957416.87 frames. ], batch size: 23, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:22:07,682 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106924.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:22:19,085 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.58 vs. limit=2.0 +2023-03-26 23:22:24,705 INFO [finetune.py:976] (4/7) Epoch 19, batch 3850, loss[loss=0.2012, simple_loss=0.2609, pruned_loss=0.07079, over 4899.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2508, pruned_loss=0.05515, over 958873.85 frames. ], batch size: 32, lr: 3.28e-03, grad_scale: 64.0 +2023-03-26 23:22:27,157 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.119e+02 1.623e+02 1.818e+02 2.255e+02 6.115e+02, threshold=3.637e+02, percent-clipped=1.0 +2023-03-26 23:22:39,187 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=106972.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:22:52,752 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106992.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:22:57,319 INFO [finetune.py:976] (4/7) Epoch 19, batch 3900, loss[loss=0.15, simple_loss=0.2311, pruned_loss=0.03446, over 4899.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2492, pruned_loss=0.05497, over 958369.51 frames. ], batch size: 35, lr: 3.28e-03, grad_scale: 64.0 +2023-03-26 23:23:24,567 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=107040.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:23:29,941 INFO [finetune.py:976] (4/7) Epoch 19, batch 3950, loss[loss=0.1749, simple_loss=0.2472, pruned_loss=0.05132, over 4895.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2462, pruned_loss=0.05384, over 959752.34 frames. ], batch size: 32, lr: 3.28e-03, grad_scale: 64.0 +2023-03-26 23:23:35,025 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.301e+01 1.519e+02 1.802e+02 2.250e+02 5.271e+02, threshold=3.605e+02, percent-clipped=1.0 +2023-03-26 23:23:35,376 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-26 23:23:45,028 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5460, 1.5956, 1.3106, 1.5955, 1.8530, 1.7577, 1.5044, 1.3559], + device='cuda:4'), covar=tensor([0.0335, 0.0310, 0.0651, 0.0313, 0.0227, 0.0513, 0.0368, 0.0435], + device='cuda:4'), in_proj_covar=tensor([0.0096, 0.0107, 0.0144, 0.0111, 0.0100, 0.0110, 0.0100, 0.0111], + device='cuda:4'), out_proj_covar=tensor([7.4872e-05, 8.2461e-05, 1.1351e-04, 8.5435e-05, 7.7767e-05, 8.1699e-05, + 7.4284e-05, 8.4770e-05], device='cuda:4') +2023-03-26 23:24:07,127 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6355, 2.4306, 2.0149, 2.7742, 2.6111, 2.4235, 3.1174, 2.7336], + device='cuda:4'), covar=tensor([0.1261, 0.2353, 0.3254, 0.2448, 0.2479, 0.1565, 0.2816, 0.1725], + device='cuda:4'), in_proj_covar=tensor([0.0185, 0.0188, 0.0236, 0.0253, 0.0248, 0.0203, 0.0215, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:24:12,977 INFO [finetune.py:976] (4/7) Epoch 19, batch 4000, loss[loss=0.181, simple_loss=0.2623, pruned_loss=0.04985, over 4731.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.246, pruned_loss=0.0539, over 960248.57 frames. ], batch size: 59, lr: 3.28e-03, grad_scale: 64.0 +2023-03-26 23:24:21,378 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=107110.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:24:21,453 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8849, 1.2482, 1.9390, 1.9028, 1.7112, 1.6820, 1.8124, 1.8059], + device='cuda:4'), covar=tensor([0.3635, 0.3793, 0.3147, 0.3376, 0.4571, 0.3385, 0.4164, 0.2904], + device='cuda:4'), in_proj_covar=tensor([0.0250, 0.0239, 0.0259, 0.0277, 0.0274, 0.0250, 0.0284, 0.0241], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:24:41,097 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107140.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 23:24:46,974 INFO [finetune.py:976] (4/7) Epoch 19, batch 4050, loss[loss=0.1777, simple_loss=0.2514, pruned_loss=0.05195, over 4817.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2508, pruned_loss=0.05577, over 960070.10 frames. ], batch size: 51, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:24:48,812 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107152.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:24:49,866 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.579e+02 1.895e+02 2.231e+02 3.900e+02, threshold=3.790e+02, percent-clipped=1.0 +2023-03-26 23:24:52,894 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=107158.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:24:56,941 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.6072, 3.1680, 2.9132, 1.4496, 3.0372, 2.5374, 2.4829, 2.7057], + device='cuda:4'), covar=tensor([0.0663, 0.0771, 0.1511, 0.1994, 0.1376, 0.1742, 0.1727, 0.0937], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0192, 0.0199, 0.0181, 0.0209, 0.0207, 0.0222, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:25:40,276 INFO [finetune.py:976] (4/7) Epoch 19, batch 4100, loss[loss=0.184, simple_loss=0.2573, pruned_loss=0.05534, over 4830.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2529, pruned_loss=0.05628, over 957963.20 frames. ], batch size: 49, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:25:41,782 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107201.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 23:25:50,476 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107213.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 23:26:13,434 INFO [finetune.py:976] (4/7) Epoch 19, batch 4150, loss[loss=0.2296, simple_loss=0.2917, pruned_loss=0.08372, over 4756.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2543, pruned_loss=0.05654, over 957967.56 frames. ], batch size: 26, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:26:21,811 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.077e+02 1.570e+02 1.970e+02 2.461e+02 5.293e+02, threshold=3.939e+02, percent-clipped=1.0 +2023-03-26 23:26:23,778 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6392, 1.4791, 1.9090, 1.2051, 1.7762, 1.8910, 1.4251, 2.0326], + device='cuda:4'), covar=tensor([0.1208, 0.2231, 0.1224, 0.1801, 0.0881, 0.1324, 0.2970, 0.0812], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0205, 0.0191, 0.0189, 0.0174, 0.0213, 0.0218, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:26:39,593 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8704, 1.6801, 1.4688, 1.2593, 1.6233, 1.5999, 1.6345, 2.1999], + device='cuda:4'), covar=tensor([0.3858, 0.3750, 0.3131, 0.3690, 0.3782, 0.2271, 0.3380, 0.1735], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0263, 0.0231, 0.0277, 0.0253, 0.0223, 0.0252, 0.0234], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:26:56,740 INFO [finetune.py:976] (4/7) Epoch 19, batch 4200, loss[loss=0.1677, simple_loss=0.2417, pruned_loss=0.04686, over 4802.00 frames. ], tot_loss[loss=0.1828, simple_loss=0.2537, pruned_loss=0.05592, over 956098.43 frames. ], batch size: 40, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:27:29,944 INFO [finetune.py:976] (4/7) Epoch 19, batch 4250, loss[loss=0.1349, simple_loss=0.205, pruned_loss=0.03244, over 4696.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2502, pruned_loss=0.05479, over 957052.58 frames. ], batch size: 23, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:27:33,464 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.503e+02 1.795e+02 2.146e+02 3.676e+02, threshold=3.590e+02, percent-clipped=0.0 +2023-03-26 23:27:43,138 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-26 23:28:03,396 INFO [finetune.py:976] (4/7) Epoch 19, batch 4300, loss[loss=0.2126, simple_loss=0.2806, pruned_loss=0.07232, over 4834.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2472, pruned_loss=0.05423, over 959186.35 frames. ], batch size: 39, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:28:30,821 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3423, 2.0539, 2.1917, 1.0488, 2.4728, 2.7688, 2.2953, 1.9647], + device='cuda:4'), covar=tensor([0.0764, 0.0796, 0.0535, 0.0632, 0.0643, 0.0573, 0.0405, 0.0692], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0150, 0.0125, 0.0124, 0.0131, 0.0128, 0.0141, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.0402e-05, 1.0846e-04, 8.9140e-05, 8.8193e-05, 9.2039e-05, 9.1433e-05, + 1.0119e-04, 1.0553e-04], device='cuda:4') +2023-03-26 23:28:32,473 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.34 vs. limit=5.0 +2023-03-26 23:28:36,191 INFO [finetune.py:976] (4/7) Epoch 19, batch 4350, loss[loss=0.1594, simple_loss=0.2332, pruned_loss=0.04279, over 4763.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2452, pruned_loss=0.05366, over 958023.77 frames. ], batch size: 28, lr: 3.28e-03, grad_scale: 32.0 +2023-03-26 23:28:40,179 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.532e+02 1.813e+02 2.231e+02 3.395e+02, threshold=3.625e+02, percent-clipped=1.0 +2023-03-26 23:29:21,287 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=107496.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 23:29:23,018 INFO [finetune.py:976] (4/7) Epoch 19, batch 4400, loss[loss=0.176, simple_loss=0.2437, pruned_loss=0.05416, over 4754.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2468, pruned_loss=0.05443, over 958390.37 frames. ], batch size: 27, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:29:29,602 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=107508.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 23:29:31,353 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107510.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:29:56,828 INFO [finetune.py:976] (4/7) Epoch 19, batch 4450, loss[loss=0.2016, simple_loss=0.2831, pruned_loss=0.06, over 4811.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2507, pruned_loss=0.05519, over 955774.54 frames. ], batch size: 38, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:29:59,905 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.983e+01 1.601e+02 1.972e+02 2.467e+02 3.942e+02, threshold=3.944e+02, percent-clipped=4.0 +2023-03-26 23:30:01,738 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.68 vs. limit=2.0 +2023-03-26 23:30:10,250 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.68 vs. limit=2.0 +2023-03-26 23:30:12,263 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107571.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:30:42,901 INFO [finetune.py:976] (4/7) Epoch 19, batch 4500, loss[loss=0.185, simple_loss=0.2605, pruned_loss=0.05476, over 4859.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2517, pruned_loss=0.0553, over 955026.66 frames. ], batch size: 44, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:31:25,201 INFO [finetune.py:976] (4/7) Epoch 19, batch 4550, loss[loss=0.1865, simple_loss=0.2611, pruned_loss=0.05592, over 4884.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2537, pruned_loss=0.0563, over 955763.11 frames. ], batch size: 43, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:31:28,195 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.552e+02 1.832e+02 2.186e+02 5.352e+02, threshold=3.664e+02, percent-clipped=1.0 +2023-03-26 23:31:28,332 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8689, 1.7818, 1.8601, 1.1529, 1.8982, 1.8499, 1.8451, 1.5026], + device='cuda:4'), covar=tensor([0.0577, 0.0654, 0.0640, 0.0909, 0.0743, 0.0704, 0.0600, 0.1197], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0134, 0.0140, 0.0120, 0.0125, 0.0139, 0.0140, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:31:31,346 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107659.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:32:03,427 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0926, 1.7866, 2.4639, 3.9800, 2.7275, 2.7643, 0.9079, 3.3571], + device='cuda:4'), covar=tensor([0.1651, 0.1486, 0.1427, 0.0568, 0.0760, 0.1710, 0.1977, 0.0426], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0115, 0.0132, 0.0163, 0.0100, 0.0135, 0.0123, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 23:32:07,547 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5818, 1.5162, 1.4634, 1.6060, 1.0744, 3.4943, 1.3899, 1.8755], + device='cuda:4'), covar=tensor([0.3473, 0.2678, 0.2214, 0.2360, 0.1863, 0.0177, 0.2601, 0.1253], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0114, 0.0119, 0.0122, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 23:32:12,110 INFO [finetune.py:976] (4/7) Epoch 19, batch 4600, loss[loss=0.1558, simple_loss=0.2345, pruned_loss=0.03852, over 4839.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2532, pruned_loss=0.05587, over 955993.76 frames. ], batch size: 49, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:32:12,451 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-03-26 23:32:26,323 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107720.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:32:37,559 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107737.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:32:45,691 INFO [finetune.py:976] (4/7) Epoch 19, batch 4650, loss[loss=0.1716, simple_loss=0.2281, pruned_loss=0.05758, over 4829.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2497, pruned_loss=0.05489, over 955910.51 frames. ], batch size: 25, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:32:48,739 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.855e+01 1.504e+02 1.713e+02 2.086e+02 4.043e+02, threshold=3.426e+02, percent-clipped=2.0 +2023-03-26 23:32:55,919 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-03-26 23:33:17,158 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=107796.0, num_to_drop=1, layers_to_drop={2} +2023-03-26 23:33:18,396 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107798.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 23:33:19,352 INFO [finetune.py:976] (4/7) Epoch 19, batch 4700, loss[loss=0.1891, simple_loss=0.2516, pruned_loss=0.06332, over 4754.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2453, pruned_loss=0.05303, over 956666.01 frames. ], batch size: 27, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:33:25,040 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=107808.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:33:49,096 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=107844.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 23:33:54,086 INFO [finetune.py:976] (4/7) Epoch 19, batch 4750, loss[loss=0.1595, simple_loss=0.2405, pruned_loss=0.03926, over 4900.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2435, pruned_loss=0.0526, over 955970.74 frames. ], batch size: 35, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:33:57,615 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.048e+02 1.438e+02 1.688e+02 2.143e+02 3.806e+02, threshold=3.376e+02, percent-clipped=2.0 +2023-03-26 23:33:58,881 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=107856.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:34:04,995 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=107866.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:34:37,227 INFO [finetune.py:976] (4/7) Epoch 19, batch 4800, loss[loss=0.1735, simple_loss=0.2321, pruned_loss=0.05739, over 4212.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2451, pruned_loss=0.0531, over 955310.07 frames. ], batch size: 18, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:34:54,343 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-26 23:35:10,749 INFO [finetune.py:976] (4/7) Epoch 19, batch 4850, loss[loss=0.2026, simple_loss=0.2491, pruned_loss=0.07802, over 4272.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2503, pruned_loss=0.05574, over 955042.05 frames. ], batch size: 18, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:35:13,742 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.163e+02 1.610e+02 1.895e+02 2.225e+02 4.035e+02, threshold=3.790e+02, percent-clipped=2.0 +2023-03-26 23:35:19,085 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6078, 1.5793, 1.5153, 1.6215, 1.0592, 3.4974, 1.3832, 1.8518], + device='cuda:4'), covar=tensor([0.3255, 0.2455, 0.2124, 0.2292, 0.1810, 0.0208, 0.2767, 0.1239], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0115, 0.0120, 0.0123, 0.0114, 0.0097, 0.0095, 0.0096], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 23:35:29,623 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.69 vs. limit=5.0 +2023-03-26 23:35:45,818 INFO [finetune.py:976] (4/7) Epoch 19, batch 4900, loss[loss=0.2098, simple_loss=0.2833, pruned_loss=0.06816, over 4777.00 frames. ], tot_loss[loss=0.1825, simple_loss=0.2527, pruned_loss=0.05618, over 957011.74 frames. ], batch size: 51, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:35:56,295 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-26 23:35:57,380 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108015.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:35:58,181 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-26 23:36:28,166 INFO [finetune.py:976] (4/7) Epoch 19, batch 4950, loss[loss=0.1921, simple_loss=0.2662, pruned_loss=0.059, over 4819.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.254, pruned_loss=0.05654, over 955647.28 frames. ], batch size: 40, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:36:31,629 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.136e+02 1.572e+02 1.807e+02 2.323e+02 4.539e+02, threshold=3.614e+02, percent-clipped=1.0 +2023-03-26 23:37:04,012 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108093.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 23:37:10,960 INFO [finetune.py:976] (4/7) Epoch 19, batch 5000, loss[loss=0.166, simple_loss=0.2388, pruned_loss=0.04664, over 4892.00 frames. ], tot_loss[loss=0.1827, simple_loss=0.2526, pruned_loss=0.05637, over 955207.91 frames. ], batch size: 35, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:37:26,855 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5775, 0.7222, 1.6927, 1.5410, 1.4682, 1.3561, 1.5097, 1.6043], + device='cuda:4'), covar=tensor([0.3464, 0.3461, 0.2704, 0.2919, 0.3800, 0.2996, 0.3492, 0.2557], + device='cuda:4'), in_proj_covar=tensor([0.0252, 0.0241, 0.0260, 0.0280, 0.0277, 0.0252, 0.0286, 0.0243], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:37:54,037 INFO [finetune.py:976] (4/7) Epoch 19, batch 5050, loss[loss=0.1519, simple_loss=0.2116, pruned_loss=0.04609, over 3985.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2496, pruned_loss=0.05547, over 954616.24 frames. ], batch size: 17, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:37:57,573 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.150e+02 1.577e+02 1.863e+02 2.132e+02 3.762e+02, threshold=3.725e+02, percent-clipped=1.0 +2023-03-26 23:38:03,816 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-26 23:38:05,839 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108166.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:38:27,769 INFO [finetune.py:976] (4/7) Epoch 19, batch 5100, loss[loss=0.1646, simple_loss=0.2391, pruned_loss=0.04505, over 4830.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2454, pruned_loss=0.05358, over 955773.44 frames. ], batch size: 39, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:38:38,020 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=108214.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:39:00,758 INFO [finetune.py:976] (4/7) Epoch 19, batch 5150, loss[loss=0.2087, simple_loss=0.2794, pruned_loss=0.06903, over 4818.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2435, pruned_loss=0.05292, over 953518.85 frames. ], batch size: 40, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:39:04,797 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.278e+01 1.451e+02 1.873e+02 2.231e+02 4.201e+02, threshold=3.747e+02, percent-clipped=0.0 +2023-03-26 23:39:39,543 INFO [finetune.py:976] (4/7) Epoch 19, batch 5200, loss[loss=0.2267, simple_loss=0.298, pruned_loss=0.07774, over 4811.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.248, pruned_loss=0.05418, over 951754.08 frames. ], batch size: 51, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:39:50,297 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3920, 3.3331, 3.1706, 1.4278, 3.4822, 2.6273, 0.7679, 2.2520], + device='cuda:4'), covar=tensor([0.2373, 0.1978, 0.1590, 0.3388, 0.1174, 0.1043, 0.4022, 0.1413], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0174, 0.0158, 0.0128, 0.0158, 0.0121, 0.0145, 0.0121], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-26 23:39:54,426 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108315.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:40:00,158 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-26 23:40:00,995 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108325.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:40:16,478 INFO [finetune.py:976] (4/7) Epoch 19, batch 5250, loss[loss=0.1626, simple_loss=0.2458, pruned_loss=0.03968, over 4913.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2519, pruned_loss=0.0557, over 954462.89 frames. ], batch size: 36, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:40:19,994 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.128e+02 1.566e+02 1.984e+02 2.332e+02 4.295e+02, threshold=3.968e+02, percent-clipped=2.0 +2023-03-26 23:40:26,550 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=108363.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:40:42,060 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108386.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:40:46,338 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108393.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:40:49,958 INFO [finetune.py:976] (4/7) Epoch 19, batch 5300, loss[loss=0.1555, simple_loss=0.2396, pruned_loss=0.0357, over 4757.00 frames. ], tot_loss[loss=0.1835, simple_loss=0.254, pruned_loss=0.05647, over 955631.02 frames. ], batch size: 27, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:40:53,141 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.2082, 2.7949, 2.6758, 1.2161, 2.7782, 2.3310, 2.1746, 2.5835], + device='cuda:4'), covar=tensor([0.1050, 0.0959, 0.2076, 0.2507, 0.1684, 0.2317, 0.2283, 0.1283], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0194, 0.0200, 0.0183, 0.0211, 0.0208, 0.0223, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:41:22,544 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108432.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:41:22,577 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0992, 1.8958, 1.6280, 1.8518, 1.7803, 1.7774, 1.7935, 2.5817], + device='cuda:4'), covar=tensor([0.3848, 0.4220, 0.3311, 0.3724, 0.4130, 0.2444, 0.3970, 0.1742], + device='cuda:4'), in_proj_covar=tensor([0.0288, 0.0260, 0.0230, 0.0275, 0.0251, 0.0221, 0.0253, 0.0233], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:41:32,019 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=108441.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:41:36,861 INFO [finetune.py:976] (4/7) Epoch 19, batch 5350, loss[loss=0.1845, simple_loss=0.2482, pruned_loss=0.06041, over 4863.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2521, pruned_loss=0.05543, over 953679.14 frames. ], batch size: 31, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:41:39,877 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.546e+01 1.453e+02 1.815e+02 2.266e+02 3.194e+02, threshold=3.630e+02, percent-clipped=0.0 +2023-03-26 23:42:09,118 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108493.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 23:42:12,600 INFO [finetune.py:976] (4/7) Epoch 19, batch 5400, loss[loss=0.1803, simple_loss=0.2437, pruned_loss=0.05845, over 4829.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2496, pruned_loss=0.05453, over 953685.71 frames. ], batch size: 25, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:42:53,851 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5161, 2.4958, 2.0891, 2.5718, 2.4470, 2.4267, 2.3450, 3.3468], + device='cuda:4'), covar=tensor([0.4171, 0.4686, 0.3706, 0.4253, 0.4218, 0.2738, 0.4543, 0.1732], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0261, 0.0231, 0.0276, 0.0252, 0.0222, 0.0254, 0.0233], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:42:58,628 INFO [finetune.py:976] (4/7) Epoch 19, batch 5450, loss[loss=0.1831, simple_loss=0.2422, pruned_loss=0.06205, over 4939.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2464, pruned_loss=0.05348, over 954737.89 frames. ], batch size: 33, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:43:01,646 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.000e+02 1.562e+02 1.816e+02 2.216e+02 4.232e+02, threshold=3.632e+02, percent-clipped=2.0 +2023-03-26 23:43:17,255 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108577.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:43:23,070 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5590, 1.5442, 2.1643, 3.1661, 2.0777, 2.3897, 1.0818, 2.6520], + device='cuda:4'), covar=tensor([0.1770, 0.1375, 0.1212, 0.0592, 0.0853, 0.1367, 0.1787, 0.0488], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0116, 0.0134, 0.0164, 0.0100, 0.0136, 0.0124, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 23:43:31,854 INFO [finetune.py:976] (4/7) Epoch 19, batch 5500, loss[loss=0.1672, simple_loss=0.2488, pruned_loss=0.0428, over 4760.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2442, pruned_loss=0.0532, over 954904.86 frames. ], batch size: 54, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:43:48,427 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6673, 1.6164, 1.5775, 1.6045, 1.2213, 4.0998, 1.5929, 1.9395], + device='cuda:4'), covar=tensor([0.3270, 0.2441, 0.2156, 0.2305, 0.1726, 0.0139, 0.2595, 0.1247], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0115, 0.0120, 0.0123, 0.0114, 0.0096, 0.0095, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 23:43:50,782 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4540, 2.1880, 2.3553, 1.6684, 2.4016, 2.4432, 2.5006, 1.9345], + device='cuda:4'), covar=tensor([0.0479, 0.0682, 0.0642, 0.0892, 0.0850, 0.0603, 0.0496, 0.1153], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0135, 0.0140, 0.0121, 0.0125, 0.0139, 0.0140, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:43:58,832 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108638.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:44:00,048 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.73 vs. limit=5.0 +2023-03-26 23:44:05,685 INFO [finetune.py:976] (4/7) Epoch 19, batch 5550, loss[loss=0.1785, simple_loss=0.2551, pruned_loss=0.05099, over 4841.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2453, pruned_loss=0.05306, over 955953.10 frames. ], batch size: 49, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:44:08,703 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.553e+02 1.822e+02 2.201e+02 3.552e+02, threshold=3.643e+02, percent-clipped=0.0 +2023-03-26 23:44:27,078 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108681.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:44:29,401 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8455, 1.8570, 2.2856, 2.1399, 2.0032, 4.5035, 1.7918, 2.0135], + device='cuda:4'), covar=tensor([0.0885, 0.1776, 0.1129, 0.0900, 0.1588, 0.0184, 0.1438, 0.1720], + device='cuda:4'), in_proj_covar=tensor([0.0074, 0.0081, 0.0074, 0.0076, 0.0091, 0.0080, 0.0084, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 23:44:37,477 INFO [finetune.py:976] (4/7) Epoch 19, batch 5600, loss[loss=0.1925, simple_loss=0.2617, pruned_loss=0.06162, over 4916.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2486, pruned_loss=0.05421, over 953468.13 frames. ], batch size: 42, lr: 3.27e-03, grad_scale: 32.0 +2023-03-26 23:44:46,276 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2446, 2.0631, 1.5267, 0.7006, 1.7594, 1.9833, 1.8034, 1.8655], + device='cuda:4'), covar=tensor([0.0748, 0.0713, 0.1198, 0.1713, 0.1095, 0.1814, 0.1901, 0.0720], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0194, 0.0201, 0.0183, 0.0211, 0.0209, 0.0224, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:44:47,196 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.74 vs. limit=5.0 +2023-03-26 23:44:53,987 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.4446, 1.6491, 1.6314, 0.9261, 1.7439, 1.9991, 1.9104, 1.4217], + device='cuda:4'), covar=tensor([0.1165, 0.0790, 0.0724, 0.0662, 0.0628, 0.0947, 0.0443, 0.0980], + device='cuda:4'), in_proj_covar=tensor([0.0125, 0.0151, 0.0125, 0.0126, 0.0132, 0.0129, 0.0144, 0.0148], + device='cuda:4'), out_proj_covar=tensor([9.1431e-05, 1.0905e-04, 8.9474e-05, 8.9022e-05, 9.2912e-05, 9.2473e-05, + 1.0310e-04, 1.0646e-04], device='cuda:4') +2023-03-26 23:45:09,466 INFO [finetune.py:976] (4/7) Epoch 19, batch 5650, loss[loss=0.2247, simple_loss=0.2889, pruned_loss=0.08029, over 4798.00 frames. ], tot_loss[loss=0.181, simple_loss=0.252, pruned_loss=0.05503, over 951518.98 frames. ], batch size: 45, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:45:12,317 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.584e+02 1.878e+02 2.184e+02 3.636e+02, threshold=3.756e+02, percent-clipped=0.0 +2023-03-26 23:45:32,950 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108788.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 23:45:34,786 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108791.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:45:37,232 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-26 23:45:39,455 INFO [finetune.py:976] (4/7) Epoch 19, batch 5700, loss[loss=0.1709, simple_loss=0.2339, pruned_loss=0.05391, over 4057.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2477, pruned_loss=0.05417, over 931342.62 frames. ], batch size: 17, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:45:42,665 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7018, 1.6769, 1.9224, 1.3722, 1.6742, 1.9419, 1.6282, 2.0414], + device='cuda:4'), covar=tensor([0.1123, 0.1823, 0.1314, 0.1468, 0.0895, 0.1305, 0.2451, 0.0744], + device='cuda:4'), in_proj_covar=tensor([0.0191, 0.0204, 0.0191, 0.0188, 0.0174, 0.0212, 0.0217, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:46:08,102 INFO [finetune.py:976] (4/7) Epoch 20, batch 0, loss[loss=0.1984, simple_loss=0.2573, pruned_loss=0.06978, over 4889.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2573, pruned_loss=0.06978, over 4889.00 frames. ], batch size: 35, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:46:08,103 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-26 23:46:16,069 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8907, 1.3154, 0.9423, 1.6312, 2.1634, 1.2103, 1.6184, 1.5796], + device='cuda:4'), covar=tensor([0.1404, 0.1940, 0.1795, 0.1110, 0.1854, 0.1956, 0.1287, 0.1900], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0095, 0.0111, 0.0091, 0.0120, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 23:46:17,449 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1664, 1.9498, 1.7774, 1.7940, 1.8939, 1.9041, 1.9136, 2.5525], + device='cuda:4'), covar=tensor([0.3984, 0.4796, 0.3621, 0.4033, 0.4126, 0.2629, 0.4037, 0.1952], + device='cuda:4'), in_proj_covar=tensor([0.0287, 0.0261, 0.0230, 0.0275, 0.0251, 0.0221, 0.0252, 0.0232], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:46:24,541 INFO [finetune.py:1010] (4/7) Epoch 20, validation: loss=0.158, simple_loss=0.2276, pruned_loss=0.04423, over 2265189.00 frames. +2023-03-26 23:46:24,542 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-26 23:46:28,022 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0884, 0.9768, 0.9150, 1.1181, 1.2431, 1.1704, 0.9534, 0.8959], + device='cuda:4'), covar=tensor([0.0422, 0.0333, 0.0695, 0.0332, 0.0294, 0.0434, 0.0388, 0.0441], + device='cuda:4'), in_proj_covar=tensor([0.0096, 0.0107, 0.0143, 0.0111, 0.0099, 0.0110, 0.0099, 0.0111], + device='cuda:4'), out_proj_covar=tensor([7.4468e-05, 8.2235e-05, 1.1294e-04, 8.5013e-05, 7.7084e-05, 8.1163e-05, + 7.3666e-05, 8.4546e-05], device='cuda:4') +2023-03-26 23:46:32,152 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2873, 1.3844, 0.6658, 2.0224, 2.4362, 1.7986, 1.7494, 1.7935], + device='cuda:4'), covar=tensor([0.1379, 0.2105, 0.2231, 0.1090, 0.1848, 0.1791, 0.1456, 0.1977], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0095, 0.0111, 0.0091, 0.0120, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 23:46:52,593 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108852.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:46:58,208 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.806e+01 1.424e+02 1.737e+02 2.098e+02 5.389e+02, threshold=3.475e+02, percent-clipped=2.0 +2023-03-26 23:47:17,655 INFO [finetune.py:976] (4/7) Epoch 20, batch 50, loss[loss=0.183, simple_loss=0.2492, pruned_loss=0.05838, over 4894.00 frames. ], tot_loss[loss=0.1824, simple_loss=0.2517, pruned_loss=0.05654, over 217567.06 frames. ], batch size: 36, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:47:19,465 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7343, 1.6416, 1.5419, 1.6740, 1.2748, 3.5196, 1.4998, 1.9697], + device='cuda:4'), covar=tensor([0.3104, 0.2395, 0.2007, 0.2142, 0.1533, 0.0191, 0.2535, 0.1106], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0114, 0.0119, 0.0122, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 23:47:57,360 INFO [finetune.py:976] (4/7) Epoch 20, batch 100, loss[loss=0.1645, simple_loss=0.2358, pruned_loss=0.04654, over 4864.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2476, pruned_loss=0.05458, over 383348.04 frames. ], batch size: 31, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:48:04,459 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-26 23:48:06,346 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108933.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:48:13,414 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.70 vs. limit=5.0 +2023-03-26 23:48:23,147 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.362e+02 1.754e+02 2.070e+02 5.157e+02, threshold=3.508e+02, percent-clipped=1.0 +2023-03-26 23:48:38,579 INFO [finetune.py:976] (4/7) Epoch 20, batch 150, loss[loss=0.1757, simple_loss=0.2317, pruned_loss=0.05991, over 4740.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2416, pruned_loss=0.05286, over 509662.00 frames. ], batch size: 54, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:48:41,558 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108981.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:49:04,006 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5293, 1.6275, 1.3173, 1.6136, 1.8360, 1.7633, 1.4944, 1.4081], + device='cuda:4'), covar=tensor([0.0290, 0.0288, 0.0610, 0.0292, 0.0220, 0.0504, 0.0339, 0.0379], + device='cuda:4'), in_proj_covar=tensor([0.0096, 0.0106, 0.0143, 0.0111, 0.0099, 0.0110, 0.0099, 0.0110], + device='cuda:4'), out_proj_covar=tensor([7.4520e-05, 8.1978e-05, 1.1242e-04, 8.4888e-05, 7.6935e-05, 8.1188e-05, + 7.3524e-05, 8.4272e-05], device='cuda:4') +2023-03-26 23:49:11,416 INFO [finetune.py:976] (4/7) Epoch 20, batch 200, loss[loss=0.1781, simple_loss=0.2536, pruned_loss=0.05132, over 4860.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2417, pruned_loss=0.05313, over 608209.63 frames. ], batch size: 44, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:49:11,511 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5168, 1.4657, 1.9822, 2.9227, 1.9353, 2.2431, 1.1634, 2.4621], + device='cuda:4'), covar=tensor([0.1780, 0.1449, 0.1212, 0.0727, 0.0928, 0.1353, 0.1680, 0.0555], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0116, 0.0133, 0.0164, 0.0100, 0.0136, 0.0124, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 23:49:13,182 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109029.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:49:19,015 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109037.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 23:49:29,132 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.519e+02 1.780e+02 2.129e+02 3.450e+02, threshold=3.561e+02, percent-clipped=0.0 +2023-03-26 23:49:32,729 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0512, 1.4865, 1.0459, 2.0794, 2.2529, 2.0717, 1.7371, 1.8244], + device='cuda:4'), covar=tensor([0.1212, 0.1767, 0.1847, 0.0906, 0.1869, 0.1586, 0.1166, 0.1604], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0095, 0.0111, 0.0091, 0.0120, 0.0093, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-26 23:49:42,646 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 23:49:44,471 INFO [finetune.py:976] (4/7) Epoch 20, batch 250, loss[loss=0.2107, simple_loss=0.2811, pruned_loss=0.07014, over 4817.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2449, pruned_loss=0.05379, over 685541.66 frames. ], batch size: 51, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:49:52,136 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109088.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:49:58,702 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109098.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 23:50:17,245 INFO [finetune.py:976] (4/7) Epoch 20, batch 300, loss[loss=0.1773, simple_loss=0.2524, pruned_loss=0.05104, over 4864.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2466, pruned_loss=0.05489, over 741997.94 frames. ], batch size: 34, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:50:23,615 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109136.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:50:31,197 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109147.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:50:35,400 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.121e+02 1.518e+02 1.856e+02 2.256e+02 3.204e+02, threshold=3.712e+02, percent-clipped=0.0 +2023-03-26 23:50:50,168 INFO [finetune.py:976] (4/7) Epoch 20, batch 350, loss[loss=0.2033, simple_loss=0.2725, pruned_loss=0.06705, over 4744.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2489, pruned_loss=0.05523, over 790870.92 frames. ], batch size: 59, lr: 3.26e-03, grad_scale: 64.0 +2023-03-26 23:51:00,294 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-26 23:51:01,443 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109194.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:51:01,860 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.44 vs. limit=5.0 +2023-03-26 23:51:08,326 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.3829, 3.8188, 4.0078, 4.2657, 4.1553, 3.8356, 4.4760, 1.3828], + device='cuda:4'), covar=tensor([0.0928, 0.0966, 0.0849, 0.1027, 0.1226, 0.1560, 0.0677, 0.5625], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0244, 0.0276, 0.0291, 0.0331, 0.0281, 0.0300, 0.0294], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:51:10,170 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8701, 1.3632, 0.8054, 1.7025, 2.0893, 1.4038, 1.7403, 1.6076], + device='cuda:4'), covar=tensor([0.1483, 0.2062, 0.2067, 0.1194, 0.1987, 0.1884, 0.1372, 0.2108], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0095, 0.0110, 0.0091, 0.0119, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-26 23:51:25,274 INFO [finetune.py:976] (4/7) Epoch 20, batch 400, loss[loss=0.156, simple_loss=0.2452, pruned_loss=0.03334, over 4900.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2492, pruned_loss=0.05433, over 828804.20 frames. ], batch size: 36, lr: 3.26e-03, grad_scale: 64.0 +2023-03-26 23:51:34,320 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109233.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:52:03,601 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.047e+02 1.657e+02 1.900e+02 2.185e+02 4.941e+02, threshold=3.801e+02, percent-clipped=3.0 +2023-03-26 23:52:03,751 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8917, 1.8459, 1.5324, 2.0135, 2.4483, 2.0288, 1.6966, 1.4923], + device='cuda:4'), covar=tensor([0.2103, 0.1999, 0.1944, 0.1661, 0.1584, 0.1169, 0.2297, 0.1931], + device='cuda:4'), in_proj_covar=tensor([0.0245, 0.0211, 0.0211, 0.0194, 0.0244, 0.0188, 0.0216, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:52:04,370 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109255.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:52:26,281 INFO [finetune.py:976] (4/7) Epoch 20, batch 450, loss[loss=0.2123, simple_loss=0.2737, pruned_loss=0.07548, over 4848.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2491, pruned_loss=0.05431, over 858759.16 frames. ], batch size: 44, lr: 3.26e-03, grad_scale: 64.0 +2023-03-26 23:52:29,296 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109281.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:52:38,177 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109294.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:52:54,643 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109318.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:53:00,674 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.46 vs. limit=2.0 +2023-03-26 23:53:02,104 INFO [finetune.py:976] (4/7) Epoch 20, batch 500, loss[loss=0.1813, simple_loss=0.2558, pruned_loss=0.05337, over 4872.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2463, pruned_loss=0.05328, over 878722.65 frames. ], batch size: 31, lr: 3.26e-03, grad_scale: 64.0 +2023-03-26 23:53:34,823 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.492e+02 1.802e+02 2.178e+02 4.247e+02, threshold=3.605e+02, percent-clipped=3.0 +2023-03-26 23:53:39,362 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109355.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:53:47,671 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.7496, 1.6620, 1.5523, 0.9958, 1.7421, 1.8691, 1.8146, 1.4838], + device='cuda:4'), covar=tensor([0.0950, 0.0651, 0.0501, 0.0570, 0.0437, 0.0717, 0.0341, 0.0815], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0149, 0.0124, 0.0124, 0.0130, 0.0128, 0.0142, 0.0147], + device='cuda:4'), out_proj_covar=tensor([9.0243e-05, 1.0815e-04, 8.8500e-05, 8.7808e-05, 9.1689e-05, 9.1520e-05, + 1.0162e-04, 1.0523e-04], device='cuda:4') +2023-03-26 23:53:52,985 INFO [finetune.py:976] (4/7) Epoch 20, batch 550, loss[loss=0.1899, simple_loss=0.2478, pruned_loss=0.06603, over 4873.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2448, pruned_loss=0.05353, over 894681.35 frames. ], batch size: 34, lr: 3.26e-03, grad_scale: 64.0 +2023-03-26 23:53:53,721 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5296, 1.5023, 1.2718, 1.5044, 1.7661, 1.7110, 1.4862, 1.3494], + device='cuda:4'), covar=tensor([0.0307, 0.0296, 0.0627, 0.0314, 0.0225, 0.0463, 0.0374, 0.0361], + device='cuda:4'), in_proj_covar=tensor([0.0096, 0.0106, 0.0143, 0.0110, 0.0099, 0.0110, 0.0099, 0.0110], + device='cuda:4'), out_proj_covar=tensor([7.4665e-05, 8.1787e-05, 1.1256e-04, 8.4643e-05, 7.6756e-05, 8.1074e-05, + 7.3525e-05, 8.4114e-05], device='cuda:4') +2023-03-26 23:53:54,325 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109379.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 23:54:03,682 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109393.0, num_to_drop=1, layers_to_drop={3} +2023-03-26 23:54:26,236 INFO [finetune.py:976] (4/7) Epoch 20, batch 600, loss[loss=0.1965, simple_loss=0.2648, pruned_loss=0.06407, over 4866.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2455, pruned_loss=0.05383, over 908999.58 frames. ], batch size: 44, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:54:30,584 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.79 vs. limit=5.0 +2023-03-26 23:54:31,678 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-26 23:54:39,806 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109447.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:54:42,869 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.5979, 1.6359, 1.6027, 0.9207, 1.7944, 1.9416, 1.9200, 1.4503], + device='cuda:4'), covar=tensor([0.0926, 0.0706, 0.0593, 0.0612, 0.0503, 0.0691, 0.0331, 0.0806], + device='cuda:4'), in_proj_covar=tensor([0.0124, 0.0151, 0.0125, 0.0125, 0.0131, 0.0129, 0.0143, 0.0148], + device='cuda:4'), out_proj_covar=tensor([9.1048e-05, 1.0913e-04, 8.9055e-05, 8.8457e-05, 9.2430e-05, 9.2308e-05, + 1.0237e-04, 1.0609e-04], device='cuda:4') +2023-03-26 23:54:44,549 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.417e+01 1.548e+02 1.729e+02 2.159e+02 3.434e+02, threshold=3.458e+02, percent-clipped=0.0 +2023-03-26 23:54:52,921 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4729, 1.3636, 1.3674, 1.3566, 0.7612, 2.2779, 0.7074, 1.1742], + device='cuda:4'), covar=tensor([0.3353, 0.2564, 0.2244, 0.2422, 0.2008, 0.0349, 0.2816, 0.1414], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0115, 0.0120, 0.0123, 0.0114, 0.0096, 0.0095, 0.0096], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-26 23:54:59,322 INFO [finetune.py:976] (4/7) Epoch 20, batch 650, loss[loss=0.1728, simple_loss=0.2489, pruned_loss=0.04835, over 4831.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2478, pruned_loss=0.05429, over 919318.58 frames. ], batch size: 49, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:55:01,896 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8767, 1.7150, 2.1151, 1.2719, 1.9866, 2.0791, 1.6455, 2.2591], + device='cuda:4'), covar=tensor([0.1324, 0.2144, 0.1507, 0.2092, 0.1028, 0.1708, 0.2651, 0.0825], + device='cuda:4'), in_proj_covar=tensor([0.0190, 0.0202, 0.0189, 0.0186, 0.0171, 0.0210, 0.0214, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:55:10,857 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109495.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:55:16,095 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8282, 1.0210, 1.8409, 1.7939, 1.6579, 1.6253, 1.6762, 1.7214], + device='cuda:4'), covar=tensor([0.3796, 0.3919, 0.3194, 0.3689, 0.4577, 0.3579, 0.4255, 0.2990], + device='cuda:4'), in_proj_covar=tensor([0.0252, 0.0240, 0.0261, 0.0280, 0.0278, 0.0252, 0.0288, 0.0243], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:55:33,023 INFO [finetune.py:976] (4/7) Epoch 20, batch 700, loss[loss=0.184, simple_loss=0.2579, pruned_loss=0.05504, over 4863.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2495, pruned_loss=0.05495, over 925698.11 frames. ], batch size: 31, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:55:33,768 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4621, 1.3507, 1.5016, 0.8528, 1.5151, 1.5383, 1.4726, 1.3004], + device='cuda:4'), covar=tensor([0.0679, 0.0931, 0.0756, 0.1017, 0.0951, 0.0798, 0.0693, 0.1299], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0135, 0.0140, 0.0120, 0.0125, 0.0140, 0.0139, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:55:47,502 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109550.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:55:51,355 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.095e+02 1.485e+02 1.783e+02 2.085e+02 4.380e+02, threshold=3.566e+02, percent-clipped=3.0 +2023-03-26 23:56:06,063 INFO [finetune.py:976] (4/7) Epoch 20, batch 750, loss[loss=0.1675, simple_loss=0.2429, pruned_loss=0.04609, over 4891.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2502, pruned_loss=0.05503, over 931413.07 frames. ], batch size: 37, lr: 3.26e-03, grad_scale: 32.0 +2023-03-26 23:56:07,739 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109579.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:56:21,704 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1582, 3.5885, 3.8338, 4.0538, 3.9270, 3.7507, 4.2385, 1.4317], + device='cuda:4'), covar=tensor([0.0877, 0.0952, 0.0921, 0.0996, 0.1278, 0.1473, 0.0741, 0.5473], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0244, 0.0278, 0.0291, 0.0331, 0.0281, 0.0300, 0.0295], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:56:39,568 INFO [finetune.py:976] (4/7) Epoch 20, batch 800, loss[loss=0.2041, simple_loss=0.2591, pruned_loss=0.07453, over 4797.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2493, pruned_loss=0.05414, over 936961.76 frames. ], batch size: 51, lr: 3.25e-03, grad_scale: 32.0 +2023-03-26 23:56:50,326 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109640.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:56:56,816 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109650.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:56:59,764 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.018e+02 1.497e+02 1.774e+02 2.103e+02 3.199e+02, threshold=3.548e+02, percent-clipped=0.0 +2023-03-26 23:57:03,291 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=109659.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:57:23,379 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109674.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 23:57:25,125 INFO [finetune.py:976] (4/7) Epoch 20, batch 850, loss[loss=0.2159, simple_loss=0.279, pruned_loss=0.07636, over 4879.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2467, pruned_loss=0.05323, over 941470.48 frames. ], batch size: 32, lr: 3.25e-03, grad_scale: 32.0 +2023-03-26 23:57:40,033 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109693.0, num_to_drop=1, layers_to_drop={1} +2023-03-26 23:57:42,093 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-03-26 23:58:06,023 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=109720.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:58:10,728 INFO [finetune.py:976] (4/7) Epoch 20, batch 900, loss[loss=0.1932, simple_loss=0.2433, pruned_loss=0.07153, over 4833.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2435, pruned_loss=0.05192, over 943749.23 frames. ], batch size: 33, lr: 3.25e-03, grad_scale: 32.0 +2023-03-26 23:58:13,917 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2319, 2.1973, 2.1678, 2.6924, 2.7677, 2.5545, 2.1325, 1.8011], + device='cuda:4'), covar=tensor([0.2358, 0.1917, 0.1723, 0.1478, 0.1864, 0.0997, 0.2227, 0.2022], + device='cuda:4'), in_proj_covar=tensor([0.0247, 0.0211, 0.0213, 0.0195, 0.0246, 0.0189, 0.0217, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:58:22,315 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109741.0, num_to_drop=1, layers_to_drop={0} +2023-03-26 23:58:40,325 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.530e+02 1.823e+02 2.181e+02 3.809e+02, threshold=3.647e+02, percent-clipped=1.0 +2023-03-26 23:58:54,378 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-26 23:58:58,262 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2462, 2.1403, 2.2532, 1.6409, 2.1782, 2.3409, 2.3580, 1.7906], + device='cuda:4'), covar=tensor([0.0521, 0.0585, 0.0620, 0.0810, 0.0638, 0.0591, 0.0516, 0.1073], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0134, 0.0139, 0.0119, 0.0124, 0.0139, 0.0138, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-26 23:59:03,844 INFO [finetune.py:976] (4/7) Epoch 20, batch 950, loss[loss=0.2231, simple_loss=0.2833, pruned_loss=0.08149, over 4219.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2428, pruned_loss=0.0522, over 945670.70 frames. ], batch size: 18, lr: 3.25e-03, grad_scale: 32.0 +2023-03-26 23:59:36,852 INFO [finetune.py:976] (4/7) Epoch 20, batch 1000, loss[loss=0.1967, simple_loss=0.2559, pruned_loss=0.06869, over 4135.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2458, pruned_loss=0.05308, over 946566.43 frames. ], batch size: 65, lr: 3.25e-03, grad_scale: 32.0 +2023-03-26 23:59:39,007 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-26 23:59:52,404 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109850.0, num_to_drop=0, layers_to_drop=set() +2023-03-26 23:59:55,345 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.630e+02 1.951e+02 2.313e+02 5.473e+02, threshold=3.903e+02, percent-clipped=2.0 +2023-03-27 00:00:00,720 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6653, 3.7098, 3.4719, 1.5042, 3.7993, 2.9434, 0.8358, 2.5348], + device='cuda:4'), covar=tensor([0.2461, 0.1818, 0.1571, 0.3646, 0.1119, 0.1013, 0.4440, 0.1478], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0177, 0.0160, 0.0131, 0.0161, 0.0123, 0.0147, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 00:00:08,419 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5886, 1.4696, 1.4341, 1.6018, 1.0988, 3.4308, 1.3506, 1.6506], + device='cuda:4'), covar=tensor([0.3291, 0.2486, 0.2266, 0.2290, 0.1852, 0.0222, 0.2769, 0.1306], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0115, 0.0120, 0.0122, 0.0114, 0.0096, 0.0095, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 00:00:10,589 INFO [finetune.py:976] (4/7) Epoch 20, batch 1050, loss[loss=0.1821, simple_loss=0.246, pruned_loss=0.0591, over 4734.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2503, pruned_loss=0.05419, over 949980.87 frames. ], batch size: 59, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:00:10,883 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 00:00:24,790 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109898.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:00:43,710 INFO [finetune.py:976] (4/7) Epoch 20, batch 1100, loss[loss=0.1741, simple_loss=0.2619, pruned_loss=0.04315, over 4789.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2509, pruned_loss=0.05448, over 951742.40 frames. ], batch size: 51, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:00:49,672 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=109935.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:00:59,778 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109950.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:01:01,344 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-27 00:01:02,702 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.532e+02 1.890e+02 2.271e+02 3.423e+02, threshold=3.780e+02, percent-clipped=0.0 +2023-03-27 00:01:13,322 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-27 00:01:14,936 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7653, 1.2953, 1.9150, 1.7507, 1.6067, 1.5404, 1.7354, 1.7324], + device='cuda:4'), covar=tensor([0.3670, 0.3581, 0.2788, 0.3270, 0.4299, 0.3374, 0.3869, 0.2721], + device='cuda:4'), in_proj_covar=tensor([0.0253, 0.0241, 0.0262, 0.0281, 0.0279, 0.0254, 0.0290, 0.0244], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:01:15,527 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=109974.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:01:17,225 INFO [finetune.py:976] (4/7) Epoch 20, batch 1150, loss[loss=0.1451, simple_loss=0.2265, pruned_loss=0.0318, over 4885.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2519, pruned_loss=0.05464, over 952341.43 frames. ], batch size: 43, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:01:31,933 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=109998.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:01:44,414 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110015.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:01:45,677 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110017.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:01:48,680 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110022.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:01:48,997 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-03-27 00:01:52,168 INFO [finetune.py:976] (4/7) Epoch 20, batch 1200, loss[loss=0.1487, simple_loss=0.2269, pruned_loss=0.03525, over 4808.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2503, pruned_loss=0.05405, over 954271.21 frames. ], batch size: 51, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:02:04,533 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110044.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:02:08,835 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-27 00:02:11,658 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.682e+01 1.527e+02 1.776e+02 2.166e+02 4.163e+02, threshold=3.552e+02, percent-clipped=2.0 +2023-03-27 00:02:32,492 INFO [finetune.py:976] (4/7) Epoch 20, batch 1250, loss[loss=0.2638, simple_loss=0.3038, pruned_loss=0.1119, over 4936.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.248, pruned_loss=0.05387, over 953483.66 frames. ], batch size: 38, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:02:33,238 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110078.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:03:02,204 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110105.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:03:09,740 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.69 vs. limit=5.0 +2023-03-27 00:03:09,985 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110110.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:03:13,064 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110115.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:03:23,865 INFO [finetune.py:976] (4/7) Epoch 20, batch 1300, loss[loss=0.1312, simple_loss=0.1997, pruned_loss=0.03132, over 4812.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2451, pruned_loss=0.05295, over 956086.04 frames. ], batch size: 25, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:03:25,175 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-03-27 00:03:33,295 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0595, 1.9838, 2.0815, 1.3847, 2.0724, 2.1796, 2.1957, 1.6629], + device='cuda:4'), covar=tensor([0.0510, 0.0598, 0.0616, 0.0799, 0.0686, 0.0610, 0.0483, 0.1059], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0135, 0.0141, 0.0121, 0.0125, 0.0140, 0.0140, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:03:45,427 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.072e+02 1.532e+02 1.884e+02 2.318e+02 4.682e+02, threshold=3.767e+02, percent-clipped=2.0 +2023-03-27 00:04:01,787 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6533, 1.5317, 1.5555, 1.5617, 0.9527, 2.9052, 1.1028, 1.5612], + device='cuda:4'), covar=tensor([0.3251, 0.2481, 0.2117, 0.2390, 0.1924, 0.0286, 0.2637, 0.1283], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0115, 0.0119, 0.0122, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 00:04:04,756 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110171.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:04:12,711 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110176.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:04:13,180 INFO [finetune.py:976] (4/7) Epoch 20, batch 1350, loss[loss=0.1887, simple_loss=0.2541, pruned_loss=0.0617, over 4756.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2461, pruned_loss=0.05395, over 954342.60 frames. ], batch size: 27, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:05:04,710 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6522, 1.5088, 1.0010, 0.3216, 1.2774, 1.4439, 1.3265, 1.4294], + device='cuda:4'), covar=tensor([0.0868, 0.0715, 0.1329, 0.1804, 0.1233, 0.2292, 0.2350, 0.0845], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0193, 0.0200, 0.0182, 0.0210, 0.0207, 0.0223, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:05:15,847 INFO [finetune.py:976] (4/7) Epoch 20, batch 1400, loss[loss=0.2089, simple_loss=0.2824, pruned_loss=0.06771, over 4797.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2485, pruned_loss=0.05399, over 954845.29 frames. ], batch size: 45, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:05:26,731 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110235.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:05:48,220 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.073e+02 1.545e+02 1.824e+02 2.176e+02 3.637e+02, threshold=3.648e+02, percent-clipped=0.0 +2023-03-27 00:05:56,185 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.45 vs. limit=5.0 +2023-03-27 00:06:02,045 INFO [finetune.py:976] (4/7) Epoch 20, batch 1450, loss[loss=0.1824, simple_loss=0.2718, pruned_loss=0.04651, over 4799.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2511, pruned_loss=0.05459, over 956200.27 frames. ], batch size: 41, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:06:06,236 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110283.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:06:06,305 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110283.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:06:10,366 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.3118, 3.7278, 3.9792, 4.1423, 4.0691, 3.7877, 4.4005, 1.4460], + device='cuda:4'), covar=tensor([0.0771, 0.0862, 0.0807, 0.1010, 0.1098, 0.1431, 0.0643, 0.5631], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0245, 0.0280, 0.0292, 0.0333, 0.0283, 0.0302, 0.0297], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:06:20,384 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110303.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:06:28,605 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110315.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:06:30,460 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110318.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:06:35,829 INFO [finetune.py:976] (4/7) Epoch 20, batch 1500, loss[loss=0.1499, simple_loss=0.2161, pruned_loss=0.04181, over 3985.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2513, pruned_loss=0.0546, over 953479.32 frames. ], batch size: 17, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:06:47,044 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5040, 2.4057, 1.9166, 0.9691, 2.0457, 1.9042, 1.7998, 2.1795], + device='cuda:4'), covar=tensor([0.1018, 0.0663, 0.1559, 0.1872, 0.1336, 0.2358, 0.2041, 0.0956], + device='cuda:4'), in_proj_covar=tensor([0.0167, 0.0191, 0.0199, 0.0181, 0.0208, 0.0206, 0.0222, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:06:47,641 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110344.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:06:55,111 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.590e+02 1.939e+02 2.244e+02 3.777e+02, threshold=3.878e+02, percent-clipped=2.0 +2023-03-27 00:07:00,486 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110363.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:07:01,158 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110364.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:07:07,072 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110373.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:07:09,449 INFO [finetune.py:976] (4/7) Epoch 20, batch 1550, loss[loss=0.1349, simple_loss=0.2114, pruned_loss=0.02919, over 4711.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2512, pruned_loss=0.05454, over 955002.28 frames. ], batch size: 23, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:07:10,802 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110379.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:07:21,943 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0812, 1.7417, 2.3863, 1.4701, 2.1279, 2.3311, 1.6346, 2.4059], + device='cuda:4'), covar=tensor([0.1370, 0.2106, 0.1372, 0.2238, 0.0898, 0.1552, 0.2922, 0.0971], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0206, 0.0192, 0.0191, 0.0176, 0.0215, 0.0220, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:07:25,519 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110400.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:07:38,433 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3837, 2.2720, 1.9639, 1.0670, 2.0997, 1.8970, 1.7410, 2.1305], + device='cuda:4'), covar=tensor([0.0832, 0.0677, 0.1376, 0.1932, 0.1251, 0.1998, 0.1918, 0.0953], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0191, 0.0199, 0.0181, 0.0208, 0.0206, 0.0222, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:07:42,143 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 00:07:43,128 INFO [finetune.py:976] (4/7) Epoch 20, batch 1600, loss[loss=0.1925, simple_loss=0.2629, pruned_loss=0.06108, over 4903.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2498, pruned_loss=0.05483, over 953869.15 frames. ], batch size: 35, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:08:13,510 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.758e+01 1.502e+02 1.787e+02 2.306e+02 4.709e+02, threshold=3.574e+02, percent-clipped=2.0 +2023-03-27 00:08:29,708 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110466.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:08:33,171 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110471.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:08:40,393 INFO [finetune.py:976] (4/7) Epoch 20, batch 1650, loss[loss=0.1759, simple_loss=0.2338, pruned_loss=0.05898, over 4854.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.247, pruned_loss=0.05428, over 954390.85 frames. ], batch size: 44, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:08:41,713 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3429, 1.3036, 1.3679, 0.7278, 1.3511, 1.3998, 1.3714, 1.2417], + device='cuda:4'), covar=tensor([0.0531, 0.0670, 0.0680, 0.0816, 0.1012, 0.0589, 0.0539, 0.1082], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0134, 0.0139, 0.0119, 0.0124, 0.0138, 0.0139, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:09:24,061 INFO [finetune.py:976] (4/7) Epoch 20, batch 1700, loss[loss=0.2156, simple_loss=0.2751, pruned_loss=0.07804, over 4836.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2455, pruned_loss=0.05412, over 956025.59 frames. ], batch size: 33, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:09:37,992 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4347, 2.2751, 1.9947, 2.4218, 2.2266, 2.2394, 2.1882, 3.1275], + device='cuda:4'), covar=tensor([0.3844, 0.4767, 0.3451, 0.4293, 0.4348, 0.2595, 0.4197, 0.1548], + device='cuda:4'), in_proj_covar=tensor([0.0288, 0.0261, 0.0231, 0.0275, 0.0251, 0.0221, 0.0252, 0.0232], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:09:42,048 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7551, 1.4798, 1.4875, 1.6054, 1.9109, 1.8679, 1.5933, 1.5441], + device='cuda:4'), covar=tensor([0.0361, 0.0344, 0.0565, 0.0336, 0.0267, 0.0419, 0.0385, 0.0372], + device='cuda:4'), in_proj_covar=tensor([0.0097, 0.0107, 0.0144, 0.0111, 0.0100, 0.0111, 0.0100, 0.0111], + device='cuda:4'), out_proj_covar=tensor([7.5184e-05, 8.2552e-05, 1.1353e-04, 8.5551e-05, 7.7848e-05, 8.2262e-05, + 7.4117e-05, 8.4887e-05], device='cuda:4') +2023-03-27 00:09:42,520 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.500e+02 1.773e+02 2.246e+02 3.830e+02, threshold=3.546e+02, percent-clipped=2.0 +2023-03-27 00:09:57,624 INFO [finetune.py:976] (4/7) Epoch 20, batch 1750, loss[loss=0.1556, simple_loss=0.2357, pruned_loss=0.03775, over 4824.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2479, pruned_loss=0.05513, over 954653.31 frames. ], batch size: 30, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:10:10,421 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0303, 2.0058, 2.0138, 1.3650, 2.0805, 2.0366, 2.1390, 1.6920], + device='cuda:4'), covar=tensor([0.0609, 0.0566, 0.0669, 0.0844, 0.0621, 0.0691, 0.0543, 0.1075], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0134, 0.0138, 0.0119, 0.0123, 0.0138, 0.0138, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:10:20,421 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110611.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:10:30,987 INFO [finetune.py:976] (4/7) Epoch 20, batch 1800, loss[loss=0.2127, simple_loss=0.2812, pruned_loss=0.07207, over 4927.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2522, pruned_loss=0.05582, over 956893.89 frames. ], batch size: 38, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:10:38,898 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110639.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:10:47,411 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.67 vs. limit=2.0 +2023-03-27 00:10:53,428 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.5367, 3.8970, 4.1845, 4.3946, 4.2758, 4.0016, 4.6135, 1.6052], + device='cuda:4'), covar=tensor([0.0795, 0.0890, 0.0913, 0.0925, 0.1268, 0.1631, 0.0795, 0.5166], + device='cuda:4'), in_proj_covar=tensor([0.0351, 0.0246, 0.0278, 0.0293, 0.0332, 0.0284, 0.0303, 0.0297], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:10:55,666 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.630e+02 1.887e+02 2.220e+02 3.285e+02, threshold=3.774e+02, percent-clipped=0.0 +2023-03-27 00:10:55,836 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2629, 1.8003, 2.2461, 2.2133, 1.9563, 1.9938, 2.1453, 2.0914], + device='cuda:4'), covar=tensor([0.4311, 0.4246, 0.3400, 0.3903, 0.5239, 0.4092, 0.5043, 0.3281], + device='cuda:4'), in_proj_covar=tensor([0.0252, 0.0240, 0.0262, 0.0281, 0.0278, 0.0253, 0.0289, 0.0243], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:11:01,585 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110659.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:11:03,342 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.7116, 1.4992, 1.4385, 0.8568, 1.5702, 1.7583, 1.7331, 1.3771], + device='cuda:4'), covar=tensor([0.0810, 0.0582, 0.0505, 0.0558, 0.0458, 0.0519, 0.0306, 0.0606], + device='cuda:4'), in_proj_covar=tensor([0.0124, 0.0150, 0.0125, 0.0125, 0.0131, 0.0129, 0.0142, 0.0148], + device='cuda:4'), out_proj_covar=tensor([9.0930e-05, 1.0884e-04, 8.9388e-05, 8.8388e-05, 9.2352e-05, 9.2119e-05, + 1.0179e-04, 1.0589e-04], device='cuda:4') +2023-03-27 00:11:11,032 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110672.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:11:11,632 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110673.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:11:12,217 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110674.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:11:14,002 INFO [finetune.py:976] (4/7) Epoch 20, batch 1850, loss[loss=0.1886, simple_loss=0.2663, pruned_loss=0.05543, over 4805.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2529, pruned_loss=0.0557, over 956897.68 frames. ], batch size: 40, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:11:29,602 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110700.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:11:44,148 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110721.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:11:47,773 INFO [finetune.py:976] (4/7) Epoch 20, batch 1900, loss[loss=0.1692, simple_loss=0.2535, pruned_loss=0.04241, over 4899.00 frames. ], tot_loss[loss=0.1829, simple_loss=0.2536, pruned_loss=0.0561, over 955408.01 frames. ], batch size: 46, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:11:54,352 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110736.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:11:58,354 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-27 00:12:01,611 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110748.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:12:06,289 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.024e+02 1.472e+02 1.826e+02 2.198e+02 3.929e+02, threshold=3.651e+02, percent-clipped=1.0 +2023-03-27 00:12:14,052 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110766.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:12:17,567 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110771.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:12:21,629 INFO [finetune.py:976] (4/7) Epoch 20, batch 1950, loss[loss=0.1739, simple_loss=0.2346, pruned_loss=0.05658, over 4835.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2502, pruned_loss=0.05439, over 955220.99 frames. ], batch size: 30, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:12:23,951 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-27 00:12:34,793 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110797.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 00:12:42,729 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.0218, 2.8661, 2.5780, 3.1341, 2.9527, 2.6318, 3.5102, 3.0810], + device='cuda:4'), covar=tensor([0.1062, 0.1801, 0.2610, 0.2184, 0.2199, 0.1439, 0.2141, 0.1583], + device='cuda:4'), in_proj_covar=tensor([0.0186, 0.0189, 0.0235, 0.0253, 0.0248, 0.0205, 0.0215, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:12:46,184 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110814.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:12:49,734 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110819.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:12:55,562 INFO [finetune.py:976] (4/7) Epoch 20, batch 2000, loss[loss=0.1752, simple_loss=0.2409, pruned_loss=0.05477, over 4931.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.248, pruned_loss=0.05376, over 955145.44 frames. ], batch size: 38, lr: 3.25e-03, grad_scale: 32.0 +2023-03-27 00:13:17,584 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.482e+02 1.740e+02 2.016e+02 2.901e+02, threshold=3.480e+02, percent-clipped=0.0 +2023-03-27 00:13:30,417 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110868.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:13:38,677 INFO [finetune.py:976] (4/7) Epoch 20, batch 2050, loss[loss=0.1443, simple_loss=0.202, pruned_loss=0.04324, over 4303.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2457, pruned_loss=0.05343, over 955195.89 frames. ], batch size: 18, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:14:28,248 INFO [finetune.py:976] (4/7) Epoch 20, batch 2100, loss[loss=0.2196, simple_loss=0.2717, pruned_loss=0.0838, over 4917.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2448, pruned_loss=0.05303, over 955317.93 frames. ], batch size: 36, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:14:29,626 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=110929.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:14:39,987 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110939.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:14:50,629 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.458e+01 1.520e+02 1.862e+02 2.217e+02 3.516e+02, threshold=3.725e+02, percent-clipped=1.0 +2023-03-27 00:14:52,593 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=110958.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:14:53,172 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110959.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:14:58,414 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=110967.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:15:03,729 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=110974.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:15:05,917 INFO [finetune.py:976] (4/7) Epoch 20, batch 2150, loss[loss=0.2274, simple_loss=0.3009, pruned_loss=0.07689, over 4807.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2474, pruned_loss=0.05394, over 953485.74 frames. ], batch size: 39, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:15:12,590 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=110987.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:15:25,801 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111007.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:15:33,630 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=111019.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:15:35,882 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111022.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:15:38,011 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.00 vs. limit=5.0 +2023-03-27 00:15:38,857 INFO [finetune.py:976] (4/7) Epoch 20, batch 2200, loss[loss=0.2088, simple_loss=0.2722, pruned_loss=0.07274, over 4925.00 frames. ], tot_loss[loss=0.1778, simple_loss=0.2488, pruned_loss=0.0534, over 953826.54 frames. ], batch size: 33, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:15:50,039 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.0402, 2.8324, 2.5753, 1.3511, 2.6285, 2.1129, 2.1944, 2.5225], + device='cuda:4'), covar=tensor([0.1184, 0.0788, 0.1923, 0.2241, 0.1896, 0.2461, 0.2034, 0.1240], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0192, 0.0201, 0.0183, 0.0211, 0.0207, 0.0224, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:16:00,235 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.523e+02 1.854e+02 2.195e+02 4.707e+02, threshold=3.708e+02, percent-clipped=2.0 +2023-03-27 00:16:10,018 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7685, 1.8440, 1.4649, 1.8457, 2.1922, 1.9115, 1.5944, 1.4158], + device='cuda:4'), covar=tensor([0.2116, 0.1779, 0.1879, 0.1558, 0.1630, 0.1143, 0.2236, 0.1938], + device='cuda:4'), in_proj_covar=tensor([0.0244, 0.0210, 0.0212, 0.0194, 0.0243, 0.0188, 0.0216, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:16:23,022 INFO [finetune.py:976] (4/7) Epoch 20, batch 2250, loss[loss=0.1624, simple_loss=0.2291, pruned_loss=0.04782, over 4795.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2499, pruned_loss=0.05426, over 955438.72 frames. ], batch size: 26, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:16:27,091 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0042, 1.4840, 0.6934, 1.8772, 2.3907, 1.7475, 1.7247, 1.9731], + device='cuda:4'), covar=tensor([0.1401, 0.1950, 0.2056, 0.1173, 0.1888, 0.1878, 0.1379, 0.1767], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0095, 0.0111, 0.0092, 0.0120, 0.0093, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 00:16:33,699 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111092.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 00:16:56,460 INFO [finetune.py:976] (4/7) Epoch 20, batch 2300, loss[loss=0.1792, simple_loss=0.2392, pruned_loss=0.05962, over 4778.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2503, pruned_loss=0.05355, over 957352.16 frames. ], batch size: 26, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:17:15,912 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.510e+02 1.791e+02 2.077e+02 4.254e+02, threshold=3.582e+02, percent-clipped=2.0 +2023-03-27 00:17:30,238 INFO [finetune.py:976] (4/7) Epoch 20, batch 2350, loss[loss=0.1898, simple_loss=0.2596, pruned_loss=0.05998, over 4924.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2496, pruned_loss=0.05396, over 956832.07 frames. ], batch size: 38, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:18:01,645 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111224.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:18:02,314 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=111225.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:18:03,427 INFO [finetune.py:976] (4/7) Epoch 20, batch 2400, loss[loss=0.1794, simple_loss=0.2552, pruned_loss=0.05178, over 4853.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2469, pruned_loss=0.05347, over 956878.97 frames. ], batch size: 44, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:18:22,739 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.826e+01 1.548e+02 1.775e+02 2.063e+02 3.363e+02, threshold=3.550e+02, percent-clipped=0.0 +2023-03-27 00:18:30,980 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111267.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:18:36,919 INFO [finetune.py:976] (4/7) Epoch 20, batch 2450, loss[loss=0.2071, simple_loss=0.271, pruned_loss=0.07156, over 4902.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2452, pruned_loss=0.05355, over 958249.96 frames. ], batch size: 35, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:18:50,049 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=111286.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:19:21,758 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111314.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:19:22,367 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111315.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:19:23,044 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9723, 1.9727, 1.7104, 2.0897, 2.5084, 2.0171, 2.0165, 1.5194], + device='cuda:4'), covar=tensor([0.2405, 0.2000, 0.2055, 0.1720, 0.1953, 0.1303, 0.2234, 0.2102], + device='cuda:4'), in_proj_covar=tensor([0.0245, 0.0210, 0.0212, 0.0194, 0.0243, 0.0188, 0.0217, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:19:33,327 INFO [finetune.py:976] (4/7) Epoch 20, batch 2500, loss[loss=0.1515, simple_loss=0.2278, pruned_loss=0.03759, over 4778.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2468, pruned_loss=0.05428, over 954868.64 frames. ], batch size: 26, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:19:54,425 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3511, 1.4494, 1.5628, 0.8374, 1.5180, 1.7787, 1.8199, 1.3872], + device='cuda:4'), covar=tensor([0.0877, 0.0572, 0.0494, 0.0519, 0.0430, 0.0593, 0.0316, 0.0728], + device='cuda:4'), in_proj_covar=tensor([0.0125, 0.0151, 0.0125, 0.0125, 0.0131, 0.0129, 0.0142, 0.0148], + device='cuda:4'), out_proj_covar=tensor([9.1185e-05, 1.0913e-04, 8.9385e-05, 8.8454e-05, 9.2091e-05, 9.2528e-05, + 1.0176e-04, 1.0634e-04], device='cuda:4') +2023-03-27 00:20:03,160 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.143e+02 1.591e+02 1.841e+02 2.119e+02 4.112e+02, threshold=3.683e+02, percent-clipped=1.0 +2023-03-27 00:20:17,415 INFO [finetune.py:976] (4/7) Epoch 20, batch 2550, loss[loss=0.1457, simple_loss=0.2163, pruned_loss=0.03749, over 4890.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2489, pruned_loss=0.05411, over 955341.84 frames. ], batch size: 32, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:20:24,835 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-27 00:20:27,523 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111392.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 00:20:51,267 INFO [finetune.py:976] (4/7) Epoch 20, batch 2600, loss[loss=0.1275, simple_loss=0.1944, pruned_loss=0.03028, over 3992.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2495, pruned_loss=0.05407, over 954247.49 frames. ], batch size: 17, lr: 3.24e-03, grad_scale: 64.0 +2023-03-27 00:20:59,729 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111440.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:21:10,163 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.233e+02 1.591e+02 1.878e+02 2.220e+02 5.233e+02, threshold=3.757e+02, percent-clipped=1.0 +2023-03-27 00:21:20,903 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=111468.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:21:31,399 INFO [finetune.py:976] (4/7) Epoch 20, batch 2650, loss[loss=0.1396, simple_loss=0.2146, pruned_loss=0.03228, over 4750.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2501, pruned_loss=0.05374, over 952945.05 frames. ], batch size: 27, lr: 3.24e-03, grad_scale: 64.0 +2023-03-27 00:21:36,681 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4720, 1.1056, 0.7656, 1.3229, 1.8974, 0.7654, 1.1735, 1.3877], + device='cuda:4'), covar=tensor([0.1518, 0.2122, 0.1847, 0.1248, 0.2037, 0.2052, 0.1609, 0.1980], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0111, 0.0091, 0.0120, 0.0093, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 00:21:43,506 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-27 00:22:06,259 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111524.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:22:07,979 INFO [finetune.py:976] (4/7) Epoch 20, batch 2700, loss[loss=0.2168, simple_loss=0.2762, pruned_loss=0.07873, over 4890.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2507, pruned_loss=0.05419, over 955171.95 frames. ], batch size: 35, lr: 3.24e-03, grad_scale: 64.0 +2023-03-27 00:22:10,279 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=111529.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:22:27,297 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.466e+01 1.581e+02 1.832e+02 2.307e+02 3.346e+02, threshold=3.664e+02, percent-clipped=0.0 +2023-03-27 00:22:33,298 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.3531, 3.7817, 3.9803, 4.1926, 4.1143, 3.8765, 4.4641, 1.5305], + device='cuda:4'), covar=tensor([0.0722, 0.0749, 0.0758, 0.0922, 0.1182, 0.1462, 0.0626, 0.5152], + device='cuda:4'), in_proj_covar=tensor([0.0349, 0.0244, 0.0277, 0.0290, 0.0332, 0.0281, 0.0302, 0.0297], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:22:38,119 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111572.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:22:41,153 INFO [finetune.py:976] (4/7) Epoch 20, batch 2750, loss[loss=0.1575, simple_loss=0.2349, pruned_loss=0.04003, over 4869.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2479, pruned_loss=0.05352, over 954604.95 frames. ], batch size: 31, lr: 3.24e-03, grad_scale: 64.0 +2023-03-27 00:22:42,765 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-27 00:22:44,074 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111581.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:22:55,101 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-27 00:23:06,498 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2476, 3.7635, 4.0238, 3.9139, 3.8344, 3.6515, 4.3604, 1.5027], + device='cuda:4'), covar=tensor([0.1302, 0.1384, 0.1334, 0.1851, 0.1992, 0.2227, 0.1153, 0.7651], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0245, 0.0278, 0.0292, 0.0333, 0.0282, 0.0303, 0.0298], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:23:06,521 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111614.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:23:06,565 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1668, 2.0254, 1.7511, 1.9368, 1.8847, 1.8441, 1.9339, 2.6983], + device='cuda:4'), covar=tensor([0.3685, 0.3922, 0.3186, 0.3553, 0.3885, 0.2484, 0.3590, 0.1649], + device='cuda:4'), in_proj_covar=tensor([0.0286, 0.0262, 0.0231, 0.0277, 0.0251, 0.0222, 0.0252, 0.0233], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:23:14,368 INFO [finetune.py:976] (4/7) Epoch 20, batch 2800, loss[loss=0.1443, simple_loss=0.2124, pruned_loss=0.03808, over 4398.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2453, pruned_loss=0.05296, over 953638.38 frames. ], batch size: 19, lr: 3.24e-03, grad_scale: 64.0 +2023-03-27 00:23:32,758 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.029e+01 1.565e+02 1.748e+02 2.177e+02 3.583e+02, threshold=3.496e+02, percent-clipped=0.0 +2023-03-27 00:23:38,054 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111662.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:23:42,133 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9767, 1.7152, 1.9710, 1.2953, 1.9670, 1.9651, 1.9370, 1.5500], + device='cuda:4'), covar=tensor([0.0587, 0.0801, 0.0652, 0.0919, 0.0943, 0.0709, 0.0657, 0.1326], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0137, 0.0141, 0.0121, 0.0125, 0.0141, 0.0142, 0.0164], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:23:48,046 INFO [finetune.py:976] (4/7) Epoch 20, batch 2850, loss[loss=0.1737, simple_loss=0.2535, pruned_loss=0.04697, over 4737.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2436, pruned_loss=0.05251, over 954846.52 frames. ], batch size: 54, lr: 3.24e-03, grad_scale: 64.0 +2023-03-27 00:24:41,628 INFO [finetune.py:976] (4/7) Epoch 20, batch 2900, loss[loss=0.1859, simple_loss=0.2599, pruned_loss=0.05594, over 4857.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2485, pruned_loss=0.05427, over 955621.25 frames. ], batch size: 44, lr: 3.24e-03, grad_scale: 64.0 +2023-03-27 00:25:12,983 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.101e+02 1.542e+02 1.820e+02 2.254e+02 5.949e+02, threshold=3.641e+02, percent-clipped=1.0 +2023-03-27 00:25:31,431 INFO [finetune.py:976] (4/7) Epoch 20, batch 2950, loss[loss=0.1714, simple_loss=0.2443, pruned_loss=0.0492, over 4783.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2524, pruned_loss=0.05506, over 956920.44 frames. ], batch size: 29, lr: 3.24e-03, grad_scale: 64.0 +2023-03-27 00:25:40,083 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=111791.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 00:26:03,394 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=111824.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:26:05,122 INFO [finetune.py:976] (4/7) Epoch 20, batch 3000, loss[loss=0.1978, simple_loss=0.2633, pruned_loss=0.06619, over 4837.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.2545, pruned_loss=0.05611, over 957299.50 frames. ], batch size: 47, lr: 3.24e-03, grad_scale: 64.0 +2023-03-27 00:26:05,122 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 00:26:20,305 INFO [finetune.py:1010] (4/7) Epoch 20, validation: loss=0.1563, simple_loss=0.2257, pruned_loss=0.04344, over 2265189.00 frames. +2023-03-27 00:26:20,306 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-27 00:26:21,980 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-03-27 00:26:31,133 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1649, 2.0552, 1.6612, 1.9721, 2.1644, 1.8547, 2.4003, 2.2201], + device='cuda:4'), covar=tensor([0.1428, 0.2006, 0.3049, 0.2556, 0.2362, 0.1671, 0.3126, 0.1782], + device='cuda:4'), in_proj_covar=tensor([0.0184, 0.0188, 0.0232, 0.0251, 0.0245, 0.0202, 0.0213, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:26:31,253 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-27 00:26:40,753 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=111852.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 00:26:48,435 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.202e+02 1.618e+02 1.875e+02 2.230e+02 3.575e+02, threshold=3.749e+02, percent-clipped=0.0 +2023-03-27 00:27:11,153 INFO [finetune.py:976] (4/7) Epoch 20, batch 3050, loss[loss=0.1939, simple_loss=0.2634, pruned_loss=0.06222, over 4886.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.253, pruned_loss=0.05483, over 957202.26 frames. ], batch size: 43, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:27:18,715 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=111881.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:28:00,386 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7231, 1.2668, 0.9181, 1.5664, 2.0271, 1.3168, 1.4647, 1.5863], + device='cuda:4'), covar=tensor([0.1338, 0.1852, 0.1820, 0.1119, 0.1841, 0.1936, 0.1309, 0.1743], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0094, 0.0111, 0.0092, 0.0119, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 00:28:14,418 INFO [finetune.py:976] (4/7) Epoch 20, batch 3100, loss[loss=0.2099, simple_loss=0.268, pruned_loss=0.07588, over 4815.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2508, pruned_loss=0.05426, over 956354.19 frames. ], batch size: 30, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:28:15,706 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=111929.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:28:32,998 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.507e+02 1.737e+02 2.301e+02 5.151e+02, threshold=3.474e+02, percent-clipped=3.0 +2023-03-27 00:28:47,145 INFO [finetune.py:976] (4/7) Epoch 20, batch 3150, loss[loss=0.1527, simple_loss=0.2261, pruned_loss=0.03966, over 4934.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2477, pruned_loss=0.05304, over 955938.76 frames. ], batch size: 33, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:29:23,073 INFO [finetune.py:976] (4/7) Epoch 20, batch 3200, loss[loss=0.2124, simple_loss=0.2625, pruned_loss=0.08113, over 4849.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2442, pruned_loss=0.05204, over 954521.58 frames. ], batch size: 47, lr: 3.24e-03, grad_scale: 32.0 +2023-03-27 00:29:23,816 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0083, 1.5861, 2.2326, 1.5779, 2.1126, 2.2844, 1.5598, 2.4036], + device='cuda:4'), covar=tensor([0.1239, 0.2085, 0.1329, 0.1881, 0.0828, 0.1268, 0.2847, 0.0777], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0204, 0.0192, 0.0189, 0.0175, 0.0212, 0.0219, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:29:53,983 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7438, 3.7844, 3.6248, 1.6303, 3.8310, 2.9196, 0.9807, 2.8027], + device='cuda:4'), covar=tensor([0.2311, 0.2018, 0.1375, 0.3593, 0.1086, 0.0916, 0.4299, 0.1470], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0178, 0.0161, 0.0131, 0.0163, 0.0123, 0.0147, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 00:29:56,972 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.705e+01 1.565e+02 1.739e+02 2.228e+02 3.922e+02, threshold=3.479e+02, percent-clipped=1.0 +2023-03-27 00:30:06,305 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112066.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:30:13,835 INFO [finetune.py:976] (4/7) Epoch 20, batch 3250, loss[loss=0.182, simple_loss=0.2516, pruned_loss=0.05622, over 4936.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2457, pruned_loss=0.05324, over 954824.09 frames. ], batch size: 33, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:30:35,795 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-27 00:30:43,275 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3043, 1.3386, 1.6214, 1.4898, 1.4718, 2.9228, 1.2608, 1.4787], + device='cuda:4'), covar=tensor([0.1002, 0.1812, 0.1124, 0.0999, 0.1597, 0.0275, 0.1530, 0.1769], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0082, 0.0074, 0.0077, 0.0092, 0.0081, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 00:30:54,518 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=112124.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:30:56,198 INFO [finetune.py:976] (4/7) Epoch 20, batch 3300, loss[loss=0.1491, simple_loss=0.223, pruned_loss=0.03755, over 4773.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2481, pruned_loss=0.05333, over 955199.98 frames. ], batch size: 26, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:30:56,332 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112127.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:31:10,260 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112147.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 00:31:16,114 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.319e+01 1.653e+02 1.982e+02 2.472e+02 3.934e+02, threshold=3.965e+02, percent-clipped=2.0 +2023-03-27 00:31:26,939 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=112172.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:31:29,979 INFO [finetune.py:976] (4/7) Epoch 20, batch 3350, loss[loss=0.2038, simple_loss=0.2721, pruned_loss=0.06775, over 4910.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2494, pruned_loss=0.05357, over 953751.47 frames. ], batch size: 37, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:31:45,305 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1566, 1.2475, 1.4083, 0.4839, 1.3353, 1.5982, 1.5924, 1.2931], + device='cuda:4'), covar=tensor([0.1136, 0.0847, 0.0583, 0.0733, 0.0607, 0.0734, 0.0459, 0.0913], + device='cuda:4'), in_proj_covar=tensor([0.0124, 0.0150, 0.0125, 0.0125, 0.0131, 0.0129, 0.0141, 0.0148], + device='cuda:4'), out_proj_covar=tensor([9.0653e-05, 1.0863e-04, 8.9327e-05, 8.8176e-05, 9.2119e-05, 9.2539e-05, + 1.0149e-04, 1.0617e-04], device='cuda:4') +2023-03-27 00:31:45,730 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-27 00:31:52,716 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9421, 1.7624, 1.6661, 1.7589, 2.1851, 2.1206, 1.8261, 1.7828], + device='cuda:4'), covar=tensor([0.0315, 0.0324, 0.0597, 0.0320, 0.0238, 0.0526, 0.0342, 0.0359], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0108, 0.0146, 0.0112, 0.0101, 0.0113, 0.0100, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.5646e-05, 8.3081e-05, 1.1471e-04, 8.6040e-05, 7.8672e-05, 8.3303e-05, + 7.4785e-05, 8.6706e-05], device='cuda:4') +2023-03-27 00:32:11,816 INFO [finetune.py:976] (4/7) Epoch 20, batch 3400, loss[loss=0.1689, simple_loss=0.2361, pruned_loss=0.05083, over 4151.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2501, pruned_loss=0.0536, over 953919.05 frames. ], batch size: 65, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:32:24,157 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-27 00:32:27,701 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.49 vs. limit=5.0 +2023-03-27 00:32:29,474 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 00:32:31,202 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.676e+02 1.964e+02 2.392e+02 4.564e+02, threshold=3.928e+02, percent-clipped=2.0 +2023-03-27 00:32:34,355 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.0722, 2.5227, 3.0359, 1.9343, 2.8709, 3.1121, 2.2964, 3.3077], + device='cuda:4'), covar=tensor([0.1215, 0.1723, 0.1525, 0.2253, 0.0948, 0.1334, 0.2587, 0.0832], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0205, 0.0192, 0.0190, 0.0175, 0.0213, 0.0219, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:32:44,382 INFO [finetune.py:976] (4/7) Epoch 20, batch 3450, loss[loss=0.1724, simple_loss=0.2482, pruned_loss=0.04829, over 4799.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2494, pruned_loss=0.0529, over 955487.21 frames. ], batch size: 29, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:32:55,948 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-27 00:33:19,431 INFO [finetune.py:976] (4/7) Epoch 20, batch 3500, loss[loss=0.247, simple_loss=0.304, pruned_loss=0.09495, over 4253.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2476, pruned_loss=0.05245, over 954355.43 frames. ], batch size: 65, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:33:56,445 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.856e+01 1.454e+02 1.799e+02 2.150e+02 5.052e+02, threshold=3.598e+02, percent-clipped=2.0 +2023-03-27 00:34:18,959 INFO [finetune.py:976] (4/7) Epoch 20, batch 3550, loss[loss=0.1677, simple_loss=0.232, pruned_loss=0.05164, over 4760.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.246, pruned_loss=0.05255, over 954569.50 frames. ], batch size: 26, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:34:36,596 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112389.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:35:17,939 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112422.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:35:20,889 INFO [finetune.py:976] (4/7) Epoch 20, batch 3600, loss[loss=0.162, simple_loss=0.2189, pruned_loss=0.05262, over 4823.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.243, pruned_loss=0.05176, over 951770.25 frames. ], batch size: 39, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:35:37,947 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=112447.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 00:35:39,842 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112450.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:35:44,348 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.527e+02 1.833e+02 2.137e+02 4.874e+02, threshold=3.667e+02, percent-clipped=1.0 +2023-03-27 00:36:00,111 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6105, 1.4556, 1.4653, 1.5434, 1.0107, 3.6352, 1.3459, 1.8400], + device='cuda:4'), covar=tensor([0.3484, 0.2660, 0.2262, 0.2483, 0.1889, 0.0156, 0.2804, 0.1297], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0115, 0.0120, 0.0123, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 00:36:08,333 INFO [finetune.py:976] (4/7) Epoch 20, batch 3650, loss[loss=0.174, simple_loss=0.2552, pruned_loss=0.0464, over 4835.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2455, pruned_loss=0.05263, over 954166.88 frames. ], batch size: 49, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:36:20,409 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=112495.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 00:36:38,758 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-27 00:36:41,251 INFO [finetune.py:976] (4/7) Epoch 20, batch 3700, loss[loss=0.1964, simple_loss=0.2709, pruned_loss=0.06098, over 4742.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2496, pruned_loss=0.05376, over 952847.13 frames. ], batch size: 27, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:37:01,298 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.006e+02 1.568e+02 1.941e+02 2.323e+02 3.962e+02, threshold=3.882e+02, percent-clipped=1.0 +2023-03-27 00:37:05,005 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0513, 2.1027, 1.8005, 2.2399, 2.8608, 2.1507, 2.1669, 1.5864], + device='cuda:4'), covar=tensor([0.2120, 0.1854, 0.1772, 0.1529, 0.1508, 0.1095, 0.1855, 0.1841], + device='cuda:4'), in_proj_covar=tensor([0.0244, 0.0210, 0.0212, 0.0194, 0.0243, 0.0188, 0.0217, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:37:05,590 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112562.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:37:10,849 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.9811, 0.9002, 0.9019, 1.0392, 1.1690, 1.1137, 0.9942, 0.9186], + device='cuda:4'), covar=tensor([0.0389, 0.0320, 0.0580, 0.0300, 0.0268, 0.0464, 0.0349, 0.0409], + device='cuda:4'), in_proj_covar=tensor([0.0097, 0.0107, 0.0144, 0.0111, 0.0100, 0.0111, 0.0099, 0.0112], + device='cuda:4'), out_proj_covar=tensor([7.5094e-05, 8.2293e-05, 1.1336e-04, 8.5197e-05, 7.7940e-05, 8.2365e-05, + 7.4049e-05, 8.5713e-05], device='cuda:4') +2023-03-27 00:37:15,915 INFO [finetune.py:976] (4/7) Epoch 20, batch 3750, loss[loss=0.167, simple_loss=0.2427, pruned_loss=0.04561, over 4892.00 frames. ], tot_loss[loss=0.1797, simple_loss=0.2508, pruned_loss=0.05432, over 954081.19 frames. ], batch size: 35, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:37:16,035 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112577.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:37:50,119 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112617.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:37:50,414 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.10 vs. limit=5.0 +2023-03-27 00:37:54,227 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112623.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:37:56,882 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8174, 4.7734, 4.5406, 2.5233, 4.8779, 3.7101, 0.8949, 3.4716], + device='cuda:4'), covar=tensor([0.2255, 0.1789, 0.1338, 0.3013, 0.0890, 0.0869, 0.4699, 0.1256], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0177, 0.0160, 0.0129, 0.0162, 0.0123, 0.0147, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 00:37:57,419 INFO [finetune.py:976] (4/7) Epoch 20, batch 3800, loss[loss=0.1481, simple_loss=0.2252, pruned_loss=0.03555, over 4829.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2521, pruned_loss=0.05474, over 954523.16 frames. ], batch size: 30, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:38:03,776 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-27 00:38:04,232 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112638.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 00:38:16,045 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.507e+01 1.515e+02 1.800e+02 2.233e+02 3.828e+02, threshold=3.600e+02, percent-clipped=0.0 +2023-03-27 00:38:30,667 INFO [finetune.py:976] (4/7) Epoch 20, batch 3850, loss[loss=0.1635, simple_loss=0.2442, pruned_loss=0.04144, over 4840.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.25, pruned_loss=0.05387, over 953656.55 frames. ], batch size: 47, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:38:31,390 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112678.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 00:38:59,757 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=112722.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:39:03,156 INFO [finetune.py:976] (4/7) Epoch 20, batch 3900, loss[loss=0.1785, simple_loss=0.2462, pruned_loss=0.05541, over 4764.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2471, pruned_loss=0.05285, over 952897.47 frames. ], batch size: 59, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:39:15,037 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112745.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:39:21,554 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.403e+01 1.612e+02 1.959e+02 2.410e+02 4.123e+02, threshold=3.918e+02, percent-clipped=1.0 +2023-03-27 00:39:32,077 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=112770.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:39:36,268 INFO [finetune.py:976] (4/7) Epoch 20, batch 3950, loss[loss=0.1784, simple_loss=0.2405, pruned_loss=0.05809, over 4813.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2435, pruned_loss=0.05171, over 953940.73 frames. ], batch size: 51, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:39:51,977 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9819, 2.6727, 2.5263, 1.3748, 2.7328, 2.1787, 2.1176, 2.4767], + device='cuda:4'), covar=tensor([0.0926, 0.0637, 0.1574, 0.1909, 0.1210, 0.2034, 0.1758, 0.0961], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0191, 0.0198, 0.0181, 0.0210, 0.0208, 0.0221, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:40:18,210 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.48 vs. limit=5.0 +2023-03-27 00:40:19,042 INFO [finetune.py:976] (4/7) Epoch 20, batch 4000, loss[loss=0.1832, simple_loss=0.2494, pruned_loss=0.05854, over 4040.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2431, pruned_loss=0.05205, over 953472.59 frames. ], batch size: 17, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:40:48,438 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.638e+02 1.904e+02 2.320e+02 3.891e+02, threshold=3.808e+02, percent-clipped=0.0 +2023-03-27 00:40:50,446 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3262, 2.1744, 1.8466, 2.1893, 2.1139, 2.0643, 2.1637, 2.7915], + device='cuda:4'), covar=tensor([0.3830, 0.4691, 0.3383, 0.4006, 0.4237, 0.2562, 0.3760, 0.1912], + device='cuda:4'), in_proj_covar=tensor([0.0286, 0.0263, 0.0231, 0.0276, 0.0253, 0.0222, 0.0251, 0.0233], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:40:56,972 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-03-27 00:41:04,949 INFO [finetune.py:976] (4/7) Epoch 20, batch 4050, loss[loss=0.2118, simple_loss=0.289, pruned_loss=0.06733, over 4744.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2474, pruned_loss=0.05377, over 955095.37 frames. ], batch size: 54, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:41:23,157 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5614, 3.7193, 3.4833, 1.8221, 3.8642, 2.8602, 0.9468, 2.6206], + device='cuda:4'), covar=tensor([0.2450, 0.2191, 0.1517, 0.3223, 0.1106, 0.1015, 0.4347, 0.1350], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0176, 0.0160, 0.0129, 0.0161, 0.0123, 0.0147, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 00:41:40,680 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0758, 1.9048, 1.6610, 1.7973, 1.8662, 1.8164, 1.8844, 2.5258], + device='cuda:4'), covar=tensor([0.3587, 0.4250, 0.3175, 0.3796, 0.4077, 0.2348, 0.3957, 0.1770], + device='cuda:4'), in_proj_covar=tensor([0.0287, 0.0263, 0.0232, 0.0277, 0.0253, 0.0223, 0.0252, 0.0234], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:41:41,261 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112918.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:41:43,130 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=112921.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:41:47,212 INFO [finetune.py:976] (4/7) Epoch 20, batch 4100, loss[loss=0.1293, simple_loss=0.205, pruned_loss=0.02686, over 4763.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2488, pruned_loss=0.05374, over 952551.90 frames. ], batch size: 26, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:41:51,356 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112933.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 00:42:06,082 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.570e+02 1.869e+02 2.355e+02 4.214e+02, threshold=3.739e+02, percent-clipped=0.0 +2023-03-27 00:42:07,470 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0556, 2.0543, 1.8194, 2.1986, 2.6951, 2.1128, 2.0432, 1.6367], + device='cuda:4'), covar=tensor([0.2270, 0.2015, 0.1983, 0.1671, 0.1754, 0.1207, 0.2082, 0.1933], + device='cuda:4'), in_proj_covar=tensor([0.0244, 0.0209, 0.0211, 0.0193, 0.0241, 0.0187, 0.0216, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:42:17,469 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=112973.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 00:42:19,820 INFO [finetune.py:976] (4/7) Epoch 20, batch 4150, loss[loss=0.1493, simple_loss=0.2053, pruned_loss=0.0466, over 4264.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2518, pruned_loss=0.05495, over 953716.79 frames. ], batch size: 18, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:42:24,009 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=112982.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:43:03,966 INFO [finetune.py:976] (4/7) Epoch 20, batch 4200, loss[loss=0.2161, simple_loss=0.291, pruned_loss=0.07062, over 4307.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.253, pruned_loss=0.05522, over 955456.21 frames. ], batch size: 66, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:43:16,410 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113045.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:43:23,436 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.027e+02 1.578e+02 1.900e+02 2.258e+02 5.235e+02, threshold=3.800e+02, percent-clipped=4.0 +2023-03-27 00:43:36,993 INFO [finetune.py:976] (4/7) Epoch 20, batch 4250, loss[loss=0.162, simple_loss=0.2322, pruned_loss=0.04587, over 4715.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2514, pruned_loss=0.05481, over 957197.85 frames. ], batch size: 54, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:43:47,718 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113093.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:44:10,286 INFO [finetune.py:976] (4/7) Epoch 20, batch 4300, loss[loss=0.1524, simple_loss=0.2253, pruned_loss=0.03979, over 4908.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2487, pruned_loss=0.05374, over 957500.67 frames. ], batch size: 36, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:44:30,855 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.679e+01 1.396e+02 1.741e+02 2.023e+02 4.034e+02, threshold=3.482e+02, percent-clipped=1.0 +2023-03-27 00:44:43,636 INFO [finetune.py:976] (4/7) Epoch 20, batch 4350, loss[loss=0.1521, simple_loss=0.2262, pruned_loss=0.03904, over 4747.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2456, pruned_loss=0.0535, over 954750.07 frames. ], batch size: 28, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:45:12,212 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113218.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:45:17,619 INFO [finetune.py:976] (4/7) Epoch 20, batch 4400, loss[loss=0.2294, simple_loss=0.2944, pruned_loss=0.08219, over 4906.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.246, pruned_loss=0.05355, over 955483.40 frames. ], batch size: 36, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:45:21,834 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113233.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:45:22,447 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=113234.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:45:33,056 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-27 00:45:49,232 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.729e+01 1.603e+02 1.846e+02 2.173e+02 5.642e+02, threshold=3.692e+02, percent-clipped=4.0 +2023-03-27 00:46:00,752 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113266.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:46:08,337 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113273.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 00:46:10,701 INFO [finetune.py:976] (4/7) Epoch 20, batch 4450, loss[loss=0.2127, simple_loss=0.2915, pruned_loss=0.06699, over 4831.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2499, pruned_loss=0.0544, over 956258.35 frames. ], batch size: 33, lr: 3.23e-03, grad_scale: 32.0 +2023-03-27 00:46:10,768 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=113277.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:46:13,203 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113281.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:46:23,174 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=113295.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:46:32,667 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6751, 1.5741, 1.9215, 1.2761, 1.7568, 1.8948, 1.4773, 2.0585], + device='cuda:4'), covar=tensor([0.1153, 0.2005, 0.1236, 0.1736, 0.0825, 0.1214, 0.3132, 0.0685], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0205, 0.0190, 0.0190, 0.0175, 0.0213, 0.0219, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:46:39,689 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5066, 1.3931, 1.0325, 0.4255, 1.1649, 1.3477, 1.2658, 1.3511], + device='cuda:4'), covar=tensor([0.0757, 0.0635, 0.1002, 0.1489, 0.1151, 0.1712, 0.1869, 0.0735], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0192, 0.0200, 0.0182, 0.0210, 0.0209, 0.0223, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:46:50,260 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113321.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:46:53,842 INFO [finetune.py:976] (4/7) Epoch 20, batch 4500, loss[loss=0.2285, simple_loss=0.28, pruned_loss=0.08855, over 4259.00 frames. ], tot_loss[loss=0.1801, simple_loss=0.2508, pruned_loss=0.05472, over 954373.32 frames. ], batch size: 65, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:46:56,402 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1989, 2.6370, 2.5184, 1.2489, 2.7860, 2.1543, 0.8536, 1.8273], + device='cuda:4'), covar=tensor([0.2856, 0.2165, 0.2002, 0.3084, 0.1484, 0.1031, 0.3517, 0.1597], + device='cuda:4'), in_proj_covar=tensor([0.0154, 0.0178, 0.0161, 0.0130, 0.0162, 0.0124, 0.0148, 0.0125], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 00:47:13,425 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.617e+02 2.027e+02 2.372e+02 5.258e+02, threshold=4.055e+02, percent-clipped=2.0 +2023-03-27 00:47:27,581 INFO [finetune.py:976] (4/7) Epoch 20, batch 4550, loss[loss=0.1867, simple_loss=0.2665, pruned_loss=0.05342, over 4918.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2522, pruned_loss=0.05463, over 956026.85 frames. ], batch size: 42, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:47:40,405 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.49 vs. limit=5.0 +2023-03-27 00:47:47,441 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6964, 1.7296, 1.7094, 1.0335, 1.8797, 2.1220, 1.9479, 1.4794], + device='cuda:4'), covar=tensor([0.1014, 0.0642, 0.0561, 0.0592, 0.0432, 0.0584, 0.0349, 0.0790], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0149, 0.0124, 0.0123, 0.0129, 0.0128, 0.0140, 0.0147], + device='cuda:4'), out_proj_covar=tensor([8.9582e-05, 1.0771e-04, 8.8539e-05, 8.6679e-05, 9.0437e-05, 9.1342e-05, + 1.0022e-04, 1.0523e-04], device='cuda:4') +2023-03-27 00:47:59,880 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-27 00:48:03,326 INFO [finetune.py:976] (4/7) Epoch 20, batch 4600, loss[loss=0.2003, simple_loss=0.255, pruned_loss=0.07283, over 4906.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.251, pruned_loss=0.05401, over 955664.34 frames. ], batch size: 32, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:48:31,337 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.757e+01 1.614e+02 1.832e+02 2.145e+02 3.668e+02, threshold=3.664e+02, percent-clipped=0.0 +2023-03-27 00:48:45,584 INFO [finetune.py:976] (4/7) Epoch 20, batch 4650, loss[loss=0.1956, simple_loss=0.2601, pruned_loss=0.06558, over 4904.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2488, pruned_loss=0.0538, over 956003.72 frames. ], batch size: 35, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:49:19,010 INFO [finetune.py:976] (4/7) Epoch 20, batch 4700, loss[loss=0.137, simple_loss=0.2221, pruned_loss=0.02596, over 4766.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2449, pruned_loss=0.05218, over 954906.95 frames. ], batch size: 28, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:49:37,208 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.713e+01 1.518e+02 1.819e+02 2.172e+02 5.123e+02, threshold=3.639e+02, percent-clipped=2.0 +2023-03-27 00:49:51,489 INFO [finetune.py:976] (4/7) Epoch 20, batch 4750, loss[loss=0.1841, simple_loss=0.2617, pruned_loss=0.05322, over 4853.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2423, pruned_loss=0.05117, over 956616.84 frames. ], batch size: 47, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:49:52,071 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113577.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:50:00,073 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=113590.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:50:07,430 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.65 vs. limit=5.0 +2023-03-27 00:50:10,350 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-03-27 00:50:18,366 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=113617.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:50:23,693 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113625.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:50:24,845 INFO [finetune.py:976] (4/7) Epoch 20, batch 4800, loss[loss=0.2245, simple_loss=0.2981, pruned_loss=0.07551, over 4911.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2452, pruned_loss=0.05252, over 955544.47 frames. ], batch size: 42, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:50:30,820 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4680, 1.6002, 1.2876, 1.5341, 1.8964, 1.7064, 1.5904, 1.3678], + device='cuda:4'), covar=tensor([0.0385, 0.0322, 0.0578, 0.0298, 0.0226, 0.0526, 0.0357, 0.0404], + device='cuda:4'), in_proj_covar=tensor([0.0097, 0.0107, 0.0144, 0.0111, 0.0100, 0.0111, 0.0100, 0.0112], + device='cuda:4'), out_proj_covar=tensor([7.5126e-05, 8.2368e-05, 1.1318e-04, 8.5082e-05, 7.7839e-05, 8.2069e-05, + 7.4501e-05, 8.5448e-05], device='cuda:4') +2023-03-27 00:50:37,454 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=113646.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:50:43,816 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.144e+02 1.583e+02 1.918e+02 2.188e+02 4.674e+02, threshold=3.836e+02, percent-clipped=2.0 +2023-03-27 00:51:01,326 INFO [finetune.py:976] (4/7) Epoch 20, batch 4850, loss[loss=0.1864, simple_loss=0.2604, pruned_loss=0.05623, over 4849.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2489, pruned_loss=0.05394, over 954492.89 frames. ], batch size: 47, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:51:02,040 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=113678.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:51:14,816 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6526, 1.3415, 2.0767, 3.3805, 2.2609, 2.4667, 0.9074, 2.8337], + device='cuda:4'), covar=tensor([0.1671, 0.1530, 0.1376, 0.0600, 0.0826, 0.1397, 0.1946, 0.0431], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0117, 0.0134, 0.0165, 0.0101, 0.0137, 0.0125, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 00:51:34,440 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=113707.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:51:57,946 INFO [finetune.py:976] (4/7) Epoch 20, batch 4900, loss[loss=0.1568, simple_loss=0.2339, pruned_loss=0.03986, over 4800.00 frames. ], tot_loss[loss=0.1796, simple_loss=0.2503, pruned_loss=0.05442, over 954726.70 frames. ], batch size: 29, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:52:20,128 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.649e+02 2.053e+02 2.498e+02 4.513e+02, threshold=4.105e+02, percent-clipped=3.0 +2023-03-27 00:52:34,817 INFO [finetune.py:976] (4/7) Epoch 20, batch 4950, loss[loss=0.1927, simple_loss=0.2622, pruned_loss=0.06157, over 4825.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2522, pruned_loss=0.05508, over 954982.27 frames. ], batch size: 30, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:53:07,555 INFO [finetune.py:976] (4/7) Epoch 20, batch 5000, loss[loss=0.1862, simple_loss=0.2396, pruned_loss=0.06639, over 4825.00 frames. ], tot_loss[loss=0.1798, simple_loss=0.2507, pruned_loss=0.05443, over 955817.09 frames. ], batch size: 30, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:53:26,538 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 1.516e+02 1.779e+02 2.023e+02 5.156e+02, threshold=3.559e+02, percent-clipped=2.0 +2023-03-27 00:53:39,282 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4256, 2.1322, 2.7240, 1.5407, 2.2980, 2.5052, 1.9720, 2.7158], + device='cuda:4'), covar=tensor([0.1263, 0.1827, 0.1632, 0.2241, 0.0917, 0.1585, 0.2592, 0.0849], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0206, 0.0190, 0.0189, 0.0175, 0.0213, 0.0219, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:53:42,068 INFO [finetune.py:976] (4/7) Epoch 20, batch 5050, loss[loss=0.1863, simple_loss=0.2451, pruned_loss=0.06372, over 4117.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2487, pruned_loss=0.05421, over 957297.84 frames. ], batch size: 65, lr: 3.22e-03, grad_scale: 64.0 +2023-03-27 00:53:51,010 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=113890.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:54:08,487 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3998, 1.3855, 1.2691, 1.3931, 1.6655, 1.5913, 1.3414, 1.2464], + device='cuda:4'), covar=tensor([0.0324, 0.0282, 0.0602, 0.0286, 0.0228, 0.0453, 0.0339, 0.0383], + device='cuda:4'), in_proj_covar=tensor([0.0097, 0.0107, 0.0143, 0.0111, 0.0100, 0.0111, 0.0100, 0.0112], + device='cuda:4'), out_proj_covar=tensor([7.4825e-05, 8.2160e-05, 1.1248e-04, 8.4814e-05, 7.7456e-05, 8.1871e-05, + 7.4082e-05, 8.5356e-05], device='cuda:4') +2023-03-27 00:54:14,769 INFO [finetune.py:976] (4/7) Epoch 20, batch 5100, loss[loss=0.1909, simple_loss=0.2549, pruned_loss=0.06348, over 4893.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2443, pruned_loss=0.0524, over 956810.32 frames. ], batch size: 35, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:54:23,012 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=113938.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:54:35,763 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.073e+02 1.567e+02 1.826e+02 2.193e+02 3.507e+02, threshold=3.652e+02, percent-clipped=0.0 +2023-03-27 00:54:38,924 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6473, 1.6983, 2.2227, 1.9064, 1.7607, 4.0871, 1.5344, 1.7257], + device='cuda:4'), covar=tensor([0.0871, 0.1654, 0.1167, 0.0921, 0.1520, 0.0203, 0.1456, 0.1725], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0082, 0.0075, 0.0077, 0.0092, 0.0081, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 00:54:40,128 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6011, 1.4760, 1.9306, 3.2277, 2.2212, 2.2478, 1.2499, 2.7472], + device='cuda:4'), covar=tensor([0.1781, 0.1481, 0.1422, 0.0752, 0.0820, 0.1344, 0.1672, 0.0468], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0117, 0.0135, 0.0166, 0.0101, 0.0138, 0.0126, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 00:54:46,070 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=113973.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:54:48,407 INFO [finetune.py:976] (4/7) Epoch 20, batch 5150, loss[loss=0.1861, simple_loss=0.2569, pruned_loss=0.05766, over 4769.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2447, pruned_loss=0.05265, over 957408.49 frames. ], batch size: 28, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:55:07,781 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=114002.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:55:07,869 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8276, 1.2428, 1.8571, 1.8451, 1.6521, 1.5795, 1.7321, 1.6959], + device='cuda:4'), covar=tensor([0.3681, 0.3712, 0.3194, 0.3135, 0.4491, 0.3470, 0.4170, 0.2837], + device='cuda:4'), in_proj_covar=tensor([0.0254, 0.0243, 0.0263, 0.0282, 0.0279, 0.0256, 0.0290, 0.0245], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:55:23,295 INFO [finetune.py:976] (4/7) Epoch 20, batch 5200, loss[loss=0.182, simple_loss=0.2624, pruned_loss=0.05083, over 4724.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.249, pruned_loss=0.05458, over 957124.59 frames. ], batch size: 23, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:55:43,820 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.106e+02 1.581e+02 1.814e+02 2.243e+02 3.815e+02, threshold=3.628e+02, percent-clipped=1.0 +2023-03-27 00:55:56,414 INFO [finetune.py:976] (4/7) Epoch 20, batch 5250, loss[loss=0.2082, simple_loss=0.2745, pruned_loss=0.07095, over 4921.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2514, pruned_loss=0.05539, over 956301.62 frames. ], batch size: 38, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:56:07,015 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6468, 1.5002, 0.9381, 0.2966, 1.3323, 1.4957, 1.3330, 1.4912], + device='cuda:4'), covar=tensor([0.0841, 0.0909, 0.1468, 0.2030, 0.1335, 0.2042, 0.2503, 0.0862], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0191, 0.0200, 0.0182, 0.0210, 0.0209, 0.0224, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:56:11,910 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-27 00:56:36,234 INFO [finetune.py:976] (4/7) Epoch 20, batch 5300, loss[loss=0.175, simple_loss=0.2588, pruned_loss=0.04558, over 4886.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2516, pruned_loss=0.05504, over 956011.96 frames. ], batch size: 43, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:56:36,358 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=114127.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:57:12,907 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7977, 1.3148, 1.8721, 1.7935, 1.6015, 1.5604, 1.7348, 1.7239], + device='cuda:4'), covar=tensor([0.3864, 0.3677, 0.3032, 0.3561, 0.4662, 0.3864, 0.4056, 0.2875], + device='cuda:4'), in_proj_covar=tensor([0.0253, 0.0241, 0.0261, 0.0280, 0.0278, 0.0255, 0.0288, 0.0244], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:57:13,340 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.589e+01 1.501e+02 1.834e+02 2.338e+02 4.244e+02, threshold=3.669e+02, percent-clipped=1.0 +2023-03-27 00:57:30,109 INFO [finetune.py:976] (4/7) Epoch 20, batch 5350, loss[loss=0.1685, simple_loss=0.2426, pruned_loss=0.04721, over 4892.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2514, pruned_loss=0.05436, over 956486.05 frames. ], batch size: 43, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:57:37,351 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=114188.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:57:57,981 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.54 vs. limit=5.0 +2023-03-27 00:58:03,281 INFO [finetune.py:976] (4/7) Epoch 20, batch 5400, loss[loss=0.1902, simple_loss=0.2604, pruned_loss=0.06002, over 4905.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2496, pruned_loss=0.05401, over 954992.32 frames. ], batch size: 43, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:58:23,337 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.275e+01 1.521e+02 1.786e+02 2.103e+02 5.074e+02, threshold=3.573e+02, percent-clipped=1.0 +2023-03-27 00:58:33,642 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=114273.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:58:34,874 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4117, 2.2584, 2.3443, 1.6297, 2.2728, 2.4431, 2.4734, 1.8884], + device='cuda:4'), covar=tensor([0.0567, 0.0568, 0.0644, 0.0897, 0.0633, 0.0656, 0.0563, 0.1082], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0136, 0.0140, 0.0121, 0.0125, 0.0139, 0.0140, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:58:35,980 INFO [finetune.py:976] (4/7) Epoch 20, batch 5450, loss[loss=0.1593, simple_loss=0.2291, pruned_loss=0.04479, over 4762.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2474, pruned_loss=0.05353, over 956175.24 frames. ], batch size: 27, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:58:51,607 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=114302.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:58:53,292 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.7518, 3.2725, 3.4707, 3.5878, 3.5312, 3.3456, 3.8441, 1.2630], + device='cuda:4'), covar=tensor([0.0974, 0.0913, 0.0976, 0.1329, 0.1318, 0.1432, 0.0873, 0.5204], + device='cuda:4'), in_proj_covar=tensor([0.0348, 0.0242, 0.0278, 0.0291, 0.0330, 0.0284, 0.0303, 0.0297], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:59:05,065 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=114321.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:59:08,649 INFO [finetune.py:976] (4/7) Epoch 20, batch 5500, loss[loss=0.1671, simple_loss=0.2361, pruned_loss=0.04903, over 4820.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2446, pruned_loss=0.05271, over 956973.41 frames. ], batch size: 39, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:59:16,049 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6874, 2.4733, 2.0456, 1.0733, 2.2365, 2.0865, 1.8876, 2.2547], + device='cuda:4'), covar=tensor([0.1056, 0.0797, 0.1807, 0.2090, 0.1370, 0.2290, 0.2139, 0.0999], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0191, 0.0199, 0.0182, 0.0209, 0.0209, 0.0223, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 00:59:23,096 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=114350.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 00:59:27,722 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.653e+01 1.524e+02 1.755e+02 2.254e+02 4.886e+02, threshold=3.510e+02, percent-clipped=3.0 +2023-03-27 00:59:42,372 INFO [finetune.py:976] (4/7) Epoch 20, batch 5550, loss[loss=0.1589, simple_loss=0.2381, pruned_loss=0.03989, over 4778.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2464, pruned_loss=0.0536, over 958119.27 frames. ], batch size: 29, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 00:59:49,171 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6789, 3.4668, 3.3557, 1.6307, 3.5845, 2.8483, 0.9781, 2.4737], + device='cuda:4'), covar=tensor([0.2568, 0.1698, 0.1516, 0.3302, 0.1118, 0.0882, 0.4104, 0.1434], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0177, 0.0160, 0.0130, 0.0161, 0.0123, 0.0148, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 01:00:00,266 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-03-27 01:00:14,062 INFO [finetune.py:976] (4/7) Epoch 20, batch 5600, loss[loss=0.155, simple_loss=0.2356, pruned_loss=0.03715, over 4819.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2509, pruned_loss=0.05502, over 956264.03 frames. ], batch size: 39, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 01:00:15,118 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-27 01:00:31,887 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.587e+02 1.911e+02 2.256e+02 4.682e+02, threshold=3.822e+02, percent-clipped=4.0 +2023-03-27 01:00:43,490 INFO [finetune.py:976] (4/7) Epoch 20, batch 5650, loss[loss=0.1266, simple_loss=0.2092, pruned_loss=0.02203, over 4760.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2526, pruned_loss=0.05494, over 957375.21 frames. ], batch size: 26, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 01:00:43,564 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7458, 1.1679, 0.8870, 1.5558, 1.9189, 1.4160, 1.3347, 1.5444], + device='cuda:4'), covar=tensor([0.1441, 0.2107, 0.1923, 0.1167, 0.2076, 0.2039, 0.1458, 0.2015], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0110, 0.0091, 0.0120, 0.0093, 0.0097, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 01:00:46,982 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=114483.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:00:50,090 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.55 vs. limit=2.0 +2023-03-27 01:00:53,859 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2712, 1.7810, 2.2609, 2.1824, 2.0186, 1.9426, 2.1337, 2.0967], + device='cuda:4'), covar=tensor([0.3677, 0.3611, 0.3526, 0.3284, 0.4639, 0.3615, 0.4309, 0.3196], + device='cuda:4'), in_proj_covar=tensor([0.0254, 0.0242, 0.0262, 0.0281, 0.0279, 0.0255, 0.0290, 0.0245], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:00:56,325 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5259, 1.0705, 0.8562, 1.3462, 1.7619, 0.8699, 1.2489, 1.4000], + device='cuda:4'), covar=tensor([0.1224, 0.1819, 0.1559, 0.1018, 0.1899, 0.2024, 0.1300, 0.1719], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0091, 0.0120, 0.0093, 0.0097, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 01:00:57,509 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7780, 1.5348, 2.3800, 3.4674, 2.3089, 2.5375, 1.2090, 2.9069], + device='cuda:4'), covar=tensor([0.1698, 0.1455, 0.1199, 0.0524, 0.0791, 0.1209, 0.1890, 0.0457], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0163, 0.0100, 0.0135, 0.0123, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 01:01:13,265 INFO [finetune.py:976] (4/7) Epoch 20, batch 5700, loss[loss=0.1657, simple_loss=0.2054, pruned_loss=0.063, over 4061.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2486, pruned_loss=0.05357, over 945084.07 frames. ], batch size: 17, lr: 3.22e-03, grad_scale: 32.0 +2023-03-27 01:01:24,539 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0072, 2.6286, 2.5689, 1.2557, 2.7244, 1.9883, 0.7818, 1.8211], + device='cuda:4'), covar=tensor([0.2154, 0.2091, 0.1570, 0.3114, 0.1287, 0.1074, 0.3724, 0.1407], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0177, 0.0160, 0.0129, 0.0161, 0.0123, 0.0148, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 01:01:39,100 INFO [finetune.py:976] (4/7) Epoch 21, batch 0, loss[loss=0.1865, simple_loss=0.2593, pruned_loss=0.05688, over 4772.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2593, pruned_loss=0.05688, over 4772.00 frames. ], batch size: 28, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:01:39,100 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 01:01:46,394 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.6576, 3.2561, 3.4059, 3.5578, 3.4344, 3.3064, 3.7137, 1.5873], + device='cuda:4'), covar=tensor([0.0776, 0.0766, 0.0856, 0.0838, 0.1268, 0.1387, 0.0715, 0.4483], + device='cuda:4'), in_proj_covar=tensor([0.0346, 0.0241, 0.0277, 0.0289, 0.0330, 0.0283, 0.0301, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:01:52,342 INFO [finetune.py:1010] (4/7) Epoch 21, validation: loss=0.1598, simple_loss=0.2277, pruned_loss=0.0459, over 2265189.00 frames. +2023-03-27 01:01:52,343 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-27 01:01:56,940 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.573e+01 1.356e+02 1.658e+02 2.014e+02 3.472e+02, threshold=3.316e+02, percent-clipped=0.0 +2023-03-27 01:02:47,783 INFO [finetune.py:976] (4/7) Epoch 21, batch 50, loss[loss=0.1945, simple_loss=0.2583, pruned_loss=0.06539, over 4723.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.252, pruned_loss=0.05462, over 217766.98 frames. ], batch size: 54, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:03:21,579 INFO [finetune.py:976] (4/7) Epoch 21, batch 100, loss[loss=0.1536, simple_loss=0.215, pruned_loss=0.04611, over 4080.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2435, pruned_loss=0.05245, over 378776.36 frames. ], batch size: 66, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:03:23,377 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.865e+01 1.560e+02 1.971e+02 2.354e+02 5.080e+02, threshold=3.943e+02, percent-clipped=2.0 +2023-03-27 01:03:31,240 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5747, 1.6809, 1.3487, 1.6412, 2.0613, 1.8754, 1.6448, 1.5171], + device='cuda:4'), covar=tensor([0.0345, 0.0311, 0.0649, 0.0318, 0.0193, 0.0581, 0.0355, 0.0390], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0108, 0.0145, 0.0112, 0.0101, 0.0112, 0.0101, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.5821e-05, 8.2983e-05, 1.1420e-04, 8.5985e-05, 7.8449e-05, 8.2662e-05, + 7.4946e-05, 8.6504e-05], device='cuda:4') +2023-03-27 01:03:54,252 INFO [finetune.py:976] (4/7) Epoch 21, batch 150, loss[loss=0.1814, simple_loss=0.2488, pruned_loss=0.05702, over 4899.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2409, pruned_loss=0.05281, over 507555.11 frames. ], batch size: 36, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:04:02,529 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=114716.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:04:03,119 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7817, 1.3476, 0.6108, 1.6325, 2.1726, 1.5610, 1.6009, 1.8161], + device='cuda:4'), covar=tensor([0.1554, 0.2064, 0.2206, 0.1316, 0.2008, 0.1997, 0.1415, 0.2022], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0091, 0.0119, 0.0093, 0.0097, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 01:04:26,909 INFO [finetune.py:976] (4/7) Epoch 21, batch 200, loss[loss=0.1422, simple_loss=0.2249, pruned_loss=0.02972, over 4768.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2396, pruned_loss=0.05181, over 607505.87 frames. ], batch size: 26, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:04:29,191 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.002e+02 1.562e+02 1.886e+02 2.300e+02 5.249e+02, threshold=3.772e+02, percent-clipped=1.0 +2023-03-27 01:04:31,583 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1184, 3.6189, 3.7768, 3.9966, 3.8788, 3.6509, 4.2159, 1.3666], + device='cuda:4'), covar=tensor([0.0777, 0.0857, 0.0862, 0.0873, 0.1265, 0.1572, 0.0781, 0.5363], + device='cuda:4'), in_proj_covar=tensor([0.0346, 0.0240, 0.0276, 0.0289, 0.0328, 0.0281, 0.0301, 0.0295], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:04:40,027 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9360, 2.1495, 1.9076, 1.7946, 2.6710, 2.7226, 2.2233, 2.1880], + device='cuda:4'), covar=tensor([0.0474, 0.0336, 0.0600, 0.0389, 0.0249, 0.0479, 0.0414, 0.0354], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0108, 0.0145, 0.0112, 0.0101, 0.0112, 0.0101, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.6036e-05, 8.3214e-05, 1.1433e-04, 8.6149e-05, 7.8420e-05, 8.2886e-05, + 7.5127e-05, 8.6427e-05], device='cuda:4') +2023-03-27 01:04:42,952 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=114777.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:04:44,170 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1643, 2.0571, 1.7674, 2.1658, 2.7630, 2.3076, 2.1501, 1.6649], + device='cuda:4'), covar=tensor([0.2068, 0.2009, 0.1918, 0.1687, 0.1574, 0.1047, 0.2064, 0.1881], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0211, 0.0214, 0.0195, 0.0243, 0.0188, 0.0218, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:04:46,577 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=114783.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:05:00,771 INFO [finetune.py:976] (4/7) Epoch 21, batch 250, loss[loss=0.2303, simple_loss=0.2952, pruned_loss=0.08272, over 4796.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2435, pruned_loss=0.05301, over 684795.68 frames. ], batch size: 45, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:05:19,223 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=114831.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:05:25,343 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1880, 1.3523, 1.3605, 0.7461, 1.2888, 1.5410, 1.6082, 1.2704], + device='cuda:4'), covar=tensor([0.0815, 0.0532, 0.0509, 0.0437, 0.0459, 0.0592, 0.0311, 0.0674], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0149, 0.0125, 0.0122, 0.0129, 0.0128, 0.0141, 0.0148], + device='cuda:4'), out_proj_covar=tensor([8.9353e-05, 1.0777e-04, 8.9032e-05, 8.6324e-05, 9.0566e-05, 9.1643e-05, + 1.0146e-04, 1.0591e-04], device='cuda:4') +2023-03-27 01:05:33,143 INFO [finetune.py:976] (4/7) Epoch 21, batch 300, loss[loss=0.1604, simple_loss=0.2192, pruned_loss=0.05081, over 4698.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2468, pruned_loss=0.05396, over 744118.92 frames. ], batch size: 23, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:05:36,360 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.093e+02 1.500e+02 1.787e+02 2.137e+02 3.935e+02, threshold=3.575e+02, percent-clipped=3.0 +2023-03-27 01:05:46,367 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=114871.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:06:06,530 INFO [finetune.py:976] (4/7) Epoch 21, batch 350, loss[loss=0.16, simple_loss=0.233, pruned_loss=0.04349, over 4759.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2507, pruned_loss=0.05588, over 791906.02 frames. ], batch size: 27, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:06:20,014 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.78 vs. limit=5.0 +2023-03-27 01:06:26,469 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=114932.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:06:38,943 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5166, 1.2820, 1.9193, 3.1679, 2.1670, 2.4386, 0.6626, 2.7334], + device='cuda:4'), covar=tensor([0.2094, 0.2242, 0.1701, 0.0966, 0.0966, 0.1466, 0.2483, 0.0685], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0116, 0.0133, 0.0164, 0.0100, 0.0135, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 01:06:40,072 INFO [finetune.py:976] (4/7) Epoch 21, batch 400, loss[loss=0.152, simple_loss=0.2246, pruned_loss=0.03968, over 4893.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2493, pruned_loss=0.05394, over 828775.99 frames. ], batch size: 43, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:06:41,872 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.016e+02 1.627e+02 1.950e+02 2.383e+02 4.205e+02, threshold=3.900e+02, percent-clipped=3.0 +2023-03-27 01:07:02,951 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2927, 2.2309, 2.2424, 1.5688, 2.2573, 2.3726, 2.3907, 1.8063], + device='cuda:4'), covar=tensor([0.0584, 0.0594, 0.0658, 0.0878, 0.0574, 0.0662, 0.0576, 0.1124], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0137, 0.0140, 0.0121, 0.0125, 0.0140, 0.0141, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:07:20,652 INFO [finetune.py:976] (4/7) Epoch 21, batch 450, loss[loss=0.1605, simple_loss=0.2237, pruned_loss=0.04866, over 4745.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2491, pruned_loss=0.05364, over 856035.54 frames. ], batch size: 27, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:08:01,546 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0865, 1.7864, 2.1319, 1.5031, 1.8327, 2.1814, 1.6795, 2.3951], + device='cuda:4'), covar=tensor([0.1255, 0.1923, 0.1560, 0.1903, 0.1097, 0.1397, 0.2800, 0.0780], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0205, 0.0189, 0.0189, 0.0174, 0.0212, 0.0218, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:08:11,192 INFO [finetune.py:976] (4/7) Epoch 21, batch 500, loss[loss=0.1549, simple_loss=0.2317, pruned_loss=0.03909, over 4827.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2465, pruned_loss=0.05255, over 877304.90 frames. ], batch size: 41, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:08:11,368 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.55 vs. limit=5.0 +2023-03-27 01:08:13,017 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.154e+02 1.445e+02 1.728e+02 2.124e+02 2.919e+02, threshold=3.456e+02, percent-clipped=0.0 +2023-03-27 01:08:24,214 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=115072.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:08:33,769 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=115086.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:08:45,012 INFO [finetune.py:976] (4/7) Epoch 21, batch 550, loss[loss=0.1848, simple_loss=0.2538, pruned_loss=0.05787, over 4830.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.245, pruned_loss=0.05288, over 894838.70 frames. ], batch size: 30, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:09:14,132 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=115147.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:09:18,266 INFO [finetune.py:976] (4/7) Epoch 21, batch 600, loss[loss=0.1411, simple_loss=0.2107, pruned_loss=0.03574, over 4767.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2444, pruned_loss=0.05244, over 910317.02 frames. ], batch size: 27, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:09:20,110 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.558e+02 1.835e+02 2.263e+02 4.639e+02, threshold=3.670e+02, percent-clipped=5.0 +2023-03-27 01:09:41,562 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4506, 1.2685, 1.9917, 1.8160, 1.3749, 3.3106, 1.2020, 1.3754], + device='cuda:4'), covar=tensor([0.1178, 0.2432, 0.1419, 0.1149, 0.2077, 0.0347, 0.2052, 0.2515], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0076, 0.0092, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 01:09:51,869 INFO [finetune.py:976] (4/7) Epoch 21, batch 650, loss[loss=0.169, simple_loss=0.2471, pruned_loss=0.04543, over 4927.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2461, pruned_loss=0.05256, over 919044.53 frames. ], batch size: 38, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:10:07,424 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=115227.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:10:07,466 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9857, 1.7077, 2.3262, 1.5194, 1.9644, 2.2450, 1.5900, 2.3742], + device='cuda:4'), covar=tensor([0.1326, 0.1863, 0.1222, 0.1780, 0.0898, 0.1255, 0.2662, 0.0733], + device='cuda:4'), in_proj_covar=tensor([0.0194, 0.0206, 0.0191, 0.0190, 0.0175, 0.0213, 0.0219, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:10:25,138 INFO [finetune.py:976] (4/7) Epoch 21, batch 700, loss[loss=0.1674, simple_loss=0.2438, pruned_loss=0.04554, over 4879.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2471, pruned_loss=0.05269, over 924751.49 frames. ], batch size: 32, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:10:26,912 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.090e+02 1.664e+02 1.957e+02 2.283e+02 3.730e+02, threshold=3.914e+02, percent-clipped=1.0 +2023-03-27 01:10:58,923 INFO [finetune.py:976] (4/7) Epoch 21, batch 750, loss[loss=0.1885, simple_loss=0.2495, pruned_loss=0.06375, over 4155.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.248, pruned_loss=0.05289, over 931838.21 frames. ], batch size: 65, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:10:59,661 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8295, 1.7512, 1.4303, 1.4123, 1.8356, 1.6213, 1.8230, 1.8599], + device='cuda:4'), covar=tensor([0.1563, 0.2003, 0.3307, 0.2808, 0.2892, 0.1964, 0.3144, 0.1920], + device='cuda:4'), in_proj_covar=tensor([0.0187, 0.0189, 0.0237, 0.0255, 0.0249, 0.0205, 0.0217, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:10:59,997 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-27 01:11:31,748 INFO [finetune.py:976] (4/7) Epoch 21, batch 800, loss[loss=0.2143, simple_loss=0.2735, pruned_loss=0.07751, over 4906.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2474, pruned_loss=0.05231, over 937281.32 frames. ], batch size: 37, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:11:33,561 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.112e+02 1.426e+02 1.720e+02 2.044e+02 3.360e+02, threshold=3.441e+02, percent-clipped=0.0 +2023-03-27 01:11:41,495 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=115370.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:11:42,134 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=115371.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:11:42,700 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=115372.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:12:04,584 INFO [finetune.py:976] (4/7) Epoch 21, batch 850, loss[loss=0.1798, simple_loss=0.2435, pruned_loss=0.05803, over 4860.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2447, pruned_loss=0.05125, over 940288.92 frames. ], batch size: 31, lr: 3.21e-03, grad_scale: 32.0 +2023-03-27 01:12:14,866 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=115420.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:12:16,192 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.42 vs. limit=5.0 +2023-03-27 01:12:24,429 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=115431.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:12:25,047 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=115432.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:12:32,194 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8044, 1.7346, 1.6120, 1.7584, 1.7370, 4.4531, 1.8163, 2.0960], + device='cuda:4'), covar=tensor([0.3961, 0.2976, 0.2323, 0.2714, 0.1372, 0.0177, 0.2299, 0.1110], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0116, 0.0120, 0.0123, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 01:12:40,136 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8027, 1.6216, 1.8659, 1.2896, 1.7065, 1.8765, 1.5982, 2.0575], + device='cuda:4'), covar=tensor([0.1110, 0.2107, 0.1295, 0.1554, 0.1025, 0.1209, 0.3068, 0.0774], + device='cuda:4'), in_proj_covar=tensor([0.0194, 0.0206, 0.0192, 0.0191, 0.0176, 0.0214, 0.0220, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:12:41,981 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=115442.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:12:54,666 INFO [finetune.py:976] (4/7) Epoch 21, batch 900, loss[loss=0.1381, simple_loss=0.2183, pruned_loss=0.02888, over 4921.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.242, pruned_loss=0.05023, over 944085.17 frames. ], batch size: 43, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:13:00,781 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.494e+02 1.776e+02 2.140e+02 4.219e+02, threshold=3.551e+02, percent-clipped=3.0 +2023-03-27 01:13:37,347 INFO [finetune.py:976] (4/7) Epoch 21, batch 950, loss[loss=0.1702, simple_loss=0.2493, pruned_loss=0.04557, over 4817.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2427, pruned_loss=0.05074, over 947821.08 frames. ], batch size: 45, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:13:40,290 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-27 01:13:40,528 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6211, 1.6540, 1.5121, 0.8669, 1.6430, 1.8775, 1.8522, 1.4037], + device='cuda:4'), covar=tensor([0.0984, 0.0527, 0.0507, 0.0536, 0.0404, 0.0526, 0.0278, 0.0718], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0149, 0.0125, 0.0123, 0.0129, 0.0128, 0.0141, 0.0147], + device='cuda:4'), out_proj_covar=tensor([8.9448e-05, 1.0759e-04, 8.9529e-05, 8.6818e-05, 9.0941e-05, 9.1520e-05, + 1.0143e-04, 1.0584e-04], device='cuda:4') +2023-03-27 01:13:51,955 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=115527.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:13:53,175 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6630, 1.5982, 2.1445, 3.3104, 2.2489, 2.3645, 0.9780, 2.8105], + device='cuda:4'), covar=tensor([0.1757, 0.1333, 0.1268, 0.0509, 0.0765, 0.1252, 0.1853, 0.0423], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0163, 0.0100, 0.0135, 0.0124, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 01:14:11,292 INFO [finetune.py:976] (4/7) Epoch 21, batch 1000, loss[loss=0.2289, simple_loss=0.295, pruned_loss=0.0814, over 4746.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2461, pruned_loss=0.05253, over 945938.28 frames. ], batch size: 59, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:14:13,112 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.049e+02 1.550e+02 1.849e+02 2.159e+02 3.452e+02, threshold=3.698e+02, percent-clipped=0.0 +2023-03-27 01:14:24,464 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=115575.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:14:25,767 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3901, 2.3024, 2.0997, 2.4932, 2.9343, 2.5531, 2.3298, 1.8266], + device='cuda:4'), covar=tensor([0.2248, 0.1914, 0.1865, 0.1575, 0.1649, 0.1035, 0.2000, 0.1967], + device='cuda:4'), in_proj_covar=tensor([0.0244, 0.0211, 0.0213, 0.0194, 0.0243, 0.0188, 0.0218, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:14:44,024 INFO [finetune.py:976] (4/7) Epoch 21, batch 1050, loss[loss=0.1845, simple_loss=0.2605, pruned_loss=0.05428, over 4759.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2481, pruned_loss=0.05258, over 949364.07 frames. ], batch size: 54, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:14:52,439 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5462, 3.4570, 3.1694, 1.4933, 3.5177, 2.7569, 0.8388, 2.2466], + device='cuda:4'), covar=tensor([0.2212, 0.2207, 0.1618, 0.3553, 0.1094, 0.0988, 0.4302, 0.1704], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0178, 0.0159, 0.0130, 0.0161, 0.0123, 0.0148, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 01:15:04,582 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.8304, 4.1517, 4.3624, 4.6284, 4.5294, 4.2998, 4.9247, 1.5613], + device='cuda:4'), covar=tensor([0.0653, 0.0774, 0.0717, 0.0925, 0.1134, 0.1644, 0.0488, 0.5678], + device='cuda:4'), in_proj_covar=tensor([0.0344, 0.0237, 0.0275, 0.0288, 0.0326, 0.0279, 0.0299, 0.0294], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:15:16,670 INFO [finetune.py:976] (4/7) Epoch 21, batch 1100, loss[loss=0.2313, simple_loss=0.2988, pruned_loss=0.08187, over 4833.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2487, pruned_loss=0.05275, over 949436.56 frames. ], batch size: 47, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:15:19,447 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.043e+02 1.582e+02 1.822e+02 2.328e+02 4.675e+02, threshold=3.643e+02, percent-clipped=4.0 +2023-03-27 01:15:28,825 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.72 vs. limit=5.0 +2023-03-27 01:15:31,238 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-27 01:15:50,438 INFO [finetune.py:976] (4/7) Epoch 21, batch 1150, loss[loss=0.131, simple_loss=0.2073, pruned_loss=0.02739, over 4792.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.249, pruned_loss=0.0523, over 951570.34 frames. ], batch size: 26, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:16:05,311 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=115726.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:16:05,413 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-03-27 01:16:05,921 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=115727.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:16:14,927 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=115742.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:16:24,029 INFO [finetune.py:976] (4/7) Epoch 21, batch 1200, loss[loss=0.1887, simple_loss=0.2538, pruned_loss=0.06178, over 4811.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2471, pruned_loss=0.05183, over 953293.19 frames. ], batch size: 45, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:16:25,823 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.936e+01 1.465e+02 1.737e+02 2.048e+02 4.574e+02, threshold=3.475e+02, percent-clipped=2.0 +2023-03-27 01:16:35,987 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6771, 1.5584, 1.8800, 1.2326, 1.6666, 1.8202, 1.4487, 2.0639], + device='cuda:4'), covar=tensor([0.1221, 0.2115, 0.1290, 0.1789, 0.0943, 0.1391, 0.3105, 0.0869], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0203, 0.0190, 0.0189, 0.0174, 0.0212, 0.0217, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:16:47,357 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=115790.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:16:56,824 INFO [finetune.py:976] (4/7) Epoch 21, batch 1250, loss[loss=0.1893, simple_loss=0.2646, pruned_loss=0.057, over 4823.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2458, pruned_loss=0.05133, over 954363.70 frames. ], batch size: 39, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:17:29,533 INFO [finetune.py:976] (4/7) Epoch 21, batch 1300, loss[loss=0.1447, simple_loss=0.2237, pruned_loss=0.03283, over 4755.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2433, pruned_loss=0.0508, over 952985.13 frames. ], batch size: 28, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:17:32,371 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.563e+02 1.759e+02 2.181e+02 4.124e+02, threshold=3.519e+02, percent-clipped=1.0 +2023-03-27 01:18:05,332 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-27 01:18:12,434 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=115892.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 01:18:13,069 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0520, 0.9808, 0.8807, 1.1433, 1.1660, 1.1479, 1.0029, 0.9480], + device='cuda:4'), covar=tensor([0.0381, 0.0335, 0.0741, 0.0355, 0.0319, 0.0437, 0.0359, 0.0438], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0108, 0.0145, 0.0113, 0.0100, 0.0111, 0.0101, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.5667e-05, 8.2964e-05, 1.1414e-04, 8.6255e-05, 7.8024e-05, 8.2152e-05, + 7.4896e-05, 8.6183e-05], device='cuda:4') +2023-03-27 01:18:23,975 INFO [finetune.py:976] (4/7) Epoch 21, batch 1350, loss[loss=0.2063, simple_loss=0.2814, pruned_loss=0.0656, over 4802.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2434, pruned_loss=0.05121, over 953047.82 frames. ], batch size: 45, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:19:00,568 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=115953.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 01:19:01,043 INFO [finetune.py:976] (4/7) Epoch 21, batch 1400, loss[loss=0.2537, simple_loss=0.3209, pruned_loss=0.09321, over 4270.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.246, pruned_loss=0.05214, over 949437.64 frames. ], batch size: 65, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:19:02,751 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-27 01:19:02,859 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.541e+02 1.818e+02 2.071e+02 3.575e+02, threshold=3.635e+02, percent-clipped=1.0 +2023-03-27 01:19:09,186 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9377, 4.6417, 4.4386, 2.1121, 4.8061, 3.6767, 1.0055, 3.1995], + device='cuda:4'), covar=tensor([0.2209, 0.1593, 0.1252, 0.3239, 0.0685, 0.0792, 0.4649, 0.1316], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0176, 0.0158, 0.0128, 0.0159, 0.0122, 0.0146, 0.0122], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 01:19:35,504 INFO [finetune.py:976] (4/7) Epoch 21, batch 1450, loss[loss=0.1751, simple_loss=0.2449, pruned_loss=0.05264, over 4809.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2487, pruned_loss=0.05303, over 951226.92 frames. ], batch size: 38, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:19:42,999 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116014.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:19:51,744 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116026.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:19:52,348 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116027.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:20:03,089 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8568, 1.5872, 2.3185, 1.5410, 1.9728, 2.0592, 1.4550, 2.1356], + device='cuda:4'), covar=tensor([0.1227, 0.1679, 0.1023, 0.1536, 0.0884, 0.1268, 0.2397, 0.0857], + device='cuda:4'), in_proj_covar=tensor([0.0190, 0.0202, 0.0189, 0.0188, 0.0173, 0.0211, 0.0216, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:20:09,068 INFO [finetune.py:976] (4/7) Epoch 21, batch 1500, loss[loss=0.1985, simple_loss=0.2694, pruned_loss=0.06381, over 4812.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2491, pruned_loss=0.05314, over 951659.75 frames. ], batch size: 40, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:20:10,888 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.231e+02 1.719e+02 2.058e+02 2.312e+02 4.180e+02, threshold=4.116e+02, percent-clipped=2.0 +2023-03-27 01:20:23,169 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116074.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:20:23,792 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116075.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:20:23,859 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116075.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 01:20:42,654 INFO [finetune.py:976] (4/7) Epoch 21, batch 1550, loss[loss=0.1651, simple_loss=0.2388, pruned_loss=0.04571, over 4827.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2491, pruned_loss=0.05298, over 953284.32 frames. ], batch size: 30, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:20:43,385 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116105.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:20:50,869 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116116.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:21:15,951 INFO [finetune.py:976] (4/7) Epoch 21, batch 1600, loss[loss=0.1895, simple_loss=0.2546, pruned_loss=0.0622, over 4058.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.248, pruned_loss=0.05311, over 953986.73 frames. ], batch size: 65, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:21:17,775 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.061e+01 1.555e+02 1.841e+02 2.223e+02 4.654e+02, threshold=3.683e+02, percent-clipped=1.0 +2023-03-27 01:21:23,295 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2019, 2.9545, 2.8427, 1.3375, 3.0505, 2.2710, 0.6331, 2.0245], + device='cuda:4'), covar=tensor([0.2347, 0.2400, 0.1904, 0.3273, 0.1403, 0.1148, 0.4315, 0.1605], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0179, 0.0160, 0.0130, 0.0162, 0.0124, 0.0149, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 01:21:23,362 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116166.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:21:31,969 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116177.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:21:49,853 INFO [finetune.py:976] (4/7) Epoch 21, batch 1650, loss[loss=0.1959, simple_loss=0.2602, pruned_loss=0.06586, over 4744.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2442, pruned_loss=0.05163, over 955641.13 frames. ], batch size: 54, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:22:05,161 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4128, 1.3512, 1.5660, 2.4004, 1.6380, 2.1380, 0.8291, 2.0688], + device='cuda:4'), covar=tensor([0.1645, 0.1440, 0.1106, 0.0769, 0.0876, 0.1291, 0.1551, 0.0588], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0116, 0.0133, 0.0163, 0.0100, 0.0136, 0.0124, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 01:22:19,470 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116248.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 01:22:23,481 INFO [finetune.py:976] (4/7) Epoch 21, batch 1700, loss[loss=0.1741, simple_loss=0.2466, pruned_loss=0.0508, over 4864.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2422, pruned_loss=0.05089, over 958106.81 frames. ], batch size: 31, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:22:25,324 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.524e+01 1.502e+02 1.774e+02 2.116e+02 3.203e+02, threshold=3.548e+02, percent-clipped=0.0 +2023-03-27 01:22:44,334 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5960, 3.4916, 3.3345, 1.5123, 3.6096, 2.6594, 0.8420, 2.4009], + device='cuda:4'), covar=tensor([0.2359, 0.1908, 0.1734, 0.3572, 0.1045, 0.1128, 0.4559, 0.1543], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0178, 0.0159, 0.0129, 0.0162, 0.0123, 0.0148, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 01:22:47,288 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5401, 2.3901, 1.9232, 0.9519, 2.0779, 1.9480, 1.8076, 2.0949], + device='cuda:4'), covar=tensor([0.0821, 0.0683, 0.1483, 0.1892, 0.1205, 0.2105, 0.2116, 0.0924], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0192, 0.0198, 0.0183, 0.0210, 0.0209, 0.0223, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:22:48,536 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1429, 2.0596, 1.7218, 1.8497, 2.1082, 1.8533, 2.3027, 2.1051], + device='cuda:4'), covar=tensor([0.1315, 0.2023, 0.2980, 0.2413, 0.2621, 0.1728, 0.2541, 0.1843], + device='cuda:4'), in_proj_covar=tensor([0.0187, 0.0188, 0.0236, 0.0254, 0.0249, 0.0204, 0.0215, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:22:58,162 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6400, 1.8567, 1.4457, 1.5757, 2.2071, 2.1206, 1.8886, 1.7515], + device='cuda:4'), covar=tensor([0.0444, 0.0363, 0.0644, 0.0358, 0.0319, 0.0546, 0.0359, 0.0415], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0108, 0.0145, 0.0113, 0.0100, 0.0111, 0.0100, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.5829e-05, 8.2835e-05, 1.1412e-04, 8.6280e-05, 7.8102e-05, 8.2133e-05, + 7.4734e-05, 8.6635e-05], device='cuda:4') +2023-03-27 01:22:59,247 INFO [finetune.py:976] (4/7) Epoch 21, batch 1750, loss[loss=0.2239, simple_loss=0.288, pruned_loss=0.07987, over 4908.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2434, pruned_loss=0.05152, over 957084.14 frames. ], batch size: 36, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:23:03,992 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7747, 1.7209, 2.0787, 1.3397, 1.8392, 2.0106, 1.4820, 2.2176], + device='cuda:4'), covar=tensor([0.1328, 0.1881, 0.1341, 0.1931, 0.1038, 0.1482, 0.2894, 0.0852], + device='cuda:4'), in_proj_covar=tensor([0.0190, 0.0202, 0.0189, 0.0188, 0.0172, 0.0211, 0.0216, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:23:15,877 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6623, 2.4970, 1.9821, 2.5832, 2.3997, 2.2060, 2.9363, 2.6486], + device='cuda:4'), covar=tensor([0.1209, 0.1973, 0.2803, 0.2583, 0.2439, 0.1532, 0.2899, 0.1703], + device='cuda:4'), in_proj_covar=tensor([0.0187, 0.0188, 0.0236, 0.0255, 0.0249, 0.0204, 0.0215, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:23:19,662 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116323.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:23:58,747 INFO [finetune.py:976] (4/7) Epoch 21, batch 1800, loss[loss=0.1576, simple_loss=0.2411, pruned_loss=0.03704, over 4903.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2477, pruned_loss=0.05249, over 957974.03 frames. ], batch size: 36, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:24:00,589 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.248e+02 1.627e+02 1.938e+02 2.363e+02 5.057e+02, threshold=3.876e+02, percent-clipped=3.0 +2023-03-27 01:24:03,081 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1468, 1.4018, 0.8608, 1.8967, 2.4359, 1.8181, 1.6577, 1.7489], + device='cuda:4'), covar=tensor([0.1372, 0.2143, 0.2037, 0.1187, 0.1786, 0.1953, 0.1503, 0.2029], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0110, 0.0092, 0.0119, 0.0093, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 01:24:08,542 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116370.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 01:24:18,598 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116384.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:24:22,076 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5346, 1.2609, 1.8623, 1.9438, 1.5203, 3.3652, 1.1130, 1.4136], + device='cuda:4'), covar=tensor([0.1141, 0.2528, 0.1271, 0.1033, 0.2055, 0.0281, 0.2350, 0.2530], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0076, 0.0092, 0.0080, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 01:24:31,790 INFO [finetune.py:976] (4/7) Epoch 21, batch 1850, loss[loss=0.1577, simple_loss=0.2202, pruned_loss=0.04762, over 4761.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2494, pruned_loss=0.05335, over 958235.60 frames. ], batch size: 23, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:25:00,259 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116446.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:25:05,406 INFO [finetune.py:976] (4/7) Epoch 21, batch 1900, loss[loss=0.1317, simple_loss=0.1984, pruned_loss=0.03247, over 4722.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.25, pruned_loss=0.05387, over 954942.31 frames. ], batch size: 23, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:25:07,231 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.037e+02 1.572e+02 1.880e+02 2.123e+02 3.861e+02, threshold=3.760e+02, percent-clipped=0.0 +2023-03-27 01:25:07,385 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1139, 2.2440, 2.0300, 2.5690, 2.6570, 2.4607, 2.1779, 1.6884], + device='cuda:4'), covar=tensor([0.2216, 0.1745, 0.1743, 0.1363, 0.1768, 0.0987, 0.1932, 0.1868], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0210, 0.0213, 0.0195, 0.0243, 0.0189, 0.0218, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:25:10,209 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116461.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:25:16,989 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116472.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:25:38,740 INFO [finetune.py:976] (4/7) Epoch 21, batch 1950, loss[loss=0.133, simple_loss=0.202, pruned_loss=0.03205, over 4475.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2486, pruned_loss=0.05305, over 955434.20 frames. ], batch size: 19, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:25:39,478 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116505.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:25:40,680 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116507.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 01:25:47,814 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.4555, 4.7378, 5.0764, 5.2516, 5.2174, 5.0986, 5.6202, 1.7172], + device='cuda:4'), covar=tensor([0.0772, 0.0766, 0.0750, 0.1134, 0.1190, 0.1307, 0.0478, 0.6049], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0238, 0.0275, 0.0287, 0.0328, 0.0282, 0.0299, 0.0295], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:26:07,962 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116548.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 01:26:11,462 INFO [finetune.py:976] (4/7) Epoch 21, batch 2000, loss[loss=0.1804, simple_loss=0.2475, pruned_loss=0.05668, over 4834.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2468, pruned_loss=0.05278, over 956092.10 frames. ], batch size: 38, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:26:13,788 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 7.123e+01 1.433e+02 1.690e+02 2.067e+02 3.885e+02, threshold=3.380e+02, percent-clipped=2.0 +2023-03-27 01:26:13,924 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6426, 1.6999, 1.4693, 1.6279, 2.0744, 1.8880, 1.7631, 1.5456], + device='cuda:4'), covar=tensor([0.0323, 0.0288, 0.0650, 0.0312, 0.0216, 0.0442, 0.0283, 0.0372], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0108, 0.0146, 0.0113, 0.0101, 0.0112, 0.0101, 0.0114], + device='cuda:4'), out_proj_covar=tensor([7.6322e-05, 8.3184e-05, 1.1485e-04, 8.6913e-05, 7.8263e-05, 8.2740e-05, + 7.4914e-05, 8.7053e-05], device='cuda:4') +2023-03-27 01:26:19,738 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116566.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:26:30,072 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4004, 2.2785, 2.0550, 1.3287, 2.2026, 1.9837, 1.8090, 2.1140], + device='cuda:4'), covar=tensor([0.0978, 0.0652, 0.1388, 0.1671, 0.1166, 0.1689, 0.1935, 0.0885], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0191, 0.0197, 0.0181, 0.0208, 0.0207, 0.0222, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:26:39,353 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116596.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 01:26:44,680 INFO [finetune.py:976] (4/7) Epoch 21, batch 2050, loss[loss=0.1655, simple_loss=0.2289, pruned_loss=0.05101, over 4766.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.244, pruned_loss=0.05161, over 956827.43 frames. ], batch size: 27, lr: 3.20e-03, grad_scale: 64.0 +2023-03-27 01:26:54,221 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116618.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:27:18,449 INFO [finetune.py:976] (4/7) Epoch 21, batch 2100, loss[loss=0.2545, simple_loss=0.3088, pruned_loss=0.1001, over 4154.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2437, pruned_loss=0.05185, over 955150.61 frames. ], batch size: 65, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:27:20,847 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.617e+02 1.783e+02 2.198e+02 6.495e+02, threshold=3.567e+02, percent-clipped=4.0 +2023-03-27 01:27:29,071 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116670.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 01:27:34,426 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116679.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:27:34,489 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116679.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:27:39,310 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.2732, 2.8714, 3.0280, 3.2162, 3.0310, 2.8515, 3.3250, 1.0295], + device='cuda:4'), covar=tensor([0.1239, 0.1055, 0.1114, 0.1319, 0.1746, 0.1786, 0.1114, 0.5840], + device='cuda:4'), in_proj_covar=tensor([0.0348, 0.0240, 0.0278, 0.0289, 0.0331, 0.0284, 0.0301, 0.0297], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:27:51,925 INFO [finetune.py:976] (4/7) Epoch 21, batch 2150, loss[loss=0.1741, simple_loss=0.2525, pruned_loss=0.04784, over 4728.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.247, pruned_loss=0.05262, over 955677.46 frames. ], batch size: 59, lr: 3.20e-03, grad_scale: 32.0 +2023-03-27 01:27:52,033 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116704.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:28:00,963 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116718.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:28:04,565 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1449, 1.2940, 1.3508, 0.7029, 1.2625, 1.5550, 1.5163, 1.2665], + device='cuda:4'), covar=tensor([0.0968, 0.0659, 0.0576, 0.0552, 0.0512, 0.0689, 0.0354, 0.0705], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0150, 0.0126, 0.0124, 0.0130, 0.0129, 0.0141, 0.0148], + device='cuda:4'), out_proj_covar=tensor([8.9938e-05, 1.0833e-04, 8.9759e-05, 8.7696e-05, 9.1623e-05, 9.1928e-05, + 1.0147e-04, 1.0603e-04], device='cuda:4') +2023-03-27 01:28:26,697 INFO [finetune.py:976] (4/7) Epoch 21, batch 2200, loss[loss=0.181, simple_loss=0.256, pruned_loss=0.05301, over 4830.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2485, pruned_loss=0.05253, over 957100.03 frames. ], batch size: 39, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:28:27,467 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1443, 2.0443, 1.6721, 1.8047, 2.0684, 1.8132, 2.2731, 2.0877], + device='cuda:4'), covar=tensor([0.1270, 0.1784, 0.2968, 0.2355, 0.2493, 0.1662, 0.2700, 0.1764], + device='cuda:4'), in_proj_covar=tensor([0.0187, 0.0189, 0.0237, 0.0255, 0.0250, 0.0205, 0.0216, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:28:30,714 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.107e+02 1.636e+02 2.054e+02 2.505e+02 6.138e+02, threshold=4.108e+02, percent-clipped=5.0 +2023-03-27 01:28:32,666 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116761.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:28:39,666 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116765.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:28:44,494 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116772.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:29:19,243 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116802.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 01:29:19,293 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=116802.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:29:20,384 INFO [finetune.py:976] (4/7) Epoch 21, batch 2250, loss[loss=0.1682, simple_loss=0.2456, pruned_loss=0.04537, over 4836.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2485, pruned_loss=0.05239, over 958021.16 frames. ], batch size: 30, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:29:28,969 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116809.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:29:39,990 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=116820.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:29:54,836 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.87 vs. limit=5.0 +2023-03-27 01:30:01,468 INFO [finetune.py:976] (4/7) Epoch 21, batch 2300, loss[loss=0.1688, simple_loss=0.2471, pruned_loss=0.04525, over 4819.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2488, pruned_loss=0.05233, over 958638.72 frames. ], batch size: 40, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:30:04,887 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.160e+02 1.477e+02 1.740e+02 2.172e+02 4.454e+02, threshold=3.479e+02, percent-clipped=1.0 +2023-03-27 01:30:06,832 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116861.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:30:08,089 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=116863.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:30:27,048 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6888, 2.4882, 1.9802, 2.8846, 2.6787, 2.4059, 3.2065, 2.7745], + device='cuda:4'), covar=tensor([0.1276, 0.2324, 0.3210, 0.2574, 0.2774, 0.1719, 0.2614, 0.1758], + device='cuda:4'), in_proj_covar=tensor([0.0186, 0.0188, 0.0235, 0.0253, 0.0247, 0.0203, 0.0215, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:30:34,064 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-27 01:30:34,136 INFO [finetune.py:976] (4/7) Epoch 21, batch 2350, loss[loss=0.2371, simple_loss=0.2878, pruned_loss=0.09323, over 4192.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2473, pruned_loss=0.05255, over 956452.06 frames. ], batch size: 65, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:30:49,680 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-27 01:30:59,671 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2449, 2.3229, 2.3245, 2.5885, 2.6855, 2.5034, 2.4278, 1.8194], + device='cuda:4'), covar=tensor([0.2064, 0.1729, 0.1568, 0.1467, 0.1777, 0.0995, 0.1872, 0.1795], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0210, 0.0214, 0.0195, 0.0244, 0.0189, 0.0219, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:31:05,083 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7727, 1.3164, 0.7562, 1.6442, 2.2134, 1.4899, 1.5782, 1.5898], + device='cuda:4'), covar=tensor([0.1571, 0.2246, 0.2123, 0.1285, 0.2045, 0.1927, 0.1600, 0.2034], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0110, 0.0092, 0.0120, 0.0094, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 01:31:07,400 INFO [finetune.py:976] (4/7) Epoch 21, batch 2400, loss[loss=0.1939, simple_loss=0.2598, pruned_loss=0.06395, over 4875.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2454, pruned_loss=0.05247, over 957118.70 frames. ], batch size: 31, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:31:09,766 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.566e+02 1.863e+02 2.219e+02 3.648e+02, threshold=3.726e+02, percent-clipped=1.0 +2023-03-27 01:31:21,638 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=116974.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:31:25,171 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=116979.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:31:40,997 INFO [finetune.py:976] (4/7) Epoch 21, batch 2450, loss[loss=0.1345, simple_loss=0.2131, pruned_loss=0.02799, over 4906.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2425, pruned_loss=0.05164, over 955314.26 frames. ], batch size: 37, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:31:50,977 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7939, 1.7634, 1.5427, 1.9935, 2.3428, 1.9679, 1.8508, 1.4668], + device='cuda:4'), covar=tensor([0.2287, 0.1967, 0.1872, 0.1545, 0.1736, 0.1193, 0.2245, 0.1931], + device='cuda:4'), in_proj_covar=tensor([0.0245, 0.0210, 0.0213, 0.0194, 0.0243, 0.0188, 0.0218, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:31:51,745 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-27 01:31:57,424 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117027.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:31:59,968 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-27 01:32:14,704 INFO [finetune.py:976] (4/7) Epoch 21, batch 2500, loss[loss=0.2231, simple_loss=0.3014, pruned_loss=0.07238, over 4889.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2444, pruned_loss=0.05263, over 955049.58 frames. ], batch size: 32, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:32:17,113 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.025e+02 1.510e+02 1.797e+02 2.340e+02 3.968e+02, threshold=3.593e+02, percent-clipped=1.0 +2023-03-27 01:32:18,390 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=117060.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:32:31,932 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0307, 1.8630, 1.8667, 0.8803, 2.1860, 2.4332, 2.0397, 1.7818], + device='cuda:4'), covar=tensor([0.0972, 0.0808, 0.0560, 0.0706, 0.0495, 0.0678, 0.0669, 0.0780], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0148, 0.0124, 0.0123, 0.0129, 0.0127, 0.0140, 0.0147], + device='cuda:4'), out_proj_covar=tensor([8.9202e-05, 1.0734e-04, 8.8894e-05, 8.6925e-05, 9.0724e-05, 9.0869e-05, + 1.0049e-04, 1.0520e-04], device='cuda:4') +2023-03-27 01:32:46,802 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117102.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:32:47,903 INFO [finetune.py:976] (4/7) Epoch 21, batch 2550, loss[loss=0.2178, simple_loss=0.2893, pruned_loss=0.07313, over 4811.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2476, pruned_loss=0.05353, over 954229.75 frames. ], batch size: 38, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:33:03,671 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-27 01:33:19,399 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117150.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:33:21,776 INFO [finetune.py:976] (4/7) Epoch 21, batch 2600, loss[loss=0.1667, simple_loss=0.238, pruned_loss=0.04765, over 4819.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2488, pruned_loss=0.05368, over 951591.44 frames. ], batch size: 39, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:33:24,217 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.102e+02 1.533e+02 1.830e+02 2.226e+02 4.351e+02, threshold=3.661e+02, percent-clipped=3.0 +2023-03-27 01:33:24,315 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=117158.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:33:26,117 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117161.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:34:06,272 INFO [finetune.py:976] (4/7) Epoch 21, batch 2650, loss[loss=0.162, simple_loss=0.2385, pruned_loss=0.04277, over 4747.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2493, pruned_loss=0.0536, over 950350.85 frames. ], batch size: 54, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:34:09,392 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117209.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:34:17,632 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0609, 0.9956, 0.9555, 0.4630, 0.9122, 1.1766, 1.2045, 1.0002], + device='cuda:4'), covar=tensor([0.0977, 0.0665, 0.0628, 0.0576, 0.0599, 0.0689, 0.0458, 0.0823], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0150, 0.0126, 0.0124, 0.0130, 0.0129, 0.0142, 0.0148], + device='cuda:4'), out_proj_covar=tensor([9.0227e-05, 1.0866e-04, 8.9926e-05, 8.7658e-05, 9.1756e-05, 9.1815e-05, + 1.0181e-04, 1.0633e-04], device='cuda:4') +2023-03-27 01:34:26,119 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=117221.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:34:45,201 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9340, 1.5249, 2.0244, 1.9618, 1.7727, 1.7536, 1.8885, 1.8885], + device='cuda:4'), covar=tensor([0.3712, 0.3922, 0.3239, 0.3729, 0.4809, 0.3541, 0.4306, 0.3023], + device='cuda:4'), in_proj_covar=tensor([0.0254, 0.0242, 0.0263, 0.0282, 0.0280, 0.0256, 0.0290, 0.0244], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:35:03,768 INFO [finetune.py:976] (4/7) Epoch 21, batch 2700, loss[loss=0.1551, simple_loss=0.2379, pruned_loss=0.03614, over 4785.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2488, pruned_loss=0.05286, over 952417.73 frames. ], batch size: 29, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:35:06,200 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.072e+02 1.522e+02 1.732e+02 2.127e+02 4.053e+02, threshold=3.464e+02, percent-clipped=3.0 +2023-03-27 01:35:23,782 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117274.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:35:30,657 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=117282.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:35:41,245 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0753, 1.7835, 2.3767, 1.5108, 2.0489, 2.2760, 1.6473, 2.4679], + device='cuda:4'), covar=tensor([0.1311, 0.2058, 0.1376, 0.2035, 0.1061, 0.1371, 0.2986, 0.0905], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0204, 0.0191, 0.0190, 0.0174, 0.0213, 0.0219, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:35:45,278 INFO [finetune.py:976] (4/7) Epoch 21, batch 2750, loss[loss=0.1774, simple_loss=0.2444, pruned_loss=0.0552, over 4786.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2468, pruned_loss=0.05244, over 950858.29 frames. ], batch size: 29, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:35:52,640 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.6669, 4.8827, 5.2260, 5.5350, 5.3654, 5.0779, 5.7067, 2.5300], + device='cuda:4'), covar=tensor([0.0551, 0.0806, 0.0662, 0.0602, 0.0993, 0.1289, 0.0423, 0.4694], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0240, 0.0277, 0.0289, 0.0330, 0.0281, 0.0301, 0.0296], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:35:56,250 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117322.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:36:01,565 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=117329.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:36:01,608 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2904, 2.1475, 1.7406, 2.0048, 2.0115, 1.9549, 2.0418, 2.7393], + device='cuda:4'), covar=tensor([0.3531, 0.4026, 0.3269, 0.3717, 0.3868, 0.2414, 0.3603, 0.1677], + device='cuda:4'), in_proj_covar=tensor([0.0287, 0.0262, 0.0233, 0.0277, 0.0253, 0.0223, 0.0252, 0.0234], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:36:18,643 INFO [finetune.py:976] (4/7) Epoch 21, batch 2800, loss[loss=0.1172, simple_loss=0.1934, pruned_loss=0.02054, over 4880.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.243, pruned_loss=0.05097, over 949477.50 frames. ], batch size: 34, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:36:21,558 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.020e+02 1.495e+02 1.752e+02 2.115e+02 2.888e+02, threshold=3.503e+02, percent-clipped=0.0 +2023-03-27 01:36:22,915 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117360.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:36:42,956 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=117390.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:36:52,478 INFO [finetune.py:976] (4/7) Epoch 21, batch 2850, loss[loss=0.1792, simple_loss=0.2531, pruned_loss=0.05267, over 4917.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2419, pruned_loss=0.05114, over 949633.69 frames. ], batch size: 37, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:36:54,972 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117408.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:37:25,543 INFO [finetune.py:976] (4/7) Epoch 21, batch 2900, loss[loss=0.2016, simple_loss=0.2739, pruned_loss=0.06467, over 4814.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2454, pruned_loss=0.05228, over 950707.68 frames. ], batch size: 40, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:37:28,394 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.016e+02 1.554e+02 1.875e+02 2.295e+02 6.888e+02, threshold=3.749e+02, percent-clipped=2.0 +2023-03-27 01:37:28,519 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117458.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:37:49,798 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4447, 2.3743, 1.9952, 2.4175, 2.2668, 2.2629, 2.2032, 3.2088], + device='cuda:4'), covar=tensor([0.3353, 0.4741, 0.3226, 0.3842, 0.4037, 0.2361, 0.4417, 0.1469], + device='cuda:4'), in_proj_covar=tensor([0.0287, 0.0262, 0.0232, 0.0276, 0.0252, 0.0223, 0.0252, 0.0233], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:37:50,363 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.5339, 1.6942, 1.6677, 0.9232, 1.8631, 2.0246, 1.9507, 1.4838], + device='cuda:4'), covar=tensor([0.0933, 0.0601, 0.0519, 0.0524, 0.0486, 0.0519, 0.0340, 0.0706], + device='cuda:4'), in_proj_covar=tensor([0.0124, 0.0151, 0.0126, 0.0125, 0.0131, 0.0129, 0.0142, 0.0149], + device='cuda:4'), out_proj_covar=tensor([9.0456e-05, 1.0891e-04, 9.0122e-05, 8.7978e-05, 9.1860e-05, 9.1899e-05, + 1.0174e-04, 1.0659e-04], device='cuda:4') +2023-03-27 01:37:59,201 INFO [finetune.py:976] (4/7) Epoch 21, batch 2950, loss[loss=0.1688, simple_loss=0.2465, pruned_loss=0.04554, over 4929.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2484, pruned_loss=0.05282, over 950876.86 frames. ], batch size: 33, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:38:00,493 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117506.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:38:32,248 INFO [finetune.py:976] (4/7) Epoch 21, batch 3000, loss[loss=0.1837, simple_loss=0.2592, pruned_loss=0.05412, over 4884.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.25, pruned_loss=0.05363, over 952684.62 frames. ], batch size: 32, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:38:32,249 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 01:38:38,557 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2517, 2.0718, 1.7903, 2.0268, 2.2101, 1.9533, 2.3239, 2.2233], + device='cuda:4'), covar=tensor([0.1422, 0.2230, 0.3127, 0.2340, 0.2631, 0.1734, 0.2946, 0.1910], + device='cuda:4'), in_proj_covar=tensor([0.0186, 0.0187, 0.0234, 0.0252, 0.0246, 0.0203, 0.0214, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:38:42,800 INFO [finetune.py:1010] (4/7) Epoch 21, validation: loss=0.1567, simple_loss=0.2253, pruned_loss=0.04408, over 2265189.00 frames. +2023-03-27 01:38:42,801 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-27 01:38:45,672 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.002e+02 1.534e+02 1.924e+02 2.362e+02 3.621e+02, threshold=3.849e+02, percent-clipped=0.0 +2023-03-27 01:39:00,139 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=117577.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:39:11,709 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1519, 1.5435, 1.0269, 1.9373, 2.4185, 1.6604, 1.8789, 1.9041], + device='cuda:4'), covar=tensor([0.1141, 0.1841, 0.1697, 0.1039, 0.1694, 0.1712, 0.1237, 0.1692], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0110, 0.0092, 0.0119, 0.0093, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 01:39:17,642 INFO [finetune.py:976] (4/7) Epoch 21, batch 3050, loss[loss=0.164, simple_loss=0.2433, pruned_loss=0.04232, over 4793.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2499, pruned_loss=0.05326, over 952868.18 frames. ], batch size: 29, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:39:54,619 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8167, 1.3973, 1.9452, 1.8478, 1.6572, 1.6181, 1.7926, 1.8184], + device='cuda:4'), covar=tensor([0.4339, 0.4135, 0.3117, 0.3591, 0.4499, 0.3799, 0.4457, 0.3010], + device='cuda:4'), in_proj_covar=tensor([0.0256, 0.0243, 0.0264, 0.0283, 0.0281, 0.0257, 0.0291, 0.0245], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:40:13,667 INFO [finetune.py:976] (4/7) Epoch 21, batch 3100, loss[loss=0.1795, simple_loss=0.2443, pruned_loss=0.0573, over 4149.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2486, pruned_loss=0.05237, over 954099.70 frames. ], batch size: 65, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:40:19,666 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.886e+01 1.482e+02 1.759e+02 2.208e+02 4.258e+02, threshold=3.518e+02, percent-clipped=1.0 +2023-03-27 01:40:29,996 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 01:40:31,789 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0377, 1.8686, 1.6893, 1.7096, 1.7759, 1.6883, 1.8146, 2.4244], + device='cuda:4'), covar=tensor([0.2799, 0.3399, 0.2728, 0.2895, 0.3326, 0.2011, 0.3044, 0.1423], + device='cuda:4'), in_proj_covar=tensor([0.0287, 0.0262, 0.0232, 0.0276, 0.0252, 0.0222, 0.0251, 0.0233], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:40:46,800 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=117685.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:40:55,378 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1757, 1.9053, 2.3884, 1.4788, 2.1633, 2.3870, 1.7300, 2.5338], + device='cuda:4'), covar=tensor([0.1275, 0.1953, 0.1391, 0.2211, 0.0917, 0.1417, 0.2660, 0.0832], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0205, 0.0191, 0.0190, 0.0174, 0.0214, 0.0218, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:40:57,219 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6972, 1.6274, 1.8372, 1.1573, 1.6550, 1.8779, 1.5547, 2.0843], + device='cuda:4'), covar=tensor([0.1238, 0.2227, 0.1265, 0.1897, 0.1011, 0.1369, 0.3127, 0.0910], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0205, 0.0191, 0.0190, 0.0174, 0.0214, 0.0218, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:40:58,310 INFO [finetune.py:976] (4/7) Epoch 21, batch 3150, loss[loss=0.1355, simple_loss=0.208, pruned_loss=0.03151, over 3918.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2459, pruned_loss=0.05203, over 954938.82 frames. ], batch size: 17, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:41:31,632 INFO [finetune.py:976] (4/7) Epoch 21, batch 3200, loss[loss=0.1497, simple_loss=0.226, pruned_loss=0.03669, over 4827.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2435, pruned_loss=0.05164, over 956651.20 frames. ], batch size: 38, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:41:34,036 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.569e+02 1.801e+02 2.101e+02 4.822e+02, threshold=3.602e+02, percent-clipped=2.0 +2023-03-27 01:41:34,204 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1913, 1.9480, 1.7141, 1.9083, 1.8892, 1.8605, 1.9150, 2.6568], + device='cuda:4'), covar=tensor([0.3641, 0.4446, 0.3474, 0.4096, 0.4411, 0.2449, 0.4018, 0.1620], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0264, 0.0234, 0.0278, 0.0254, 0.0224, 0.0253, 0.0235], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:41:59,981 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 01:42:00,929 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-27 01:42:05,171 INFO [finetune.py:976] (4/7) Epoch 21, batch 3250, loss[loss=0.1851, simple_loss=0.2629, pruned_loss=0.05358, over 4833.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2445, pruned_loss=0.05198, over 955100.68 frames. ], batch size: 33, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:42:20,614 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-27 01:42:27,000 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.68 vs. limit=2.0 +2023-03-27 01:42:38,400 INFO [finetune.py:976] (4/7) Epoch 21, batch 3300, loss[loss=0.2139, simple_loss=0.2747, pruned_loss=0.07657, over 4827.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2478, pruned_loss=0.0534, over 955858.45 frames. ], batch size: 33, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:42:40,846 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.148e+02 1.638e+02 1.917e+02 2.241e+02 9.038e+02, threshold=3.833e+02, percent-clipped=2.0 +2023-03-27 01:42:54,837 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117877.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:43:09,251 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4587, 2.3085, 1.9728, 1.1482, 2.1626, 1.8975, 1.8155, 2.0826], + device='cuda:4'), covar=tensor([0.1016, 0.0805, 0.1835, 0.2037, 0.1413, 0.2424, 0.2217, 0.1097], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0191, 0.0198, 0.0182, 0.0209, 0.0208, 0.0221, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:43:11,572 INFO [finetune.py:976] (4/7) Epoch 21, batch 3350, loss[loss=0.1318, simple_loss=0.214, pruned_loss=0.02478, over 4738.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2502, pruned_loss=0.05404, over 957301.55 frames. ], batch size: 27, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:43:22,355 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6211, 2.4417, 1.9901, 2.7453, 2.6543, 2.3267, 3.1233, 2.5826], + device='cuda:4'), covar=tensor([0.1328, 0.2496, 0.3167, 0.2714, 0.2629, 0.1657, 0.2995, 0.1898], + device='cuda:4'), in_proj_covar=tensor([0.0186, 0.0187, 0.0235, 0.0253, 0.0247, 0.0204, 0.0214, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:43:25,774 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=117925.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:43:45,043 INFO [finetune.py:976] (4/7) Epoch 21, batch 3400, loss[loss=0.1885, simple_loss=0.2749, pruned_loss=0.05106, over 4912.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2511, pruned_loss=0.05387, over 956571.76 frames. ], batch size: 46, lr: 3.19e-03, grad_scale: 32.0 +2023-03-27 01:43:47,451 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.189e+02 1.619e+02 1.880e+02 2.233e+02 5.629e+02, threshold=3.761e+02, percent-clipped=2.0 +2023-03-27 01:44:05,874 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=117985.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:44:19,702 INFO [finetune.py:976] (4/7) Epoch 21, batch 3450, loss[loss=0.2043, simple_loss=0.2529, pruned_loss=0.07785, over 4768.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2504, pruned_loss=0.05361, over 954951.19 frames. ], batch size: 26, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:44:23,458 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1950, 1.7329, 2.3593, 1.6164, 2.0581, 2.2999, 1.6437, 2.4798], + device='cuda:4'), covar=tensor([0.1257, 0.2142, 0.1332, 0.2007, 0.0946, 0.1580, 0.2880, 0.0901], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0204, 0.0191, 0.0189, 0.0174, 0.0214, 0.0217, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:44:39,325 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118033.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:44:44,146 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118040.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:44:59,763 INFO [finetune.py:976] (4/7) Epoch 21, batch 3500, loss[loss=0.2157, simple_loss=0.2724, pruned_loss=0.0795, over 4877.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2466, pruned_loss=0.05218, over 954620.83 frames. ], batch size: 34, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:45:02,216 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.127e+01 1.501e+02 1.833e+02 2.184e+02 3.839e+02, threshold=3.666e+02, percent-clipped=2.0 +2023-03-27 01:45:52,332 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118101.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:45:57,921 INFO [finetune.py:976] (4/7) Epoch 21, batch 3550, loss[loss=0.1547, simple_loss=0.2256, pruned_loss=0.04191, over 4911.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2448, pruned_loss=0.05165, over 955023.82 frames. ], batch size: 36, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:46:30,093 INFO [finetune.py:976] (4/7) Epoch 21, batch 3600, loss[loss=0.2525, simple_loss=0.2999, pruned_loss=0.1025, over 4730.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2428, pruned_loss=0.05151, over 955369.15 frames. ], batch size: 59, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:46:33,039 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.102e+02 1.558e+02 1.902e+02 2.180e+02 3.976e+02, threshold=3.804e+02, percent-clipped=2.0 +2023-03-27 01:46:39,806 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118168.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:46:59,874 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2507, 1.8190, 2.6820, 4.1343, 2.9364, 2.6877, 0.5935, 3.5314], + device='cuda:4'), covar=tensor([0.1457, 0.1347, 0.1169, 0.0412, 0.0625, 0.1638, 0.2061, 0.0292], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0116, 0.0133, 0.0162, 0.0100, 0.0137, 0.0124, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 01:47:03,625 INFO [finetune.py:976] (4/7) Epoch 21, batch 3650, loss[loss=0.1989, simple_loss=0.2814, pruned_loss=0.05821, over 4901.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2452, pruned_loss=0.05261, over 955521.62 frames. ], batch size: 35, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:47:19,815 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118229.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 01:47:22,387 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-27 01:47:26,158 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6272, 1.4226, 2.1609, 3.3813, 2.3430, 2.3697, 1.0841, 2.8283], + device='cuda:4'), covar=tensor([0.1788, 0.1514, 0.1203, 0.0528, 0.0718, 0.1388, 0.1775, 0.0443], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0117, 0.0133, 0.0163, 0.0101, 0.0137, 0.0125, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 01:47:36,721 INFO [finetune.py:976] (4/7) Epoch 21, batch 3700, loss[loss=0.1767, simple_loss=0.2507, pruned_loss=0.05131, over 4908.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2482, pruned_loss=0.05272, over 956574.78 frames. ], batch size: 43, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:47:39,057 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.607e+02 1.941e+02 2.377e+02 3.454e+02, threshold=3.882e+02, percent-clipped=0.0 +2023-03-27 01:47:46,808 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118269.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:47:47,619 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-27 01:48:00,467 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118289.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:48:10,298 INFO [finetune.py:976] (4/7) Epoch 21, batch 3750, loss[loss=0.1915, simple_loss=0.2645, pruned_loss=0.0592, over 4815.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2491, pruned_loss=0.05295, over 955218.60 frames. ], batch size: 40, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:48:25,129 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1526, 1.7948, 2.5673, 3.9789, 2.7300, 2.7351, 0.6588, 3.3573], + device='cuda:4'), covar=tensor([0.1585, 0.1399, 0.1320, 0.0441, 0.0719, 0.1592, 0.2219, 0.0370], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0117, 0.0133, 0.0164, 0.0101, 0.0137, 0.0125, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 01:48:26,348 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118329.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:48:26,920 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5990, 1.4584, 2.1129, 2.9634, 1.9911, 2.1840, 0.9258, 2.4633], + device='cuda:4'), covar=tensor([0.1730, 0.1485, 0.1222, 0.0648, 0.0893, 0.1346, 0.1943, 0.0562], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0117, 0.0133, 0.0164, 0.0101, 0.0137, 0.0125, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 01:48:26,966 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118330.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:48:40,956 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118350.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:48:43,714 INFO [finetune.py:976] (4/7) Epoch 21, batch 3800, loss[loss=0.1507, simple_loss=0.2328, pruned_loss=0.03429, over 4919.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2499, pruned_loss=0.05301, over 953965.39 frames. ], batch size: 42, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:48:46,091 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.364e+01 1.561e+02 1.815e+02 2.293e+02 4.441e+02, threshold=3.631e+02, percent-clipped=1.0 +2023-03-27 01:49:06,610 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118390.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:49:09,551 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5567, 1.6005, 2.0615, 1.9063, 1.7158, 3.5476, 1.5012, 1.6562], + device='cuda:4'), covar=tensor([0.0934, 0.1759, 0.1016, 0.0879, 0.1503, 0.0291, 0.1487, 0.1717], + device='cuda:4'), in_proj_covar=tensor([0.0074, 0.0081, 0.0074, 0.0076, 0.0091, 0.0081, 0.0086, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 01:49:11,214 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118396.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:49:17,067 INFO [finetune.py:976] (4/7) Epoch 21, batch 3850, loss[loss=0.2108, simple_loss=0.2725, pruned_loss=0.07456, over 4863.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2483, pruned_loss=0.0518, over 954443.87 frames. ], batch size: 31, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:49:28,521 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3231, 2.2005, 1.8280, 2.2285, 2.1720, 1.9945, 2.5338, 2.3963], + device='cuda:4'), covar=tensor([0.1055, 0.2007, 0.2436, 0.2572, 0.2116, 0.1396, 0.3507, 0.1354], + device='cuda:4'), in_proj_covar=tensor([0.0187, 0.0188, 0.0236, 0.0253, 0.0247, 0.0205, 0.0215, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:49:34,212 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-27 01:49:50,287 INFO [finetune.py:976] (4/7) Epoch 21, batch 3900, loss[loss=0.1418, simple_loss=0.2196, pruned_loss=0.03199, over 4788.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2456, pruned_loss=0.05118, over 955729.15 frames. ], batch size: 29, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:49:52,687 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.530e+02 1.819e+02 2.327e+02 4.856e+02, threshold=3.639e+02, percent-clipped=2.0 +2023-03-27 01:50:09,215 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8016, 4.0020, 3.8200, 1.9397, 4.0980, 3.2371, 1.1221, 2.9845], + device='cuda:4'), covar=tensor([0.2416, 0.2067, 0.1404, 0.3375, 0.0972, 0.0856, 0.4266, 0.1423], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0177, 0.0158, 0.0130, 0.0160, 0.0122, 0.0147, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 01:50:11,752 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9883, 1.4418, 1.8830, 1.8567, 1.7169, 1.7111, 1.7721, 1.8471], + device='cuda:4'), covar=tensor([0.4701, 0.4245, 0.3985, 0.4250, 0.5763, 0.4546, 0.5196, 0.3813], + device='cuda:4'), in_proj_covar=tensor([0.0255, 0.0243, 0.0264, 0.0283, 0.0281, 0.0257, 0.0291, 0.0245], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:50:12,309 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6734, 1.6538, 2.1577, 2.0039, 1.7092, 3.7931, 1.6773, 1.6905], + device='cuda:4'), covar=tensor([0.0985, 0.1843, 0.1014, 0.0928, 0.1641, 0.0223, 0.1468, 0.1797], + device='cuda:4'), in_proj_covar=tensor([0.0074, 0.0081, 0.0074, 0.0076, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 01:50:25,013 INFO [finetune.py:976] (4/7) Epoch 21, batch 3950, loss[loss=0.1685, simple_loss=0.2403, pruned_loss=0.04836, over 4904.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2429, pruned_loss=0.0506, over 957635.04 frames. ], batch size: 32, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:50:38,354 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-27 01:50:48,091 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118524.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 01:50:49,870 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1433, 1.8219, 2.1611, 2.1151, 1.8381, 1.8721, 2.1006, 2.0366], + device='cuda:4'), covar=tensor([0.3690, 0.3771, 0.2991, 0.3927, 0.5038, 0.3964, 0.4443, 0.2836], + device='cuda:4'), in_proj_covar=tensor([0.0256, 0.0244, 0.0265, 0.0284, 0.0282, 0.0258, 0.0292, 0.0246], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:51:16,459 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9115, 1.6997, 2.0076, 1.4342, 1.9585, 2.1253, 1.5852, 2.2826], + device='cuda:4'), covar=tensor([0.1227, 0.2094, 0.1426, 0.1884, 0.0929, 0.1394, 0.2754, 0.0873], + device='cuda:4'), in_proj_covar=tensor([0.0194, 0.0207, 0.0192, 0.0191, 0.0176, 0.0216, 0.0219, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:51:19,655 INFO [finetune.py:976] (4/7) Epoch 21, batch 4000, loss[loss=0.176, simple_loss=0.247, pruned_loss=0.05255, over 4830.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2419, pruned_loss=0.05079, over 958489.96 frames. ], batch size: 33, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:51:19,755 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6670, 1.1557, 0.8315, 1.5569, 2.0701, 1.4046, 1.4999, 1.5466], + device='cuda:4'), covar=tensor([0.1496, 0.2257, 0.2061, 0.1313, 0.1942, 0.1990, 0.1484, 0.1970], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0092, 0.0119, 0.0093, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 01:51:26,592 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.742e+01 1.543e+02 1.809e+02 2.201e+02 4.154e+02, threshold=3.618e+02, percent-clipped=2.0 +2023-03-27 01:51:27,582 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 01:51:56,643 INFO [finetune.py:976] (4/7) Epoch 21, batch 4050, loss[loss=0.1838, simple_loss=0.2526, pruned_loss=0.05749, over 4916.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2453, pruned_loss=0.05201, over 958058.66 frames. ], batch size: 37, lr: 3.18e-03, grad_scale: 32.0 +2023-03-27 01:52:11,489 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118625.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:52:15,686 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=118631.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:52:22,235 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4108, 1.6006, 0.7481, 2.2448, 2.7236, 1.9438, 1.9715, 2.0563], + device='cuda:4'), covar=tensor([0.1330, 0.2026, 0.2283, 0.1121, 0.1658, 0.1906, 0.1422, 0.1939], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0110, 0.0092, 0.0119, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 01:52:24,639 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118645.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:52:30,075 INFO [finetune.py:976] (4/7) Epoch 21, batch 4100, loss[loss=0.2013, simple_loss=0.2703, pruned_loss=0.06615, over 4824.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2465, pruned_loss=0.05192, over 957967.87 frames. ], batch size: 33, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:52:33,000 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2246, 3.5999, 3.8446, 4.0260, 3.9599, 3.7252, 4.3081, 1.4158], + device='cuda:4'), covar=tensor([0.0778, 0.0917, 0.0918, 0.1102, 0.1211, 0.1751, 0.0701, 0.5634], + device='cuda:4'), in_proj_covar=tensor([0.0351, 0.0245, 0.0282, 0.0293, 0.0336, 0.0286, 0.0306, 0.0301], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:52:33,507 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.828e+01 1.601e+02 1.824e+02 2.338e+02 3.980e+02, threshold=3.647e+02, percent-clipped=1.0 +2023-03-27 01:52:51,558 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118685.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:52:56,341 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=118692.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:52:58,686 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118696.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:53:03,515 INFO [finetune.py:976] (4/7) Epoch 21, batch 4150, loss[loss=0.1948, simple_loss=0.2555, pruned_loss=0.06699, over 4821.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2484, pruned_loss=0.05276, over 955634.71 frames. ], batch size: 33, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:53:31,226 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118744.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:53:37,401 INFO [finetune.py:976] (4/7) Epoch 21, batch 4200, loss[loss=0.1327, simple_loss=0.2053, pruned_loss=0.03008, over 4280.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2487, pruned_loss=0.05231, over 952585.24 frames. ], batch size: 65, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:53:39,822 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.932e+01 1.492e+02 1.761e+02 2.066e+02 3.902e+02, threshold=3.521e+02, percent-clipped=1.0 +2023-03-27 01:53:48,087 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.99 vs. limit=5.0 +2023-03-27 01:54:00,192 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 01:54:02,337 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2516, 3.6702, 3.8255, 4.0431, 4.0170, 3.7479, 4.3067, 1.3591], + device='cuda:4'), covar=tensor([0.0805, 0.0895, 0.0838, 0.0970, 0.1292, 0.1524, 0.0745, 0.5786], + device='cuda:4'), in_proj_covar=tensor([0.0351, 0.0244, 0.0281, 0.0293, 0.0335, 0.0285, 0.0305, 0.0300], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:54:11,364 INFO [finetune.py:976] (4/7) Epoch 21, batch 4250, loss[loss=0.2087, simple_loss=0.2594, pruned_loss=0.07896, over 4932.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2465, pruned_loss=0.05139, over 954186.99 frames. ], batch size: 33, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:54:25,965 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118824.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:54:45,142 INFO [finetune.py:976] (4/7) Epoch 21, batch 4300, loss[loss=0.139, simple_loss=0.2061, pruned_loss=0.03592, over 4829.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2441, pruned_loss=0.05059, over 955803.43 frames. ], batch size: 39, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:54:47,576 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.108e+02 1.563e+02 1.855e+02 2.179e+02 3.656e+02, threshold=3.709e+02, percent-clipped=1.0 +2023-03-27 01:54:57,553 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118872.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:55:18,873 INFO [finetune.py:976] (4/7) Epoch 21, batch 4350, loss[loss=0.1766, simple_loss=0.2463, pruned_loss=0.05344, over 4829.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2419, pruned_loss=0.05034, over 956065.24 frames. ], batch size: 30, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:55:33,252 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118925.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:55:48,466 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118945.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:55:59,294 INFO [finetune.py:976] (4/7) Epoch 21, batch 4400, loss[loss=0.2372, simple_loss=0.3049, pruned_loss=0.08471, over 4833.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2437, pruned_loss=0.05171, over 955478.65 frames. ], batch size: 40, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:56:01,710 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.753e+01 1.466e+02 1.745e+02 2.136e+02 3.634e+02, threshold=3.490e+02, percent-clipped=0.0 +2023-03-27 01:56:03,056 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3803, 1.3987, 2.2652, 1.7635, 1.7234, 4.0149, 1.3119, 1.5651], + device='cuda:4'), covar=tensor([0.1057, 0.1840, 0.1257, 0.1067, 0.1608, 0.0207, 0.1659, 0.1856], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0092, 0.0081, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 01:56:07,263 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-27 01:56:18,413 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118973.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:56:18,457 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3456, 1.4493, 1.8830, 1.6786, 1.4474, 3.2590, 1.3759, 1.4856], + device='cuda:4'), covar=tensor([0.1014, 0.1881, 0.1180, 0.1012, 0.1743, 0.0245, 0.1531, 0.1826], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0092, 0.0081, 0.0086, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 01:56:31,289 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=118985.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:56:36,932 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=118987.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:56:40,528 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=118993.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:56:46,452 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9620, 1.6087, 2.0258, 1.4231, 1.9009, 2.0837, 1.4807, 2.2163], + device='cuda:4'), covar=tensor([0.1157, 0.2135, 0.1465, 0.1850, 0.0935, 0.1355, 0.3169, 0.0919], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0205, 0.0191, 0.0189, 0.0174, 0.0213, 0.0216, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:56:51,290 INFO [finetune.py:976] (4/7) Epoch 21, batch 4450, loss[loss=0.2107, simple_loss=0.2911, pruned_loss=0.06512, over 4853.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2465, pruned_loss=0.0522, over 956875.86 frames. ], batch size: 44, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:57:11,421 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=119033.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:57:25,003 INFO [finetune.py:976] (4/7) Epoch 21, batch 4500, loss[loss=0.1917, simple_loss=0.2531, pruned_loss=0.06514, over 4896.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2471, pruned_loss=0.05247, over 954575.31 frames. ], batch size: 35, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:57:25,733 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4936, 1.3341, 1.2504, 1.4857, 1.6821, 1.5324, 1.1168, 1.2774], + device='cuda:4'), covar=tensor([0.2347, 0.2056, 0.2065, 0.1768, 0.1569, 0.1320, 0.2384, 0.1861], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0210, 0.0214, 0.0196, 0.0243, 0.0189, 0.0218, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:57:27,416 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.079e+02 1.548e+02 1.910e+02 2.429e+02 4.520e+02, threshold=3.820e+02, percent-clipped=3.0 +2023-03-27 01:57:31,708 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3642, 1.2882, 1.2430, 1.3241, 1.6014, 1.5138, 1.3708, 1.1903], + device='cuda:4'), covar=tensor([0.0340, 0.0305, 0.0639, 0.0313, 0.0258, 0.0544, 0.0313, 0.0442], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0106, 0.0143, 0.0112, 0.0099, 0.0110, 0.0100, 0.0112], + device='cuda:4'), out_proj_covar=tensor([7.5752e-05, 8.1315e-05, 1.1255e-04, 8.5667e-05, 7.6747e-05, 8.1723e-05, + 7.4382e-05, 8.5551e-05], device='cuda:4') +2023-03-27 01:57:31,721 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6916, 1.6343, 1.4370, 1.8212, 2.0213, 1.8402, 1.3039, 1.3980], + device='cuda:4'), covar=tensor([0.2360, 0.2095, 0.1946, 0.1684, 0.1652, 0.1137, 0.2487, 0.2004], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0211, 0.0214, 0.0196, 0.0243, 0.0189, 0.0218, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:57:38,311 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=119075.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:57:41,015 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7596, 3.7240, 3.5852, 1.9533, 3.8236, 2.9211, 0.8306, 2.7561], + device='cuda:4'), covar=tensor([0.2869, 0.1691, 0.1665, 0.3243, 0.1011, 0.1032, 0.4547, 0.1439], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0178, 0.0160, 0.0130, 0.0161, 0.0123, 0.0148, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 01:57:50,863 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6103, 1.6714, 1.5139, 0.9947, 1.6875, 1.8879, 1.8668, 1.4344], + device='cuda:4'), covar=tensor([0.0913, 0.0595, 0.0524, 0.0486, 0.0465, 0.0613, 0.0330, 0.0699], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0149, 0.0126, 0.0123, 0.0131, 0.0128, 0.0141, 0.0148], + device='cuda:4'), out_proj_covar=tensor([8.9403e-05, 1.0749e-04, 9.0033e-05, 8.6654e-05, 9.1821e-05, 9.1552e-05, + 1.0132e-04, 1.0581e-04], device='cuda:4') +2023-03-27 01:57:51,487 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=119093.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:57:58,451 INFO [finetune.py:976] (4/7) Epoch 21, batch 4550, loss[loss=0.1842, simple_loss=0.2547, pruned_loss=0.05679, over 4804.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2503, pruned_loss=0.0538, over 954938.61 frames. ], batch size: 45, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:58:19,748 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=119136.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 01:58:31,550 INFO [finetune.py:976] (4/7) Epoch 21, batch 4600, loss[loss=0.1788, simple_loss=0.2379, pruned_loss=0.05981, over 4745.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2491, pruned_loss=0.05308, over 955693.71 frames. ], batch size: 54, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:58:31,681 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=119154.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 01:58:34,457 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.653e+02 1.869e+02 2.317e+02 3.451e+02, threshold=3.738e+02, percent-clipped=0.0 +2023-03-27 01:58:46,605 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2462, 2.0921, 1.6641, 2.1400, 2.1763, 1.9039, 2.3592, 2.2631], + device='cuda:4'), covar=tensor([0.1238, 0.1932, 0.2842, 0.2214, 0.2292, 0.1540, 0.3048, 0.1539], + device='cuda:4'), in_proj_covar=tensor([0.0187, 0.0189, 0.0236, 0.0254, 0.0247, 0.0205, 0.0216, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:59:04,159 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.2753, 4.5353, 4.8019, 5.0908, 4.9796, 4.6316, 5.3493, 1.6347], + device='cuda:4'), covar=tensor([0.0815, 0.0872, 0.0884, 0.0885, 0.1275, 0.1801, 0.0593, 0.6110], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0244, 0.0280, 0.0293, 0.0335, 0.0285, 0.0304, 0.0299], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 01:59:05,261 INFO [finetune.py:976] (4/7) Epoch 21, batch 4650, loss[loss=0.1715, simple_loss=0.2259, pruned_loss=0.05855, over 4740.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2465, pruned_loss=0.05247, over 954225.11 frames. ], batch size: 23, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:59:37,238 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3786, 1.4746, 1.7603, 1.7072, 1.6138, 3.0755, 1.3877, 1.5508], + device='cuda:4'), covar=tensor([0.0987, 0.1706, 0.1225, 0.0946, 0.1463, 0.0265, 0.1444, 0.1659], + device='cuda:4'), in_proj_covar=tensor([0.0074, 0.0082, 0.0074, 0.0077, 0.0092, 0.0081, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 01:59:38,313 INFO [finetune.py:976] (4/7) Epoch 21, batch 4700, loss[loss=0.1653, simple_loss=0.231, pruned_loss=0.04982, over 4868.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2431, pruned_loss=0.05145, over 952231.72 frames. ], batch size: 31, lr: 3.18e-03, grad_scale: 64.0 +2023-03-27 01:59:40,728 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.102e+02 1.521e+02 1.791e+02 2.382e+02 6.096e+02, threshold=3.583e+02, percent-clipped=7.0 +2023-03-27 01:59:59,461 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=119287.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:00:10,097 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.10 vs. limit=5.0 +2023-03-27 02:00:11,568 INFO [finetune.py:976] (4/7) Epoch 21, batch 4750, loss[loss=0.2035, simple_loss=0.262, pruned_loss=0.07255, over 4893.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2408, pruned_loss=0.05057, over 949825.94 frames. ], batch size: 35, lr: 3.17e-03, grad_scale: 64.0 +2023-03-27 02:00:31,335 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=119335.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:00:44,658 INFO [finetune.py:976] (4/7) Epoch 21, batch 4800, loss[loss=0.1912, simple_loss=0.2591, pruned_loss=0.06166, over 4901.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2438, pruned_loss=0.05168, over 951597.47 frames. ], batch size: 35, lr: 3.17e-03, grad_scale: 64.0 +2023-03-27 02:00:47,500 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.052e+02 1.488e+02 1.781e+02 2.195e+02 3.360e+02, threshold=3.562e+02, percent-clipped=0.0 +2023-03-27 02:00:50,072 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8713, 1.8299, 1.4895, 1.7721, 1.8949, 1.5584, 2.1641, 1.8767], + device='cuda:4'), covar=tensor([0.1365, 0.1750, 0.2777, 0.2331, 0.2422, 0.1669, 0.2706, 0.1610], + device='cuda:4'), in_proj_covar=tensor([0.0185, 0.0186, 0.0233, 0.0251, 0.0244, 0.0202, 0.0213, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:00:54,132 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9327, 2.6602, 2.2314, 1.1854, 2.4729, 2.2250, 1.9770, 2.4266], + device='cuda:4'), covar=tensor([0.0708, 0.0640, 0.1166, 0.1807, 0.1110, 0.1698, 0.1735, 0.0756], + device='cuda:4'), in_proj_covar=tensor([0.0171, 0.0192, 0.0200, 0.0184, 0.0211, 0.0210, 0.0225, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:01:22,407 INFO [finetune.py:976] (4/7) Epoch 21, batch 4850, loss[loss=0.163, simple_loss=0.231, pruned_loss=0.04754, over 4708.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2469, pruned_loss=0.05259, over 950842.05 frames. ], batch size: 23, lr: 3.17e-03, grad_scale: 64.0 +2023-03-27 02:01:27,731 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-03-27 02:01:30,445 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5147, 1.5278, 2.0452, 1.8277, 1.5662, 3.5568, 1.4157, 1.5593], + device='cuda:4'), covar=tensor([0.0965, 0.1812, 0.1035, 0.0996, 0.1742, 0.0204, 0.1589, 0.1851], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0074, 0.0077, 0.0091, 0.0081, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 02:01:48,916 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9481, 1.8409, 1.5803, 1.4851, 1.9743, 1.7074, 1.8334, 1.9271], + device='cuda:4'), covar=tensor([0.1312, 0.1916, 0.2956, 0.2356, 0.2579, 0.1671, 0.2941, 0.1735], + device='cuda:4'), in_proj_covar=tensor([0.0185, 0.0187, 0.0234, 0.0252, 0.0245, 0.0203, 0.0214, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:01:55,525 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=119431.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:02:15,265 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=119449.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 02:02:19,177 INFO [finetune.py:976] (4/7) Epoch 21, batch 4900, loss[loss=0.221, simple_loss=0.277, pruned_loss=0.08247, over 4811.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.249, pruned_loss=0.05358, over 951848.05 frames. ], batch size: 40, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:02:25,215 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.684e+02 1.937e+02 2.365e+02 4.201e+02, threshold=3.874e+02, percent-clipped=2.0 +2023-03-27 02:02:55,580 INFO [finetune.py:976] (4/7) Epoch 21, batch 4950, loss[loss=0.146, simple_loss=0.2137, pruned_loss=0.0392, over 4223.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2494, pruned_loss=0.0529, over 953255.88 frames. ], batch size: 66, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:03:00,496 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1253, 1.2754, 1.4175, 0.6860, 1.3018, 1.5644, 1.5975, 1.2917], + device='cuda:4'), covar=tensor([0.0818, 0.0533, 0.0474, 0.0469, 0.0469, 0.0497, 0.0307, 0.0601], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0150, 0.0126, 0.0123, 0.0131, 0.0129, 0.0142, 0.0148], + device='cuda:4'), out_proj_covar=tensor([9.0017e-05, 1.0814e-04, 9.0498e-05, 8.6992e-05, 9.2010e-05, 9.1845e-05, + 1.0199e-04, 1.0618e-04], device='cuda:4') +2023-03-27 02:03:01,057 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3724, 1.2383, 1.5778, 2.4781, 1.6952, 2.2259, 0.8914, 2.1455], + device='cuda:4'), covar=tensor([0.1846, 0.1581, 0.1271, 0.0733, 0.0956, 0.1106, 0.1693, 0.0622], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0118, 0.0135, 0.0166, 0.0102, 0.0139, 0.0126, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 02:03:02,755 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6578, 1.5773, 1.5472, 1.5740, 1.3819, 3.4306, 1.4536, 1.8631], + device='cuda:4'), covar=tensor([0.3293, 0.2502, 0.2054, 0.2408, 0.1575, 0.0202, 0.2671, 0.1210], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0124, 0.0114, 0.0097, 0.0095, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 02:03:02,770 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=119514.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:03:29,008 INFO [finetune.py:976] (4/7) Epoch 21, batch 5000, loss[loss=0.1696, simple_loss=0.2484, pruned_loss=0.04538, over 4789.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.248, pruned_loss=0.05235, over 953410.65 frames. ], batch size: 26, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:03:32,983 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.554e+02 1.853e+02 2.138e+02 3.358e+02, threshold=3.705e+02, percent-clipped=0.0 +2023-03-27 02:03:43,313 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=119575.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:03:49,979 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.66 vs. limit=2.0 +2023-03-27 02:04:02,235 INFO [finetune.py:976] (4/7) Epoch 21, batch 5050, loss[loss=0.1438, simple_loss=0.2281, pruned_loss=0.02975, over 4813.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2443, pruned_loss=0.05087, over 953546.78 frames. ], batch size: 25, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:04:25,247 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8277, 1.6875, 1.4763, 1.3894, 1.8690, 1.5602, 1.7304, 1.7848], + device='cuda:4'), covar=tensor([0.1390, 0.1794, 0.2990, 0.2500, 0.2632, 0.1708, 0.3106, 0.1803], + device='cuda:4'), in_proj_covar=tensor([0.0187, 0.0188, 0.0235, 0.0253, 0.0246, 0.0204, 0.0215, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:04:31,169 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1301, 1.8637, 2.1827, 1.6366, 2.1747, 2.4015, 1.7429, 2.4985], + device='cuda:4'), covar=tensor([0.1152, 0.1908, 0.1501, 0.1777, 0.0917, 0.1361, 0.2659, 0.0804], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0205, 0.0192, 0.0190, 0.0174, 0.0214, 0.0217, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:04:35,255 INFO [finetune.py:976] (4/7) Epoch 21, batch 5100, loss[loss=0.1409, simple_loss=0.2137, pruned_loss=0.03399, over 4855.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2415, pruned_loss=0.04972, over 955413.55 frames. ], batch size: 44, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:04:39,201 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.052e+02 1.479e+02 1.750e+02 2.173e+02 3.976e+02, threshold=3.500e+02, percent-clipped=1.0 +2023-03-27 02:04:43,185 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-27 02:04:43,503 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=119665.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:05:08,857 INFO [finetune.py:976] (4/7) Epoch 21, batch 5150, loss[loss=0.2721, simple_loss=0.3325, pruned_loss=0.1058, over 4842.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2417, pruned_loss=0.04995, over 957255.48 frames. ], batch size: 47, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:05:24,645 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=119726.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:05:27,584 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=119731.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:05:30,180 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-27 02:05:38,883 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=119749.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 02:05:42,228 INFO [finetune.py:976] (4/7) Epoch 21, batch 5200, loss[loss=0.2255, simple_loss=0.2971, pruned_loss=0.0769, over 4919.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2466, pruned_loss=0.05183, over 958499.68 frames. ], batch size: 36, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:05:45,718 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.656e+02 1.877e+02 2.270e+02 4.720e+02, threshold=3.754e+02, percent-clipped=1.0 +2023-03-27 02:05:59,325 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=119779.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:06:10,787 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=119797.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:06:15,131 INFO [finetune.py:976] (4/7) Epoch 21, batch 5250, loss[loss=0.1884, simple_loss=0.2696, pruned_loss=0.0536, over 4795.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.249, pruned_loss=0.05285, over 955889.54 frames. ], batch size: 41, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:06:26,416 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 02:06:32,213 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7661, 3.6364, 3.4604, 1.7283, 3.7162, 2.8420, 1.0041, 2.5056], + device='cuda:4'), covar=tensor([0.2300, 0.1689, 0.1585, 0.3063, 0.1048, 0.0888, 0.3941, 0.1464], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0177, 0.0158, 0.0129, 0.0160, 0.0122, 0.0147, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 02:06:59,460 INFO [finetune.py:976] (4/7) Epoch 21, batch 5300, loss[loss=0.1638, simple_loss=0.2351, pruned_loss=0.04629, over 4760.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2501, pruned_loss=0.05294, over 956205.17 frames. ], batch size: 54, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:07:07,161 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.978e+01 1.512e+02 1.747e+02 2.069e+02 4.039e+02, threshold=3.495e+02, percent-clipped=1.0 +2023-03-27 02:07:19,027 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=119870.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:07:54,831 INFO [finetune.py:976] (4/7) Epoch 21, batch 5350, loss[loss=0.2009, simple_loss=0.2636, pruned_loss=0.06908, over 4739.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2497, pruned_loss=0.05242, over 955836.11 frames. ], batch size: 54, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:08:00,895 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7262, 1.3193, 0.9426, 1.6569, 2.2150, 1.2388, 1.4949, 1.5787], + device='cuda:4'), covar=tensor([0.1436, 0.1983, 0.1812, 0.1211, 0.1767, 0.1733, 0.1508, 0.2046], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0110, 0.0092, 0.0119, 0.0093, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 02:08:10,925 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9583, 1.7602, 2.2692, 1.5814, 2.0396, 2.2620, 1.6311, 2.4035], + device='cuda:4'), covar=tensor([0.1322, 0.1886, 0.1430, 0.1966, 0.0900, 0.1368, 0.2946, 0.0729], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0205, 0.0192, 0.0190, 0.0174, 0.0214, 0.0217, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:08:10,940 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.5866, 1.6642, 1.5745, 0.8421, 1.7170, 1.9477, 1.8755, 1.4558], + device='cuda:4'), covar=tensor([0.0961, 0.0591, 0.0544, 0.0593, 0.0443, 0.0525, 0.0360, 0.0634], + device='cuda:4'), in_proj_covar=tensor([0.0124, 0.0150, 0.0127, 0.0124, 0.0132, 0.0129, 0.0143, 0.0149], + device='cuda:4'), out_proj_covar=tensor([9.0179e-05, 1.0860e-04, 9.0994e-05, 8.7453e-05, 9.2694e-05, 9.2166e-05, + 1.0229e-04, 1.0667e-04], device='cuda:4') +2023-03-27 02:08:28,088 INFO [finetune.py:976] (4/7) Epoch 21, batch 5400, loss[loss=0.1401, simple_loss=0.2144, pruned_loss=0.03288, over 4884.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2469, pruned_loss=0.05184, over 956111.13 frames. ], batch size: 32, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:08:31,173 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.474e+02 1.662e+02 2.015e+02 3.492e+02, threshold=3.324e+02, percent-clipped=0.0 +2023-03-27 02:09:02,906 INFO [finetune.py:976] (4/7) Epoch 21, batch 5450, loss[loss=0.168, simple_loss=0.2344, pruned_loss=0.05078, over 4903.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2452, pruned_loss=0.05153, over 956547.49 frames. ], batch size: 36, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:09:13,780 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=120021.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:09:27,887 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3760, 2.4101, 2.3224, 1.6208, 2.3529, 2.3971, 2.4823, 2.0473], + device='cuda:4'), covar=tensor([0.0544, 0.0463, 0.0606, 0.0807, 0.0538, 0.0695, 0.0544, 0.1002], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0137, 0.0139, 0.0121, 0.0126, 0.0139, 0.0140, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:09:36,195 INFO [finetune.py:976] (4/7) Epoch 21, batch 5500, loss[loss=0.1454, simple_loss=0.2167, pruned_loss=0.03711, over 4901.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2424, pruned_loss=0.05053, over 956806.99 frames. ], batch size: 43, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:09:39,679 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.953e+01 1.422e+02 1.683e+02 2.099e+02 3.794e+02, threshold=3.366e+02, percent-clipped=2.0 +2023-03-27 02:10:09,942 INFO [finetune.py:976] (4/7) Epoch 21, batch 5550, loss[loss=0.1869, simple_loss=0.2606, pruned_loss=0.05663, over 4850.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2436, pruned_loss=0.05124, over 956423.18 frames. ], batch size: 44, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:10:23,797 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2387, 2.1544, 1.7799, 0.8198, 1.8916, 1.7149, 1.5700, 2.0020], + device='cuda:4'), covar=tensor([0.0867, 0.0677, 0.1558, 0.1949, 0.1329, 0.2345, 0.2331, 0.0833], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0190, 0.0197, 0.0183, 0.0208, 0.0207, 0.0222, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:10:34,163 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5212, 2.3962, 2.3516, 1.9180, 2.6368, 2.7945, 2.8801, 1.7902], + device='cuda:4'), covar=tensor([0.0820, 0.0904, 0.0951, 0.1065, 0.0644, 0.0810, 0.0772, 0.1901], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0137, 0.0140, 0.0121, 0.0126, 0.0139, 0.0140, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:10:38,617 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-27 02:10:42,114 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 02:10:42,285 INFO [finetune.py:976] (4/7) Epoch 21, batch 5600, loss[loss=0.1573, simple_loss=0.2287, pruned_loss=0.04299, over 4800.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2477, pruned_loss=0.05226, over 954317.81 frames. ], batch size: 25, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:10:45,197 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.550e+02 1.831e+02 2.203e+02 3.727e+02, threshold=3.662e+02, percent-clipped=1.0 +2023-03-27 02:10:52,170 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=120170.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:11:03,273 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3926, 1.6241, 1.7414, 0.9310, 1.6633, 1.8663, 1.8919, 1.5214], + device='cuda:4'), covar=tensor([0.0862, 0.0648, 0.0629, 0.0598, 0.0601, 0.0834, 0.0445, 0.0736], + device='cuda:4'), in_proj_covar=tensor([0.0124, 0.0151, 0.0128, 0.0125, 0.0132, 0.0130, 0.0143, 0.0150], + device='cuda:4'), out_proj_covar=tensor([9.0446e-05, 1.0911e-04, 9.1297e-05, 8.7895e-05, 9.3057e-05, 9.2598e-05, + 1.0244e-04, 1.0718e-04], device='cuda:4') +2023-03-27 02:11:12,066 INFO [finetune.py:976] (4/7) Epoch 21, batch 5650, loss[loss=0.1485, simple_loss=0.2295, pruned_loss=0.03378, over 4837.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2497, pruned_loss=0.05256, over 952216.90 frames. ], batch size: 30, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:11:15,163 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1346, 1.8852, 2.6559, 4.0785, 2.8450, 2.7648, 0.9587, 3.4938], + device='cuda:4'), covar=tensor([0.1655, 0.1378, 0.1314, 0.0516, 0.0741, 0.1505, 0.1960, 0.0357], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0116, 0.0134, 0.0165, 0.0101, 0.0138, 0.0125, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 02:11:16,168 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.36 vs. limit=5.0 +2023-03-27 02:11:20,664 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=120218.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:11:41,964 INFO [finetune.py:976] (4/7) Epoch 21, batch 5700, loss[loss=0.1604, simple_loss=0.2226, pruned_loss=0.04915, over 3999.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2449, pruned_loss=0.05136, over 932276.13 frames. ], batch size: 17, lr: 3.17e-03, grad_scale: 32.0 +2023-03-27 02:11:44,946 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.458e+02 1.705e+02 2.128e+02 3.595e+02, threshold=3.409e+02, percent-clipped=0.0 +2023-03-27 02:11:54,318 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 02:12:12,193 INFO [finetune.py:976] (4/7) Epoch 22, batch 0, loss[loss=0.1706, simple_loss=0.2393, pruned_loss=0.05093, over 4885.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2393, pruned_loss=0.05093, over 4885.00 frames. ], batch size: 32, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:12:12,193 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 02:12:27,780 INFO [finetune.py:1010] (4/7) Epoch 22, validation: loss=0.1597, simple_loss=0.228, pruned_loss=0.04574, over 2265189.00 frames. +2023-03-27 02:12:27,780 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-27 02:12:35,249 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5438, 2.7662, 2.4352, 1.7826, 2.5033, 2.6654, 2.8806, 2.2818], + device='cuda:4'), covar=tensor([0.0762, 0.0610, 0.0767, 0.0942, 0.0696, 0.0825, 0.0595, 0.1102], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0137, 0.0139, 0.0121, 0.0125, 0.0139, 0.0140, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:13:15,386 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=120321.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:13:26,056 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-27 02:13:27,843 INFO [finetune.py:976] (4/7) Epoch 22, batch 50, loss[loss=0.1723, simple_loss=0.2473, pruned_loss=0.04864, over 4896.00 frames. ], tot_loss[loss=0.1853, simple_loss=0.2557, pruned_loss=0.05745, over 216275.28 frames. ], batch size: 46, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:13:45,814 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1184, 1.7599, 2.1883, 2.1048, 1.8396, 1.8670, 2.0915, 2.0130], + device='cuda:4'), covar=tensor([0.4118, 0.4090, 0.3126, 0.4368, 0.5272, 0.4000, 0.4929, 0.2999], + device='cuda:4'), in_proj_covar=tensor([0.0256, 0.0243, 0.0264, 0.0284, 0.0282, 0.0259, 0.0292, 0.0247], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:13:47,641 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0499, 1.9858, 1.7047, 1.8792, 1.8741, 1.8012, 1.8555, 2.5335], + device='cuda:4'), covar=tensor([0.3245, 0.3287, 0.3126, 0.3438, 0.3567, 0.2257, 0.3477, 0.1612], + device='cuda:4'), in_proj_covar=tensor([0.0288, 0.0262, 0.0232, 0.0276, 0.0253, 0.0223, 0.0252, 0.0234], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:13:48,207 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0067, 1.7515, 2.3902, 3.9957, 2.7501, 2.8111, 0.8170, 3.2918], + device='cuda:4'), covar=tensor([0.1732, 0.1431, 0.1394, 0.0614, 0.0764, 0.1623, 0.2124, 0.0464], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0116, 0.0134, 0.0165, 0.0101, 0.0138, 0.0125, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 02:13:48,726 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.043e+02 1.631e+02 1.972e+02 2.363e+02 4.295e+02, threshold=3.943e+02, percent-clipped=3.0 +2023-03-27 02:13:55,427 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=120369.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:14:04,334 INFO [finetune.py:976] (4/7) Epoch 22, batch 100, loss[loss=0.149, simple_loss=0.2154, pruned_loss=0.04134, over 4820.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2485, pruned_loss=0.05495, over 380644.66 frames. ], batch size: 33, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:14:23,584 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.84 vs. limit=5.0 +2023-03-27 02:14:37,022 INFO [finetune.py:976] (4/7) Epoch 22, batch 150, loss[loss=0.1693, simple_loss=0.2442, pruned_loss=0.04718, over 4907.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2431, pruned_loss=0.05262, over 508252.64 frames. ], batch size: 43, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:14:54,306 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-27 02:14:55,336 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.054e+02 1.445e+02 1.737e+02 2.098e+02 4.550e+02, threshold=3.473e+02, percent-clipped=2.0 +2023-03-27 02:15:06,441 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.67 vs. limit=2.0 +2023-03-27 02:15:10,227 INFO [finetune.py:976] (4/7) Epoch 22, batch 200, loss[loss=0.2134, simple_loss=0.2764, pruned_loss=0.07521, over 4828.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2388, pruned_loss=0.05106, over 607216.41 frames. ], batch size: 30, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:15:28,521 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3015, 2.1286, 2.2806, 1.5140, 2.1918, 2.3207, 2.3442, 1.8548], + device='cuda:4'), covar=tensor([0.0545, 0.0638, 0.0671, 0.0823, 0.0672, 0.0708, 0.0579, 0.1082], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0138, 0.0141, 0.0122, 0.0127, 0.0140, 0.0141, 0.0164], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:15:42,764 INFO [finetune.py:976] (4/7) Epoch 22, batch 250, loss[loss=0.1943, simple_loss=0.2668, pruned_loss=0.06089, over 4845.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2421, pruned_loss=0.05215, over 683163.90 frames. ], batch size: 49, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:15:54,243 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4181, 1.3157, 1.2619, 1.4000, 1.6420, 1.5626, 1.3670, 1.2448], + device='cuda:4'), covar=tensor([0.0330, 0.0296, 0.0617, 0.0277, 0.0228, 0.0505, 0.0334, 0.0378], + device='cuda:4'), in_proj_covar=tensor([0.0097, 0.0106, 0.0143, 0.0111, 0.0098, 0.0110, 0.0100, 0.0111], + device='cuda:4'), out_proj_covar=tensor([7.5577e-05, 8.1232e-05, 1.1206e-04, 8.4939e-05, 7.6505e-05, 8.1533e-05, + 7.4500e-05, 8.4962e-05], device='cuda:4') +2023-03-27 02:16:01,906 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.519e+02 1.843e+02 2.180e+02 3.548e+02, threshold=3.686e+02, percent-clipped=1.0 +2023-03-27 02:16:16,408 INFO [finetune.py:976] (4/7) Epoch 22, batch 300, loss[loss=0.1713, simple_loss=0.2506, pruned_loss=0.04599, over 4899.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2458, pruned_loss=0.05279, over 742611.02 frames. ], batch size: 32, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:16:30,491 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3169, 2.2171, 1.6583, 2.2885, 2.0677, 1.8266, 2.5434, 2.3325], + device='cuda:4'), covar=tensor([0.1233, 0.1766, 0.3014, 0.2503, 0.2620, 0.1686, 0.3244, 0.1573], + device='cuda:4'), in_proj_covar=tensor([0.0186, 0.0188, 0.0235, 0.0253, 0.0247, 0.0204, 0.0214, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:16:44,582 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.8313, 3.3007, 3.5104, 3.7568, 3.5822, 3.3012, 3.9033, 1.2729], + device='cuda:4'), covar=tensor([0.0971, 0.1002, 0.1021, 0.1108, 0.1509, 0.1805, 0.0917, 0.5991], + device='cuda:4'), in_proj_covar=tensor([0.0352, 0.0244, 0.0281, 0.0294, 0.0334, 0.0284, 0.0305, 0.0300], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:16:50,473 INFO [finetune.py:976] (4/7) Epoch 22, batch 350, loss[loss=0.2073, simple_loss=0.2762, pruned_loss=0.0692, over 4792.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2464, pruned_loss=0.05238, over 789096.18 frames. ], batch size: 29, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:17:09,297 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.647e+02 1.881e+02 2.280e+02 4.594e+02, threshold=3.762e+02, percent-clipped=3.0 +2023-03-27 02:17:13,097 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-03-27 02:17:18,934 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.63 vs. limit=5.0 +2023-03-27 02:17:19,813 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=120676.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:17:23,372 INFO [finetune.py:976] (4/7) Epoch 22, batch 400, loss[loss=0.1551, simple_loss=0.2345, pruned_loss=0.03787, over 4811.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2474, pruned_loss=0.05206, over 826221.07 frames. ], batch size: 40, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:18:12,315 INFO [finetune.py:976] (4/7) Epoch 22, batch 450, loss[loss=0.1362, simple_loss=0.2052, pruned_loss=0.03354, over 4866.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2474, pruned_loss=0.05232, over 853928.28 frames. ], batch size: 31, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:18:20,747 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=120737.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:18:31,211 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 02:18:39,833 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0800, 2.0641, 2.1331, 1.4463, 2.0855, 2.1202, 2.1599, 1.7255], + device='cuda:4'), covar=tensor([0.0604, 0.0625, 0.0640, 0.0862, 0.0664, 0.0759, 0.0635, 0.1207], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0137, 0.0141, 0.0121, 0.0126, 0.0140, 0.0140, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:18:43,318 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.581e+02 1.866e+02 2.243e+02 3.725e+02, threshold=3.731e+02, percent-clipped=0.0 +2023-03-27 02:18:47,375 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9682, 1.8434, 1.6020, 1.6190, 1.7710, 1.7438, 1.8026, 2.4206], + device='cuda:4'), covar=tensor([0.3799, 0.4118, 0.3126, 0.3720, 0.3876, 0.2433, 0.3367, 0.1866], + device='cuda:4'), in_proj_covar=tensor([0.0290, 0.0264, 0.0234, 0.0278, 0.0256, 0.0225, 0.0254, 0.0235], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:19:07,341 INFO [finetune.py:976] (4/7) Epoch 22, batch 500, loss[loss=0.1298, simple_loss=0.2056, pruned_loss=0.02701, over 4765.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2454, pruned_loss=0.052, over 878905.86 frames. ], batch size: 28, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:19:14,675 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3031, 1.4374, 1.4584, 0.8776, 1.4534, 1.6473, 1.7371, 1.3088], + device='cuda:4'), covar=tensor([0.0870, 0.0568, 0.0535, 0.0531, 0.0477, 0.0577, 0.0306, 0.0764], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0149, 0.0125, 0.0123, 0.0130, 0.0128, 0.0140, 0.0147], + device='cuda:4'), out_proj_covar=tensor([8.9041e-05, 1.0732e-04, 8.9627e-05, 8.6422e-05, 9.1631e-05, 9.1369e-05, + 1.0067e-04, 1.0563e-04], device='cuda:4') +2023-03-27 02:19:20,818 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-27 02:19:40,276 INFO [finetune.py:976] (4/7) Epoch 22, batch 550, loss[loss=0.1692, simple_loss=0.2413, pruned_loss=0.04853, over 4810.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2434, pruned_loss=0.05121, over 896427.31 frames. ], batch size: 51, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:19:49,974 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.8552, 3.3901, 3.5352, 3.7355, 3.6643, 3.3620, 3.9485, 1.2095], + device='cuda:4'), covar=tensor([0.0946, 0.0977, 0.0917, 0.1103, 0.1281, 0.1633, 0.0816, 0.5826], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0241, 0.0277, 0.0290, 0.0330, 0.0281, 0.0301, 0.0297], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:19:58,121 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.063e+02 1.544e+02 1.778e+02 2.172e+02 3.720e+02, threshold=3.555e+02, percent-clipped=0.0 +2023-03-27 02:20:13,119 INFO [finetune.py:976] (4/7) Epoch 22, batch 600, loss[loss=0.165, simple_loss=0.2413, pruned_loss=0.04435, over 4897.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2433, pruned_loss=0.05134, over 909485.80 frames. ], batch size: 32, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:20:46,531 INFO [finetune.py:976] (4/7) Epoch 22, batch 650, loss[loss=0.2072, simple_loss=0.2817, pruned_loss=0.06632, over 4862.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.247, pruned_loss=0.05296, over 919611.73 frames. ], batch size: 44, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:21:04,771 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.580e+02 1.877e+02 2.237e+02 3.344e+02, threshold=3.754e+02, percent-clipped=0.0 +2023-03-27 02:21:20,031 INFO [finetune.py:976] (4/7) Epoch 22, batch 700, loss[loss=0.174, simple_loss=0.2383, pruned_loss=0.05486, over 4792.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2492, pruned_loss=0.05409, over 927824.51 frames. ], batch size: 51, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:21:32,894 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121003.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:21:43,179 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4005, 1.1901, 1.5614, 2.3758, 1.6231, 2.2408, 0.6827, 2.0436], + device='cuda:4'), covar=tensor([0.1843, 0.1755, 0.1434, 0.0937, 0.1029, 0.1218, 0.1972, 0.0652], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0117, 0.0135, 0.0166, 0.0102, 0.0138, 0.0126, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 02:21:53,221 INFO [finetune.py:976] (4/7) Epoch 22, batch 750, loss[loss=0.2106, simple_loss=0.2761, pruned_loss=0.07254, over 4889.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.251, pruned_loss=0.05476, over 933959.04 frames. ], batch size: 32, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:21:53,300 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121032.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:22:09,807 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.017e+02 1.617e+02 1.906e+02 2.410e+02 4.829e+02, threshold=3.812e+02, percent-clipped=5.0 +2023-03-27 02:22:14,821 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=121064.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:22:26,765 INFO [finetune.py:976] (4/7) Epoch 22, batch 800, loss[loss=0.1877, simple_loss=0.2622, pruned_loss=0.05659, over 4804.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2493, pruned_loss=0.05323, over 937736.31 frames. ], batch size: 40, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:23:10,265 INFO [finetune.py:976] (4/7) Epoch 22, batch 850, loss[loss=0.2657, simple_loss=0.3043, pruned_loss=0.1136, over 4150.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2484, pruned_loss=0.05312, over 938973.66 frames. ], batch size: 65, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:23:28,702 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.269e+01 1.509e+02 1.830e+02 2.168e+02 4.982e+02, threshold=3.659e+02, percent-clipped=1.0 +2023-03-27 02:23:37,194 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121164.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:23:49,211 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7172, 1.5833, 1.4996, 1.6015, 1.2419, 3.6073, 1.3361, 1.7356], + device='cuda:4'), covar=tensor([0.3386, 0.2510, 0.2280, 0.2636, 0.1726, 0.0201, 0.2746, 0.1278], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0124, 0.0114, 0.0096, 0.0095, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 02:23:58,238 INFO [finetune.py:976] (4/7) Epoch 22, batch 900, loss[loss=0.2093, simple_loss=0.2628, pruned_loss=0.07793, over 4256.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2455, pruned_loss=0.05224, over 941231.10 frames. ], batch size: 65, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:24:09,539 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1744, 1.2306, 1.3116, 0.7680, 1.3396, 1.5009, 1.5346, 1.2485], + device='cuda:4'), covar=tensor([0.0938, 0.0724, 0.0553, 0.0531, 0.0451, 0.0705, 0.0366, 0.0793], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0150, 0.0125, 0.0123, 0.0130, 0.0129, 0.0141, 0.0147], + device='cuda:4'), out_proj_covar=tensor([8.9136e-05, 1.0803e-04, 8.9656e-05, 8.6466e-05, 9.1737e-05, 9.1740e-05, + 1.0085e-04, 1.0542e-04], device='cuda:4') +2023-03-27 02:24:24,201 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-03-27 02:24:38,329 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=121225.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:24:42,942 INFO [finetune.py:976] (4/7) Epoch 22, batch 950, loss[loss=0.2235, simple_loss=0.2787, pruned_loss=0.08413, over 4835.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2439, pruned_loss=0.052, over 944103.84 frames. ], batch size: 47, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:24:59,302 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.960e+01 1.496e+02 1.804e+02 2.225e+02 4.174e+02, threshold=3.608e+02, percent-clipped=1.0 +2023-03-27 02:25:09,454 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3202, 2.2415, 1.8056, 2.2087, 2.3162, 1.9573, 2.5608, 2.3173], + device='cuda:4'), covar=tensor([0.1289, 0.2012, 0.2928, 0.2392, 0.2476, 0.1658, 0.2402, 0.1638], + device='cuda:4'), in_proj_covar=tensor([0.0186, 0.0187, 0.0234, 0.0253, 0.0246, 0.0203, 0.0213, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:25:16,227 INFO [finetune.py:976] (4/7) Epoch 22, batch 1000, loss[loss=0.1739, simple_loss=0.2502, pruned_loss=0.04883, over 4818.00 frames. ], tot_loss[loss=0.177, simple_loss=0.247, pruned_loss=0.05353, over 946287.91 frames. ], batch size: 41, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:25:33,745 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121311.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 02:25:42,368 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5442, 2.4638, 1.9614, 2.7500, 2.5691, 2.1276, 3.0254, 2.5760], + device='cuda:4'), covar=tensor([0.1350, 0.2083, 0.3175, 0.2528, 0.2542, 0.1766, 0.2939, 0.1814], + device='cuda:4'), in_proj_covar=tensor([0.0186, 0.0188, 0.0235, 0.0254, 0.0247, 0.0204, 0.0214, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:25:49,276 INFO [finetune.py:976] (4/7) Epoch 22, batch 1050, loss[loss=0.1676, simple_loss=0.2358, pruned_loss=0.0497, over 4899.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2487, pruned_loss=0.05352, over 948025.68 frames. ], batch size: 37, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:25:49,380 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=121332.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:25:51,802 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0767, 1.0041, 0.9992, 0.4388, 0.8959, 1.1650, 1.1340, 1.0247], + device='cuda:4'), covar=tensor([0.0802, 0.0551, 0.0538, 0.0488, 0.0547, 0.0637, 0.0351, 0.0641], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0150, 0.0127, 0.0124, 0.0132, 0.0130, 0.0142, 0.0149], + device='cuda:4'), out_proj_covar=tensor([8.9814e-05, 1.0860e-04, 9.0406e-05, 8.7095e-05, 9.2787e-05, 9.2661e-05, + 1.0186e-04, 1.0648e-04], device='cuda:4') +2023-03-27 02:26:05,530 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.073e+02 1.613e+02 2.056e+02 2.633e+02 6.948e+02, threshold=4.113e+02, percent-clipped=5.0 +2023-03-27 02:26:05,612 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121359.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:26:09,274 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.2034, 2.9295, 2.6255, 1.5438, 2.6950, 2.1974, 2.1212, 2.6169], + device='cuda:4'), covar=tensor([0.1228, 0.0724, 0.1654, 0.2106, 0.1537, 0.2223, 0.2281, 0.1099], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0192, 0.0199, 0.0183, 0.0210, 0.0209, 0.0223, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:26:12,284 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0618, 2.6452, 2.5331, 1.2112, 2.8259, 2.1975, 0.8934, 1.8582], + device='cuda:4'), covar=tensor([0.2835, 0.2007, 0.1672, 0.2786, 0.1294, 0.0940, 0.3244, 0.1397], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0177, 0.0158, 0.0129, 0.0159, 0.0122, 0.0147, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 02:26:13,515 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=121372.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 02:26:19,243 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=121380.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:26:20,879 INFO [finetune.py:976] (4/7) Epoch 22, batch 1100, loss[loss=0.2147, simple_loss=0.2824, pruned_loss=0.07348, over 4855.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2485, pruned_loss=0.05346, over 948551.66 frames. ], batch size: 44, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:26:31,161 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6071, 1.4313, 2.2469, 3.4563, 2.3019, 2.4393, 1.2249, 2.8726], + device='cuda:4'), covar=tensor([0.1749, 0.1502, 0.1226, 0.0490, 0.0786, 0.1683, 0.1678, 0.0403], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0118, 0.0136, 0.0167, 0.0102, 0.0139, 0.0127, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 02:26:36,793 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121407.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:26:53,106 INFO [finetune.py:976] (4/7) Epoch 22, batch 1150, loss[loss=0.1831, simple_loss=0.2608, pruned_loss=0.05266, over 4925.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2492, pruned_loss=0.05362, over 950091.82 frames. ], batch size: 33, lr: 3.16e-03, grad_scale: 32.0 +2023-03-27 02:27:02,648 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.1383, 2.9588, 2.5673, 1.5127, 2.6999, 2.0754, 2.0326, 2.5687], + device='cuda:4'), covar=tensor([0.0869, 0.0708, 0.1850, 0.2161, 0.1571, 0.2405, 0.2369, 0.1154], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0192, 0.0199, 0.0183, 0.0210, 0.0209, 0.0224, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:27:10,454 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.055e+01 1.683e+02 1.964e+02 2.424e+02 3.625e+02, threshold=3.928e+02, percent-clipped=0.0 +2023-03-27 02:27:15,981 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=121468.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:27:20,147 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6293, 3.6800, 3.5101, 1.5030, 3.8047, 2.8234, 0.8100, 2.5926], + device='cuda:4'), covar=tensor([0.2294, 0.1912, 0.1524, 0.3518, 0.0999, 0.1016, 0.4440, 0.1374], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0178, 0.0159, 0.0130, 0.0159, 0.0122, 0.0148, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 02:27:25,672 INFO [finetune.py:976] (4/7) Epoch 22, batch 1200, loss[loss=0.1793, simple_loss=0.2487, pruned_loss=0.0549, over 4925.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2476, pruned_loss=0.05243, over 949321.78 frames. ], batch size: 38, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:27:44,330 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1737, 2.2548, 1.9703, 2.3881, 2.1477, 2.2035, 2.1002, 3.0211], + device='cuda:4'), covar=tensor([0.3466, 0.4259, 0.3047, 0.3712, 0.4385, 0.2289, 0.3990, 0.1308], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0262, 0.0232, 0.0277, 0.0255, 0.0225, 0.0254, 0.0235], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:27:49,695 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121520.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:27:57,407 INFO [finetune.py:976] (4/7) Epoch 22, batch 1250, loss[loss=0.1606, simple_loss=0.2276, pruned_loss=0.04675, over 4903.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2447, pruned_loss=0.05122, over 952572.03 frames. ], batch size: 35, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:28:26,677 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.015e+02 1.487e+02 1.790e+02 2.203e+02 4.512e+02, threshold=3.581e+02, percent-clipped=1.0 +2023-03-27 02:28:40,602 INFO [finetune.py:976] (4/7) Epoch 22, batch 1300, loss[loss=0.2108, simple_loss=0.2676, pruned_loss=0.07698, over 4867.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2424, pruned_loss=0.05097, over 952728.01 frames. ], batch size: 34, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:29:08,208 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0472, 1.9929, 1.7276, 1.7511, 1.8317, 1.8506, 1.9070, 2.5895], + device='cuda:4'), covar=tensor([0.3579, 0.3404, 0.2963, 0.3459, 0.3723, 0.2311, 0.3234, 0.1553], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0263, 0.0233, 0.0278, 0.0256, 0.0225, 0.0254, 0.0235], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:29:39,001 INFO [finetune.py:976] (4/7) Epoch 22, batch 1350, loss[loss=0.1803, simple_loss=0.2344, pruned_loss=0.06306, over 4226.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2426, pruned_loss=0.05094, over 953022.73 frames. ], batch size: 18, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:30:02,099 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.135e+02 1.546e+02 1.913e+02 2.289e+02 6.231e+02, threshold=3.826e+02, percent-clipped=1.0 +2023-03-27 02:30:02,195 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=121659.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:30:07,016 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121667.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 02:30:11,397 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 02:30:16,103 INFO [finetune.py:976] (4/7) Epoch 22, batch 1400, loss[loss=0.1913, simple_loss=0.2805, pruned_loss=0.05111, over 4816.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2469, pruned_loss=0.05235, over 952807.86 frames. ], batch size: 51, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:30:16,842 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0116, 1.8807, 1.6679, 1.9930, 2.5091, 2.0826, 1.9917, 1.5659], + device='cuda:4'), covar=tensor([0.2172, 0.1863, 0.1863, 0.1595, 0.1773, 0.1164, 0.2103, 0.1794], + device='cuda:4'), in_proj_covar=tensor([0.0245, 0.0211, 0.0215, 0.0196, 0.0243, 0.0189, 0.0218, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:30:34,337 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=121707.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:30:49,372 INFO [finetune.py:976] (4/7) Epoch 22, batch 1450, loss[loss=0.1614, simple_loss=0.2444, pruned_loss=0.03915, over 4863.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2491, pruned_loss=0.053, over 953831.05 frames. ], batch size: 31, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:31:08,592 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.585e+02 1.838e+02 2.176e+02 4.319e+02, threshold=3.676e+02, percent-clipped=1.0 +2023-03-27 02:31:11,102 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=121763.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:31:22,462 INFO [finetune.py:976] (4/7) Epoch 22, batch 1500, loss[loss=0.1747, simple_loss=0.2448, pruned_loss=0.05227, over 4787.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2508, pruned_loss=0.05372, over 954479.97 frames. ], batch size: 29, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:31:44,128 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-27 02:31:49,160 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=121820.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:31:56,351 INFO [finetune.py:976] (4/7) Epoch 22, batch 1550, loss[loss=0.2065, simple_loss=0.2699, pruned_loss=0.07155, over 4923.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2496, pruned_loss=0.05322, over 954782.97 frames. ], batch size: 33, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:32:15,624 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.581e+02 1.856e+02 2.152e+02 3.350e+02, threshold=3.712e+02, percent-clipped=0.0 +2023-03-27 02:32:21,174 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=121868.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:32:29,537 INFO [finetune.py:976] (4/7) Epoch 22, batch 1600, loss[loss=0.1439, simple_loss=0.218, pruned_loss=0.03485, over 4819.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2469, pruned_loss=0.05253, over 953274.98 frames. ], batch size: 30, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:32:51,280 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4364, 2.2611, 1.9248, 2.2151, 2.3794, 2.1136, 2.6030, 2.3578], + device='cuda:4'), covar=tensor([0.1325, 0.1920, 0.2880, 0.2253, 0.2322, 0.1712, 0.2509, 0.1818], + device='cuda:4'), in_proj_covar=tensor([0.0187, 0.0188, 0.0235, 0.0254, 0.0248, 0.0204, 0.0215, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:33:02,675 INFO [finetune.py:976] (4/7) Epoch 22, batch 1650, loss[loss=0.1485, simple_loss=0.228, pruned_loss=0.03457, over 4933.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2436, pruned_loss=0.05134, over 953877.40 frames. ], batch size: 33, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:33:13,739 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=121950.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:33:22,467 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.553e+02 1.837e+02 2.107e+02 3.976e+02, threshold=3.675e+02, percent-clipped=1.0 +2023-03-27 02:33:33,823 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=121967.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 02:33:46,460 INFO [finetune.py:976] (4/7) Epoch 22, batch 1700, loss[loss=0.1945, simple_loss=0.2456, pruned_loss=0.07171, over 4062.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2416, pruned_loss=0.05095, over 953985.73 frames. ], batch size: 65, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:34:13,896 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122011.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:34:16,846 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122015.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 02:34:36,865 INFO [finetune.py:976] (4/7) Epoch 22, batch 1750, loss[loss=0.1604, simple_loss=0.2247, pruned_loss=0.04802, over 4798.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2433, pruned_loss=0.05143, over 955203.25 frames. ], batch size: 25, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:35:06,505 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.629e+02 1.889e+02 2.189e+02 5.095e+02, threshold=3.778e+02, percent-clipped=2.0 +2023-03-27 02:35:10,493 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=122063.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:35:22,846 INFO [finetune.py:976] (4/7) Epoch 22, batch 1800, loss[loss=0.1661, simple_loss=0.2422, pruned_loss=0.04497, over 4826.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2464, pruned_loss=0.05163, over 956148.47 frames. ], batch size: 30, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:35:41,159 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122111.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:35:46,330 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122117.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:35:53,547 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-03-27 02:35:56,271 INFO [finetune.py:976] (4/7) Epoch 22, batch 1850, loss[loss=0.1677, simple_loss=0.2372, pruned_loss=0.04911, over 4769.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2476, pruned_loss=0.05253, over 953817.64 frames. ], batch size: 28, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:36:12,755 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.044e+01 1.577e+02 1.909e+02 2.256e+02 5.766e+02, threshold=3.818e+02, percent-clipped=1.0 +2023-03-27 02:36:27,298 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122178.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 02:36:29,613 INFO [finetune.py:976] (4/7) Epoch 22, batch 1900, loss[loss=0.1705, simple_loss=0.2508, pruned_loss=0.04516, over 4843.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2485, pruned_loss=0.05249, over 955190.15 frames. ], batch size: 47, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:36:37,625 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122195.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:37:03,560 INFO [finetune.py:976] (4/7) Epoch 22, batch 1950, loss[loss=0.1763, simple_loss=0.2399, pruned_loss=0.05631, over 4819.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2477, pruned_loss=0.05209, over 954867.54 frames. ], batch size: 33, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:37:05,022 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-27 02:37:18,290 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122256.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:37:19,908 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-27 02:37:19,980 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.844e+01 1.567e+02 1.786e+02 2.225e+02 4.203e+02, threshold=3.573e+02, percent-clipped=2.0 +2023-03-27 02:37:36,887 INFO [finetune.py:976] (4/7) Epoch 22, batch 2000, loss[loss=0.1808, simple_loss=0.2467, pruned_loss=0.05748, over 4827.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2454, pruned_loss=0.05143, over 955130.17 frames. ], batch size: 40, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:37:51,524 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=122306.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:37:59,436 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122319.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:38:02,159 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7365, 1.5687, 1.9332, 1.3847, 1.7807, 1.9105, 1.5308, 2.1033], + device='cuda:4'), covar=tensor([0.1101, 0.2098, 0.1225, 0.1540, 0.0851, 0.1232, 0.3085, 0.0750], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0207, 0.0193, 0.0190, 0.0175, 0.0213, 0.0218, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:38:06,109 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7292, 1.5818, 1.3993, 1.7258, 2.0826, 1.9435, 1.6761, 1.4966], + device='cuda:4'), covar=tensor([0.0381, 0.0385, 0.0667, 0.0314, 0.0251, 0.0583, 0.0318, 0.0484], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0106, 0.0143, 0.0111, 0.0099, 0.0111, 0.0100, 0.0112], + device='cuda:4'), out_proj_covar=tensor([7.6497e-05, 8.1717e-05, 1.1197e-04, 8.4925e-05, 7.6971e-05, 8.2280e-05, + 7.4325e-05, 8.5579e-05], device='cuda:4') +2023-03-27 02:38:10,019 INFO [finetune.py:976] (4/7) Epoch 22, batch 2050, loss[loss=0.1383, simple_loss=0.204, pruned_loss=0.03627, over 4320.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2424, pruned_loss=0.05022, over 954905.19 frames. ], batch size: 18, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:38:26,816 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.990e+01 1.410e+02 1.787e+02 2.088e+02 3.673e+02, threshold=3.574e+02, percent-clipped=1.0 +2023-03-27 02:38:43,030 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122380.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:38:44,602 INFO [finetune.py:976] (4/7) Epoch 22, batch 2100, loss[loss=0.2175, simple_loss=0.2807, pruned_loss=0.07716, over 4868.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2431, pruned_loss=0.05063, over 954055.14 frames. ], batch size: 31, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:39:02,470 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0469, 1.9723, 1.6664, 2.0007, 2.5339, 2.1159, 2.0730, 1.6204], + device='cuda:4'), covar=tensor([0.2355, 0.2011, 0.1992, 0.1702, 0.1919, 0.1214, 0.2043, 0.1958], + device='cuda:4'), in_proj_covar=tensor([0.0247, 0.0212, 0.0215, 0.0198, 0.0245, 0.0191, 0.0219, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:39:23,319 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-27 02:39:28,191 INFO [finetune.py:976] (4/7) Epoch 22, batch 2150, loss[loss=0.1695, simple_loss=0.2385, pruned_loss=0.05025, over 4877.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.246, pruned_loss=0.05157, over 953809.46 frames. ], batch size: 34, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:40:03,768 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.048e+02 1.591e+02 1.910e+02 2.352e+02 5.051e+02, threshold=3.819e+02, percent-clipped=3.0 +2023-03-27 02:40:06,808 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.83 vs. limit=5.0 +2023-03-27 02:40:13,685 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122468.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:40:16,651 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=122473.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 02:40:26,432 INFO [finetune.py:976] (4/7) Epoch 22, batch 2200, loss[loss=0.2284, simple_loss=0.2888, pruned_loss=0.08402, over 4813.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2484, pruned_loss=0.0524, over 952710.27 frames. ], batch size: 38, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:40:27,759 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.0301, 2.7376, 2.4740, 1.3835, 2.6195, 2.1297, 2.0078, 2.3504], + device='cuda:4'), covar=tensor([0.1235, 0.0755, 0.2120, 0.2321, 0.1664, 0.2594, 0.2466, 0.1347], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0193, 0.0200, 0.0183, 0.0210, 0.0208, 0.0224, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:40:39,107 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-27 02:40:56,497 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-27 02:40:56,787 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122529.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:40:58,945 INFO [finetune.py:976] (4/7) Epoch 22, batch 2250, loss[loss=0.1489, simple_loss=0.2306, pruned_loss=0.03362, over 4777.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2502, pruned_loss=0.05303, over 953417.88 frames. ], batch size: 26, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:41:12,977 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=122551.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:41:17,777 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.977e+01 1.468e+02 1.830e+02 2.095e+02 3.153e+02, threshold=3.659e+02, percent-clipped=0.0 +2023-03-27 02:41:22,682 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4636, 1.4110, 1.5798, 0.7782, 1.5210, 1.5096, 1.5425, 1.3278], + device='cuda:4'), covar=tensor([0.0631, 0.0797, 0.0693, 0.0970, 0.0890, 0.0778, 0.0692, 0.1296], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0137, 0.0141, 0.0121, 0.0126, 0.0139, 0.0140, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:41:26,563 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-27 02:41:26,990 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4710, 2.3378, 2.0397, 1.1336, 2.1742, 1.9100, 1.7422, 2.0310], + device='cuda:4'), covar=tensor([0.0839, 0.0765, 0.1706, 0.2000, 0.1486, 0.2104, 0.2313, 0.1149], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0192, 0.0199, 0.0182, 0.0209, 0.0207, 0.0223, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:41:31,701 INFO [finetune.py:976] (4/7) Epoch 22, batch 2300, loss[loss=0.1392, simple_loss=0.2238, pruned_loss=0.02733, over 4760.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2497, pruned_loss=0.05219, over 952470.21 frames. ], batch size: 28, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:41:49,495 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=122606.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:42:05,215 INFO [finetune.py:976] (4/7) Epoch 22, batch 2350, loss[loss=0.1522, simple_loss=0.2186, pruned_loss=0.04293, over 4891.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2479, pruned_loss=0.0518, over 954164.06 frames. ], batch size: 32, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:42:09,426 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8561, 4.1338, 3.8586, 2.0321, 4.2149, 3.1953, 1.0719, 2.8437], + device='cuda:4'), covar=tensor([0.2285, 0.1837, 0.1609, 0.3267, 0.1024, 0.0995, 0.4547, 0.1380], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0178, 0.0159, 0.0130, 0.0161, 0.0123, 0.0148, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 02:42:21,477 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122654.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:42:24,443 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.047e+02 1.417e+02 1.625e+02 2.018e+02 3.172e+02, threshold=3.250e+02, percent-clipped=0.0 +2023-03-27 02:42:34,178 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=122675.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:42:38,345 INFO [finetune.py:976] (4/7) Epoch 22, batch 2400, loss[loss=0.1543, simple_loss=0.224, pruned_loss=0.0423, over 4845.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2456, pruned_loss=0.05136, over 955756.78 frames. ], batch size: 44, lr: 3.15e-03, grad_scale: 64.0 +2023-03-27 02:42:50,687 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4376, 1.3432, 1.3900, 0.7646, 1.5040, 1.4507, 1.4585, 1.3081], + device='cuda:4'), covar=tensor([0.0582, 0.0784, 0.0746, 0.0936, 0.0876, 0.0757, 0.0692, 0.1218], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0136, 0.0139, 0.0120, 0.0125, 0.0138, 0.0139, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:43:11,474 INFO [finetune.py:976] (4/7) Epoch 22, batch 2450, loss[loss=0.195, simple_loss=0.2558, pruned_loss=0.06715, over 4804.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2427, pruned_loss=0.05055, over 954679.47 frames. ], batch size: 51, lr: 3.14e-03, grad_scale: 64.0 +2023-03-27 02:43:31,105 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.588e+02 1.841e+02 2.130e+02 2.968e+02, threshold=3.682e+02, percent-clipped=0.0 +2023-03-27 02:43:37,342 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8117, 2.6060, 2.4265, 3.1123, 2.8865, 2.4837, 3.2864, 2.8407], + device='cuda:4'), covar=tensor([0.1272, 0.2259, 0.2852, 0.2188, 0.2368, 0.1581, 0.2698, 0.1691], + device='cuda:4'), in_proj_covar=tensor([0.0188, 0.0190, 0.0237, 0.0256, 0.0249, 0.0206, 0.0215, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:43:39,661 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=122773.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 02:43:45,021 INFO [finetune.py:976] (4/7) Epoch 22, batch 2500, loss[loss=0.2638, simple_loss=0.3157, pruned_loss=0.106, over 4870.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.244, pruned_loss=0.05157, over 955086.34 frames. ], batch size: 31, lr: 3.14e-03, grad_scale: 64.0 +2023-03-27 02:43:46,412 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.95 vs. limit=5.0 +2023-03-27 02:44:21,547 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122821.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:44:23,365 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=122824.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:44:28,112 INFO [finetune.py:976] (4/7) Epoch 22, batch 2550, loss[loss=0.1926, simple_loss=0.2715, pruned_loss=0.05682, over 4741.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2465, pruned_loss=0.05185, over 956100.70 frames. ], batch size: 54, lr: 3.14e-03, grad_scale: 64.0 +2023-03-27 02:44:42,581 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=122851.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:44:53,792 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.164e+02 1.566e+02 1.889e+02 2.331e+02 3.878e+02, threshold=3.777e+02, percent-clipped=1.0 +2023-03-27 02:45:07,304 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=122872.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:45:22,082 INFO [finetune.py:976] (4/7) Epoch 22, batch 2600, loss[loss=0.1436, simple_loss=0.2168, pruned_loss=0.03516, over 4749.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2481, pruned_loss=0.05199, over 955887.81 frames. ], batch size: 28, lr: 3.14e-03, grad_scale: 64.0 +2023-03-27 02:45:40,401 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=122899.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:46:02,677 INFO [finetune.py:976] (4/7) Epoch 22, batch 2650, loss[loss=0.1818, simple_loss=0.262, pruned_loss=0.05079, over 4826.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.248, pruned_loss=0.0518, over 953029.04 frames. ], batch size: 30, lr: 3.14e-03, grad_scale: 64.0 +2023-03-27 02:46:03,428 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=122933.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:46:19,589 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.898e+01 1.493e+02 1.785e+02 2.135e+02 3.458e+02, threshold=3.570e+02, percent-clipped=0.0 +2023-03-27 02:46:31,738 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=122975.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:46:35,904 INFO [finetune.py:976] (4/7) Epoch 22, batch 2700, loss[loss=0.1668, simple_loss=0.2377, pruned_loss=0.04798, over 4859.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2459, pruned_loss=0.05065, over 951937.23 frames. ], batch size: 31, lr: 3.14e-03, grad_scale: 64.0 +2023-03-27 02:47:04,318 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=123023.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:47:09,737 INFO [finetune.py:976] (4/7) Epoch 22, batch 2750, loss[loss=0.1571, simple_loss=0.2295, pruned_loss=0.04235, over 4847.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2445, pruned_loss=0.05052, over 951668.25 frames. ], batch size: 44, lr: 3.14e-03, grad_scale: 64.0 +2023-03-27 02:47:16,474 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.3463, 4.7059, 4.8780, 5.2386, 5.0229, 4.7216, 5.4586, 1.7722], + device='cuda:4'), covar=tensor([0.0633, 0.0726, 0.0807, 0.0657, 0.1334, 0.1650, 0.0547, 0.5647], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0244, 0.0280, 0.0290, 0.0335, 0.0284, 0.0305, 0.0299], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:47:24,424 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8138, 1.3513, 0.8432, 1.7286, 2.2242, 1.3489, 1.6731, 1.7087], + device='cuda:4'), covar=tensor([0.1302, 0.1913, 0.1881, 0.1092, 0.1712, 0.1916, 0.1288, 0.1808], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0091, 0.0119, 0.0094, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 02:47:26,635 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.465e+02 1.704e+02 2.045e+02 3.535e+02, threshold=3.409e+02, percent-clipped=0.0 +2023-03-27 02:47:42,977 INFO [finetune.py:976] (4/7) Epoch 22, batch 2800, loss[loss=0.1601, simple_loss=0.2408, pruned_loss=0.03967, over 4792.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2425, pruned_loss=0.05048, over 952978.31 frames. ], batch size: 26, lr: 3.14e-03, grad_scale: 64.0 +2023-03-27 02:47:48,486 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3011, 1.3982, 1.6588, 1.5450, 1.6055, 2.9817, 1.4076, 1.5460], + device='cuda:4'), covar=tensor([0.0974, 0.1760, 0.1034, 0.0935, 0.1538, 0.0313, 0.1348, 0.1665], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0092, 0.0081, 0.0086, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 02:47:56,785 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-27 02:48:11,378 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=123124.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:48:12,001 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1438, 2.2148, 1.8320, 2.2135, 2.1175, 2.0584, 2.0784, 2.9848], + device='cuda:4'), covar=tensor([0.3829, 0.4790, 0.3496, 0.4840, 0.4454, 0.2602, 0.4572, 0.1565], + device='cuda:4'), in_proj_covar=tensor([0.0287, 0.0262, 0.0233, 0.0276, 0.0254, 0.0225, 0.0253, 0.0234], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:48:16,590 INFO [finetune.py:976] (4/7) Epoch 22, batch 2850, loss[loss=0.2232, simple_loss=0.2881, pruned_loss=0.07914, over 4811.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2411, pruned_loss=0.05053, over 952741.26 frames. ], batch size: 39, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:48:33,489 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.606e+02 1.919e+02 2.375e+02 6.875e+02, threshold=3.839e+02, percent-clipped=7.0 +2023-03-27 02:48:42,229 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=123172.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:48:46,870 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4452, 1.3443, 1.3397, 1.4424, 1.0155, 2.9381, 1.0977, 1.4812], + device='cuda:4'), covar=tensor([0.3282, 0.2549, 0.2242, 0.2465, 0.1884, 0.0257, 0.2770, 0.1287], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0124, 0.0114, 0.0096, 0.0094, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 02:48:49,677 INFO [finetune.py:976] (4/7) Epoch 22, batch 2900, loss[loss=0.1957, simple_loss=0.2715, pruned_loss=0.06, over 4787.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2449, pruned_loss=0.05227, over 953790.77 frames. ], batch size: 29, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:49:22,652 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=123228.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:49:25,027 INFO [finetune.py:976] (4/7) Epoch 22, batch 2950, loss[loss=0.2165, simple_loss=0.2907, pruned_loss=0.07117, over 4941.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2485, pruned_loss=0.05329, over 953400.93 frames. ], batch size: 39, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:49:42,391 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.515e+01 1.556e+02 1.822e+02 2.269e+02 3.192e+02, threshold=3.643e+02, percent-clipped=0.0 +2023-03-27 02:49:59,813 INFO [finetune.py:976] (4/7) Epoch 22, batch 3000, loss[loss=0.2096, simple_loss=0.2809, pruned_loss=0.06913, over 4889.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2502, pruned_loss=0.05368, over 954374.45 frames. ], batch size: 35, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:49:59,813 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 02:50:18,027 INFO [finetune.py:1010] (4/7) Epoch 22, validation: loss=0.1575, simple_loss=0.2256, pruned_loss=0.04471, over 2265189.00 frames. +2023-03-27 02:50:18,028 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-27 02:50:27,870 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5803, 2.4210, 1.9385, 2.6675, 2.4962, 2.0829, 3.0327, 2.5344], + device='cuda:4'), covar=tensor([0.1296, 0.2166, 0.3152, 0.2512, 0.2381, 0.1676, 0.2848, 0.1790], + device='cuda:4'), in_proj_covar=tensor([0.0188, 0.0189, 0.0236, 0.0255, 0.0249, 0.0205, 0.0214, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:50:47,930 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=123314.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:51:07,175 INFO [finetune.py:976] (4/7) Epoch 22, batch 3050, loss[loss=0.1918, simple_loss=0.259, pruned_loss=0.06232, over 4822.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2504, pruned_loss=0.05348, over 954403.43 frames. ], batch size: 33, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:51:27,201 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.922e+01 1.647e+02 1.960e+02 2.377e+02 4.726e+02, threshold=3.920e+02, percent-clipped=6.0 +2023-03-27 02:51:36,582 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=123375.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:51:41,107 INFO [finetune.py:976] (4/7) Epoch 22, batch 3100, loss[loss=0.1345, simple_loss=0.2078, pruned_loss=0.03064, over 4726.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2481, pruned_loss=0.05261, over 954245.34 frames. ], batch size: 23, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:52:09,231 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0882, 1.2452, 1.4347, 1.3579, 1.3895, 2.4098, 1.2211, 1.4142], + device='cuda:4'), covar=tensor([0.0984, 0.1915, 0.1069, 0.0924, 0.1607, 0.0400, 0.1543, 0.1756], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0092, 0.0081, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 02:52:12,235 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=123428.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:52:14,578 INFO [finetune.py:976] (4/7) Epoch 22, batch 3150, loss[loss=0.1604, simple_loss=0.2251, pruned_loss=0.0478, over 4837.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2459, pruned_loss=0.05234, over 954509.39 frames. ], batch size: 47, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:52:18,242 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2247, 3.6589, 3.8340, 4.0575, 4.0174, 3.7953, 4.3202, 1.3891], + device='cuda:4'), covar=tensor([0.0747, 0.0816, 0.0836, 0.0929, 0.1148, 0.1402, 0.0684, 0.5397], + device='cuda:4'), in_proj_covar=tensor([0.0353, 0.0246, 0.0282, 0.0292, 0.0337, 0.0287, 0.0307, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:52:34,405 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.134e+02 1.485e+02 1.808e+02 2.093e+02 3.688e+02, threshold=3.617e+02, percent-clipped=0.0 +2023-03-27 02:52:47,870 INFO [finetune.py:976] (4/7) Epoch 22, batch 3200, loss[loss=0.1361, simple_loss=0.2188, pruned_loss=0.02671, over 4773.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2431, pruned_loss=0.05154, over 957459.77 frames. ], batch size: 28, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:52:48,009 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2165, 2.1849, 1.7249, 2.1521, 2.1703, 1.9299, 2.5205, 2.2476], + device='cuda:4'), covar=tensor([0.1252, 0.2011, 0.2860, 0.2474, 0.2289, 0.1500, 0.3010, 0.1571], + device='cuda:4'), in_proj_covar=tensor([0.0187, 0.0188, 0.0235, 0.0254, 0.0248, 0.0204, 0.0214, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:52:48,637 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3848, 1.3791, 1.3357, 0.7773, 1.3617, 1.5942, 1.6254, 1.2787], + device='cuda:4'), covar=tensor([0.0918, 0.0620, 0.0523, 0.0502, 0.0501, 0.0540, 0.0311, 0.0673], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0149, 0.0127, 0.0123, 0.0132, 0.0130, 0.0141, 0.0148], + device='cuda:4'), out_proj_covar=tensor([8.9707e-05, 1.0787e-04, 9.1042e-05, 8.6330e-05, 9.2625e-05, 9.2797e-05, + 1.0106e-04, 1.0625e-04], device='cuda:4') +2023-03-27 02:52:52,353 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=123489.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:53:18,978 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=123528.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:53:21,318 INFO [finetune.py:976] (4/7) Epoch 22, batch 3250, loss[loss=0.203, simple_loss=0.2606, pruned_loss=0.07272, over 4797.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2441, pruned_loss=0.05242, over 956475.48 frames. ], batch size: 29, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:53:41,212 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.285e+01 1.537e+02 1.740e+02 2.121e+02 3.629e+02, threshold=3.481e+02, percent-clipped=1.0 +2023-03-27 02:53:51,148 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=123576.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:53:54,732 INFO [finetune.py:976] (4/7) Epoch 22, batch 3300, loss[loss=0.1816, simple_loss=0.2632, pruned_loss=0.04998, over 4817.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2473, pruned_loss=0.05269, over 957322.91 frames. ], batch size: 39, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:54:28,105 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 02:54:28,300 INFO [finetune.py:976] (4/7) Epoch 22, batch 3350, loss[loss=0.1746, simple_loss=0.2539, pruned_loss=0.04768, over 4839.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2488, pruned_loss=0.05303, over 956976.44 frames. ], batch size: 47, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:54:35,088 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7187, 1.6300, 2.0309, 1.3830, 1.7841, 1.9946, 1.5595, 2.1813], + device='cuda:4'), covar=tensor([0.1424, 0.2186, 0.1523, 0.1928, 0.1000, 0.1508, 0.2878, 0.0891], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0207, 0.0191, 0.0189, 0.0175, 0.0213, 0.0217, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:54:47,671 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.106e+02 1.542e+02 1.809e+02 2.055e+02 5.285e+02, threshold=3.617e+02, percent-clipped=2.0 +2023-03-27 02:54:54,269 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=123670.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:54:56,124 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4913, 2.3264, 1.9631, 1.0466, 2.0797, 1.8321, 1.6205, 2.1950], + device='cuda:4'), covar=tensor([0.0779, 0.0781, 0.1382, 0.1959, 0.1258, 0.2457, 0.2450, 0.0872], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0191, 0.0197, 0.0182, 0.0210, 0.0206, 0.0222, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:55:01,514 INFO [finetune.py:976] (4/7) Epoch 22, batch 3400, loss[loss=0.1613, simple_loss=0.2337, pruned_loss=0.04442, over 4912.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2496, pruned_loss=0.05315, over 957250.88 frames. ], batch size: 33, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:55:05,293 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3815, 2.9065, 2.7640, 1.2185, 3.0050, 2.2997, 0.7777, 1.9766], + device='cuda:4'), covar=tensor([0.2407, 0.2155, 0.1886, 0.3460, 0.1540, 0.1109, 0.4078, 0.1520], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0176, 0.0157, 0.0129, 0.0160, 0.0122, 0.0147, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 02:55:54,857 INFO [finetune.py:976] (4/7) Epoch 22, batch 3450, loss[loss=0.1673, simple_loss=0.2409, pruned_loss=0.04682, over 4919.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2501, pruned_loss=0.05361, over 956692.53 frames. ], batch size: 38, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:56:26,817 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.145e+02 1.545e+02 1.898e+02 2.347e+02 3.548e+02, threshold=3.797e+02, percent-clipped=0.0 +2023-03-27 02:56:45,230 INFO [finetune.py:976] (4/7) Epoch 22, batch 3500, loss[loss=0.1649, simple_loss=0.2299, pruned_loss=0.04994, over 4833.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.247, pruned_loss=0.05259, over 955893.30 frames. ], batch size: 30, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:56:46,495 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=123784.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:57:01,657 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8878, 2.5867, 3.1137, 2.0037, 2.6243, 3.1422, 2.3668, 3.2658], + device='cuda:4'), covar=tensor([0.1198, 0.1768, 0.1547, 0.2188, 0.1068, 0.1319, 0.2293, 0.0775], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0206, 0.0191, 0.0189, 0.0174, 0.0213, 0.0216, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 02:57:18,475 INFO [finetune.py:976] (4/7) Epoch 22, batch 3550, loss[loss=0.1808, simple_loss=0.2441, pruned_loss=0.05873, over 4816.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2446, pruned_loss=0.05193, over 958485.48 frames. ], batch size: 41, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:57:27,307 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.56 vs. limit=5.0 +2023-03-27 02:57:36,085 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.884e+01 1.429e+02 1.766e+02 2.143e+02 3.754e+02, threshold=3.531e+02, percent-clipped=0.0 +2023-03-27 02:57:45,445 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=123872.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:57:52,357 INFO [finetune.py:976] (4/7) Epoch 22, batch 3600, loss[loss=0.1574, simple_loss=0.2275, pruned_loss=0.04367, over 4758.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2411, pruned_loss=0.05042, over 956045.94 frames. ], batch size: 26, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:58:00,483 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.23 vs. limit=5.0 +2023-03-27 02:58:25,719 INFO [finetune.py:976] (4/7) Epoch 22, batch 3650, loss[loss=0.198, simple_loss=0.2559, pruned_loss=0.07001, over 4893.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2424, pruned_loss=0.05108, over 956251.50 frames. ], batch size: 35, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:58:26,482 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=123933.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:58:35,812 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.90 vs. limit=5.0 +2023-03-27 02:58:43,173 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.106e+01 1.598e+02 1.925e+02 2.525e+02 5.508e+02, threshold=3.851e+02, percent-clipped=5.0 +2023-03-27 02:58:49,863 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=123970.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:58:56,167 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8237, 1.8142, 1.6302, 1.8403, 1.7513, 4.4521, 1.7465, 2.1673], + device='cuda:4'), covar=tensor([0.3156, 0.2533, 0.2196, 0.2271, 0.1483, 0.0113, 0.2506, 0.1169], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0116, 0.0122, 0.0124, 0.0114, 0.0096, 0.0095, 0.0096], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 02:58:59,577 INFO [finetune.py:976] (4/7) Epoch 22, batch 3700, loss[loss=0.1615, simple_loss=0.237, pruned_loss=0.04303, over 4743.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2458, pruned_loss=0.05204, over 955718.66 frames. ], batch size: 27, lr: 3.14e-03, grad_scale: 32.0 +2023-03-27 02:59:23,318 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124018.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:59:34,050 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5151, 1.0962, 0.7888, 1.3406, 1.9139, 0.6868, 1.2955, 1.3812], + device='cuda:4'), covar=tensor([0.1401, 0.2004, 0.1659, 0.1141, 0.1895, 0.1870, 0.1402, 0.1867], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0111, 0.0092, 0.0120, 0.0094, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 02:59:34,574 INFO [finetune.py:976] (4/7) Epoch 22, batch 3750, loss[loss=0.2043, simple_loss=0.2777, pruned_loss=0.06547, over 4817.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2468, pruned_loss=0.05178, over 955471.54 frames. ], batch size: 39, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 02:59:37,121 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124036.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:59:51,700 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.170e+02 1.541e+02 1.786e+02 2.095e+02 2.976e+02, threshold=3.572e+02, percent-clipped=0.0 +2023-03-27 02:59:53,039 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124062.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 02:59:53,933 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-03-27 03:00:06,947 INFO [finetune.py:976] (4/7) Epoch 22, batch 3800, loss[loss=0.1778, simple_loss=0.2568, pruned_loss=0.04935, over 4815.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2473, pruned_loss=0.0515, over 957266.21 frames. ], batch size: 38, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:00:08,710 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124084.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:00:17,181 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124097.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:00:39,618 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124123.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:00:51,383 INFO [finetune.py:976] (4/7) Epoch 22, batch 3850, loss[loss=0.1417, simple_loss=0.2225, pruned_loss=0.03048, over 4829.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2453, pruned_loss=0.05065, over 954676.61 frames. ], batch size: 30, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:00:51,454 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124132.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:00:54,261 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6548, 1.2196, 0.9617, 1.6490, 2.1370, 1.5487, 1.5336, 1.7602], + device='cuda:4'), covar=tensor([0.1413, 0.1987, 0.1915, 0.1086, 0.1751, 0.1944, 0.1295, 0.1664], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0094, 0.0111, 0.0092, 0.0120, 0.0094, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 03:01:02,211 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.77 vs. limit=2.0 +2023-03-27 03:01:04,467 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7183, 1.1594, 0.7270, 1.6177, 2.1854, 1.2688, 1.5391, 1.7303], + device='cuda:4'), covar=tensor([0.1386, 0.2118, 0.1935, 0.1093, 0.1825, 0.1741, 0.1377, 0.1737], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0094, 0.0111, 0.0092, 0.0120, 0.0094, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 03:01:17,675 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 03:01:20,787 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.486e+02 1.746e+02 2.060e+02 3.550e+02, threshold=3.491e+02, percent-clipped=0.0 +2023-03-27 03:01:21,937 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-27 03:01:46,863 INFO [finetune.py:976] (4/7) Epoch 22, batch 3900, loss[loss=0.1676, simple_loss=0.2283, pruned_loss=0.05343, over 4872.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2439, pruned_loss=0.05076, over 956344.84 frames. ], batch size: 34, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:01:50,418 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7763, 1.9046, 1.8010, 2.1899, 2.2429, 2.2254, 1.7559, 1.5306], + device='cuda:4'), covar=tensor([0.2085, 0.1726, 0.1549, 0.1313, 0.1740, 0.1036, 0.2127, 0.1749], + device='cuda:4'), in_proj_covar=tensor([0.0244, 0.0209, 0.0212, 0.0195, 0.0242, 0.0189, 0.0215, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:02:17,916 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124228.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:02:20,288 INFO [finetune.py:976] (4/7) Epoch 22, batch 3950, loss[loss=0.1234, simple_loss=0.1939, pruned_loss=0.02641, over 4830.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2415, pruned_loss=0.05023, over 957200.20 frames. ], batch size: 25, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:02:37,027 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0975, 1.3239, 1.5448, 1.2617, 1.5043, 2.4217, 1.3272, 1.5081], + device='cuda:4'), covar=tensor([0.0973, 0.1783, 0.0920, 0.0956, 0.1571, 0.0379, 0.1453, 0.1722], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0092, 0.0081, 0.0086, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 03:02:39,959 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.148e+02 1.559e+02 1.897e+02 2.284e+02 3.853e+02, threshold=3.794e+02, percent-clipped=3.0 +2023-03-27 03:02:47,308 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124272.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:02:49,751 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6478, 1.5670, 1.3820, 1.4795, 1.9439, 1.8846, 1.5681, 1.3993], + device='cuda:4'), covar=tensor([0.0356, 0.0339, 0.0670, 0.0364, 0.0222, 0.0417, 0.0356, 0.0455], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0106, 0.0143, 0.0111, 0.0098, 0.0111, 0.0100, 0.0111], + device='cuda:4'), out_proj_covar=tensor([7.6354e-05, 8.1654e-05, 1.1187e-04, 8.5148e-05, 7.6416e-05, 8.2074e-05, + 7.4843e-05, 8.4787e-05], device='cuda:4') +2023-03-27 03:02:53,667 INFO [finetune.py:976] (4/7) Epoch 22, batch 4000, loss[loss=0.2088, simple_loss=0.2687, pruned_loss=0.07445, over 4816.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2419, pruned_loss=0.05136, over 956304.92 frames. ], batch size: 33, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:03:20,853 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-27 03:03:26,975 INFO [finetune.py:976] (4/7) Epoch 22, batch 4050, loss[loss=0.1682, simple_loss=0.242, pruned_loss=0.04719, over 4744.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2449, pruned_loss=0.05271, over 955464.13 frames. ], batch size: 28, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:03:27,710 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124333.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:03:39,254 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124348.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 03:03:46,869 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.585e+02 1.909e+02 2.213e+02 3.315e+02, threshold=3.818e+02, percent-clipped=0.0 +2023-03-27 03:04:00,188 INFO [finetune.py:976] (4/7) Epoch 22, batch 4100, loss[loss=0.1301, simple_loss=0.1986, pruned_loss=0.0308, over 4725.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2479, pruned_loss=0.05348, over 955981.21 frames. ], batch size: 23, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:04:07,291 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124392.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:04:14,208 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2626, 1.3781, 1.5601, 1.4669, 1.5296, 2.9342, 1.4014, 1.5525], + device='cuda:4'), covar=tensor([0.1029, 0.1726, 0.1144, 0.0973, 0.1518, 0.0290, 0.1362, 0.1631], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0092, 0.0081, 0.0086, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 03:04:19,528 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124409.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 03:04:24,897 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124418.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:04:29,730 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2254, 2.0582, 2.2353, 1.5397, 2.1090, 2.3629, 2.2304, 1.7306], + device='cuda:4'), covar=tensor([0.0561, 0.0691, 0.0642, 0.0823, 0.0720, 0.0581, 0.0595, 0.1151], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0135, 0.0139, 0.0120, 0.0125, 0.0137, 0.0138, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:04:33,284 INFO [finetune.py:976] (4/7) Epoch 22, batch 4150, loss[loss=0.1827, simple_loss=0.2461, pruned_loss=0.05971, over 4064.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.2499, pruned_loss=0.05414, over 955526.04 frames. ], batch size: 65, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:04:33,387 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124432.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:04:38,488 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-27 03:04:53,596 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.819e+01 1.555e+02 1.746e+02 2.178e+02 5.076e+02, threshold=3.492e+02, percent-clipped=2.0 +2023-03-27 03:05:06,938 INFO [finetune.py:976] (4/7) Epoch 22, batch 4200, loss[loss=0.1628, simple_loss=0.2367, pruned_loss=0.04445, over 4742.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.25, pruned_loss=0.05332, over 955944.01 frames. ], batch size: 54, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:05:10,162 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1586, 2.0647, 1.7703, 2.1152, 1.9752, 1.9757, 2.0173, 2.7168], + device='cuda:4'), covar=tensor([0.3786, 0.4323, 0.3352, 0.3915, 0.4248, 0.2424, 0.3796, 0.1676], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0264, 0.0235, 0.0278, 0.0256, 0.0226, 0.0255, 0.0236], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:05:14,724 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124493.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:05:40,316 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124528.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:05:42,631 INFO [finetune.py:976] (4/7) Epoch 22, batch 4250, loss[loss=0.1693, simple_loss=0.2396, pruned_loss=0.04943, over 4849.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2484, pruned_loss=0.05265, over 957319.44 frames. ], batch size: 49, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:05:44,089 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=7.88 vs. limit=5.0 +2023-03-27 03:05:51,126 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4668, 1.3364, 1.3386, 1.3291, 0.7523, 2.3107, 0.7074, 1.2532], + device='cuda:4'), covar=tensor([0.3296, 0.2514, 0.2219, 0.2503, 0.2062, 0.0340, 0.2728, 0.1265], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0115, 0.0121, 0.0123, 0.0113, 0.0095, 0.0094, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 03:06:02,064 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.200e+01 1.490e+02 1.704e+02 2.072e+02 3.423e+02, threshold=3.408e+02, percent-clipped=0.0 +2023-03-27 03:06:12,353 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124576.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:06:17,860 INFO [finetune.py:976] (4/7) Epoch 22, batch 4300, loss[loss=0.1561, simple_loss=0.2271, pruned_loss=0.04253, over 4910.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2458, pruned_loss=0.05182, over 955675.97 frames. ], batch size: 35, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:07:10,854 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124628.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:07:17,350 INFO [finetune.py:976] (4/7) Epoch 22, batch 4350, loss[loss=0.1618, simple_loss=0.233, pruned_loss=0.04526, over 4864.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2432, pruned_loss=0.05108, over 957401.32 frames. ], batch size: 31, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:07:21,126 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124638.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:07:48,794 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.554e+01 1.414e+02 1.739e+02 2.148e+02 3.782e+02, threshold=3.479e+02, percent-clipped=2.0 +2023-03-27 03:07:51,280 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124663.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:08:06,300 INFO [finetune.py:976] (4/7) Epoch 22, batch 4400, loss[loss=0.2043, simple_loss=0.2869, pruned_loss=0.06086, over 4939.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2443, pruned_loss=0.05153, over 955860.50 frames. ], batch size: 39, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:08:12,497 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124692.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:08:17,328 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124699.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:08:20,765 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124704.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 03:08:31,154 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124718.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:08:35,328 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124724.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:08:40,102 INFO [finetune.py:976] (4/7) Epoch 22, batch 4450, loss[loss=0.1954, simple_loss=0.2548, pruned_loss=0.06802, over 4828.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2464, pruned_loss=0.05192, over 956133.24 frames. ], batch size: 30, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:08:45,010 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124740.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:08:51,006 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.57 vs. limit=5.0 +2023-03-27 03:08:58,179 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.229e+02 1.653e+02 1.910e+02 2.206e+02 5.425e+02, threshold=3.820e+02, percent-clipped=3.0 +2023-03-27 03:09:02,351 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124766.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:09:13,853 INFO [finetune.py:976] (4/7) Epoch 22, batch 4500, loss[loss=0.1385, simple_loss=0.2229, pruned_loss=0.02702, over 4906.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2474, pruned_loss=0.05226, over 954199.88 frames. ], batch size: 37, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:09:17,603 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124788.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:09:26,915 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0546, 0.9682, 0.9279, 0.4355, 0.9201, 1.0981, 1.1329, 0.8608], + device='cuda:4'), covar=tensor([0.0806, 0.0701, 0.0628, 0.0512, 0.0654, 0.0721, 0.0451, 0.0737], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0150, 0.0127, 0.0123, 0.0132, 0.0131, 0.0142, 0.0150], + device='cuda:4'), out_proj_covar=tensor([8.9824e-05, 1.0866e-04, 9.1221e-05, 8.6671e-05, 9.2432e-05, 9.3121e-05, + 1.0162e-04, 1.0729e-04], device='cuda:4') +2023-03-27 03:09:38,623 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=124820.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:09:47,292 INFO [finetune.py:976] (4/7) Epoch 22, batch 4550, loss[loss=0.19, simple_loss=0.2655, pruned_loss=0.05728, over 4812.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2487, pruned_loss=0.05271, over 954905.45 frames. ], batch size: 45, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:09:56,530 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0784, 1.9277, 1.9420, 1.9743, 1.5665, 3.8681, 1.6904, 2.2450], + device='cuda:4'), covar=tensor([0.3035, 0.2483, 0.1921, 0.2116, 0.1504, 0.0248, 0.2370, 0.1077], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0124, 0.0113, 0.0096, 0.0094, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 03:10:04,841 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.024e+02 1.469e+02 1.742e+02 1.998e+02 4.285e+02, threshold=3.484e+02, percent-clipped=1.0 +2023-03-27 03:10:12,984 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.53 vs. limit=5.0 +2023-03-27 03:10:19,584 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=124881.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:10:20,090 INFO [finetune.py:976] (4/7) Epoch 22, batch 4600, loss[loss=0.1657, simple_loss=0.2503, pruned_loss=0.04056, over 4820.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.248, pruned_loss=0.05217, over 955432.43 frames. ], batch size: 40, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:10:48,845 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2254, 2.0879, 1.7848, 2.0349, 1.9903, 1.9669, 1.9764, 2.7563], + device='cuda:4'), covar=tensor([0.3681, 0.4327, 0.3427, 0.3900, 0.3877, 0.2443, 0.3704, 0.1766], + device='cuda:4'), in_proj_covar=tensor([0.0286, 0.0261, 0.0233, 0.0275, 0.0254, 0.0223, 0.0252, 0.0234], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:10:50,954 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=124928.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:10:53,307 INFO [finetune.py:976] (4/7) Epoch 22, batch 4650, loss[loss=0.1592, simple_loss=0.2302, pruned_loss=0.04414, over 4872.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2448, pruned_loss=0.05111, over 954870.96 frames. ], batch size: 34, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:11:10,704 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.467e+02 1.675e+02 2.008e+02 3.769e+02, threshold=3.350e+02, percent-clipped=1.0 +2023-03-27 03:11:21,849 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=124976.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:11:24,379 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0972, 1.9441, 1.7498, 1.8671, 1.8097, 1.8213, 1.8362, 2.5935], + device='cuda:4'), covar=tensor([0.3227, 0.3903, 0.2957, 0.3380, 0.3538, 0.2257, 0.3381, 0.1536], + device='cuda:4'), in_proj_covar=tensor([0.0287, 0.0262, 0.0234, 0.0275, 0.0254, 0.0224, 0.0252, 0.0235], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:11:24,420 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.31 vs. limit=5.0 +2023-03-27 03:11:26,384 INFO [finetune.py:976] (4/7) Epoch 22, batch 4700, loss[loss=0.1716, simple_loss=0.2384, pruned_loss=0.05236, over 4821.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2432, pruned_loss=0.05125, over 954532.64 frames. ], batch size: 51, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:11:36,920 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=124994.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:11:43,185 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125004.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 03:11:52,769 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125019.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:12:07,771 INFO [finetune.py:976] (4/7) Epoch 22, batch 4750, loss[loss=0.2202, simple_loss=0.2773, pruned_loss=0.08156, over 4073.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2404, pruned_loss=0.05046, over 953438.89 frames. ], batch size: 65, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:12:30,814 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125052.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 03:12:39,938 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.109e+02 1.467e+02 1.710e+02 2.084e+02 3.277e+02, threshold=3.420e+02, percent-clipped=0.0 +2023-03-27 03:13:08,203 INFO [finetune.py:976] (4/7) Epoch 22, batch 4800, loss[loss=0.1657, simple_loss=0.242, pruned_loss=0.04465, over 4226.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2429, pruned_loss=0.05102, over 953582.33 frames. ], batch size: 65, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:13:16,554 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125088.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:13:42,372 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125128.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:13:45,182 INFO [finetune.py:976] (4/7) Epoch 22, batch 4850, loss[loss=0.206, simple_loss=0.2808, pruned_loss=0.06558, over 4835.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2445, pruned_loss=0.05099, over 955030.52 frames. ], batch size: 47, lr: 3.13e-03, grad_scale: 64.0 +2023-03-27 03:13:47,686 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125136.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:14:00,641 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8103, 1.7087, 1.6612, 1.7659, 1.4574, 4.3906, 1.7183, 2.0319], + device='cuda:4'), covar=tensor([0.3306, 0.2563, 0.2153, 0.2351, 0.1688, 0.0126, 0.2437, 0.1250], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0115, 0.0121, 0.0123, 0.0113, 0.0096, 0.0094, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 03:14:04,216 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.151e+02 1.665e+02 1.915e+02 2.224e+02 3.844e+02, threshold=3.831e+02, percent-clipped=1.0 +2023-03-27 03:14:10,503 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.52 vs. limit=2.0 +2023-03-27 03:14:12,852 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0519, 2.0261, 1.8354, 2.1975, 2.5722, 2.0526, 2.0038, 1.5630], + device='cuda:4'), covar=tensor([0.2114, 0.1879, 0.1796, 0.1473, 0.1742, 0.1166, 0.2065, 0.1958], + device='cuda:4'), in_proj_covar=tensor([0.0244, 0.0210, 0.0212, 0.0196, 0.0243, 0.0188, 0.0215, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:14:14,554 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125176.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:14:18,598 INFO [finetune.py:976] (4/7) Epoch 22, batch 4900, loss[loss=0.2142, simple_loss=0.2867, pruned_loss=0.07082, over 4817.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2476, pruned_loss=0.05258, over 953132.78 frames. ], batch size: 41, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:14:23,440 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125189.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:14:52,067 INFO [finetune.py:976] (4/7) Epoch 22, batch 4950, loss[loss=0.175, simple_loss=0.2447, pruned_loss=0.0527, over 4723.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2479, pruned_loss=0.05225, over 953561.29 frames. ], batch size: 23, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:15:12,019 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.098e+02 1.500e+02 1.887e+02 2.169e+02 5.445e+02, threshold=3.774e+02, percent-clipped=5.0 +2023-03-27 03:15:25,140 INFO [finetune.py:976] (4/7) Epoch 22, batch 5000, loss[loss=0.1704, simple_loss=0.2359, pruned_loss=0.05244, over 4896.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2469, pruned_loss=0.05204, over 955453.54 frames. ], batch size: 32, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:15:33,515 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125294.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:15:50,231 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125319.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:15:58,104 INFO [finetune.py:976] (4/7) Epoch 22, batch 5050, loss[loss=0.2106, simple_loss=0.2785, pruned_loss=0.07135, over 4907.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2449, pruned_loss=0.052, over 956765.13 frames. ], batch size: 36, lr: 3.13e-03, grad_scale: 32.0 +2023-03-27 03:16:05,259 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125342.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:16:15,336 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125356.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:16:18,273 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.708e+01 1.493e+02 1.877e+02 2.229e+02 3.838e+02, threshold=3.755e+02, percent-clipped=1.0 +2023-03-27 03:16:22,418 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125367.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:16:27,957 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125376.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:16:31,543 INFO [finetune.py:976] (4/7) Epoch 22, batch 5100, loss[loss=0.1547, simple_loss=0.2136, pruned_loss=0.04791, over 4772.00 frames. ], tot_loss[loss=0.172, simple_loss=0.242, pruned_loss=0.05099, over 956179.73 frames. ], batch size: 54, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:16:55,561 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125417.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 03:16:58,448 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125421.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:17:05,017 INFO [finetune.py:976] (4/7) Epoch 22, batch 5150, loss[loss=0.1904, simple_loss=0.2547, pruned_loss=0.06302, over 4914.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2424, pruned_loss=0.0513, over 956832.44 frames. ], batch size: 37, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:17:13,056 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125437.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:17:23,294 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4375, 1.9152, 2.7496, 1.8351, 2.3456, 2.6481, 1.8025, 2.6152], + device='cuda:4'), covar=tensor([0.1208, 0.1952, 0.1088, 0.1698, 0.0941, 0.1230, 0.2611, 0.1002], + device='cuda:4'), in_proj_covar=tensor([0.0191, 0.0206, 0.0190, 0.0190, 0.0173, 0.0214, 0.0216, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:17:26,818 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6455, 1.4596, 2.0434, 3.3291, 2.1825, 2.3889, 0.8577, 2.7734], + device='cuda:4'), covar=tensor([0.1707, 0.1404, 0.1362, 0.0565, 0.0808, 0.1618, 0.1920, 0.0454], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0164, 0.0101, 0.0136, 0.0124, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 03:17:30,570 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-27 03:17:33,158 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.242e+02 1.608e+02 1.768e+02 2.290e+02 4.207e+02, threshold=3.536e+02, percent-clipped=1.0 +2023-03-27 03:17:36,285 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7000, 3.6297, 3.4210, 1.7288, 3.7495, 2.8894, 0.7789, 2.4968], + device='cuda:4'), covar=tensor([0.2417, 0.2172, 0.1654, 0.3600, 0.1083, 0.0995, 0.4689, 0.1671], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0179, 0.0162, 0.0130, 0.0162, 0.0124, 0.0150, 0.0125], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 03:17:45,317 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125476.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:17:46,512 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5600, 1.1922, 0.9511, 1.4719, 1.8892, 1.1700, 1.4175, 1.5071], + device='cuda:4'), covar=tensor([0.1500, 0.2053, 0.1923, 0.1183, 0.2093, 0.2238, 0.1460, 0.1843], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0096, 0.0111, 0.0093, 0.0121, 0.0095, 0.0100, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 03:17:48,791 INFO [finetune.py:976] (4/7) Epoch 22, batch 5200, loss[loss=0.14, simple_loss=0.2233, pruned_loss=0.02834, over 4836.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2457, pruned_loss=0.05223, over 957913.74 frames. ], batch size: 33, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:17:48,941 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125482.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:17:54,882 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125484.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:18:39,411 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125524.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:18:44,716 INFO [finetune.py:976] (4/7) Epoch 22, batch 5250, loss[loss=0.1813, simple_loss=0.2601, pruned_loss=0.05122, over 4768.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2474, pruned_loss=0.05272, over 958052.93 frames. ], batch size: 28, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:18:51,572 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3281, 1.2261, 1.1910, 1.2844, 1.5952, 1.5518, 1.2983, 1.1839], + device='cuda:4'), covar=tensor([0.0354, 0.0309, 0.0587, 0.0289, 0.0209, 0.0440, 0.0371, 0.0424], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0106, 0.0142, 0.0111, 0.0098, 0.0111, 0.0100, 0.0111], + device='cuda:4'), out_proj_covar=tensor([7.6025e-05, 8.1468e-05, 1.1109e-04, 8.4986e-05, 7.6000e-05, 8.1940e-05, + 7.4749e-05, 8.4671e-05], device='cuda:4') +2023-03-27 03:19:03,922 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.075e+02 1.572e+02 1.864e+02 2.118e+02 3.675e+02, threshold=3.728e+02, percent-clipped=1.0 +2023-03-27 03:19:18,600 INFO [finetune.py:976] (4/7) Epoch 22, batch 5300, loss[loss=0.2009, simple_loss=0.2662, pruned_loss=0.06783, over 4727.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2486, pruned_loss=0.05292, over 956528.02 frames. ], batch size: 23, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:19:32,531 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125603.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:19:52,429 INFO [finetune.py:976] (4/7) Epoch 22, batch 5350, loss[loss=0.1519, simple_loss=0.2286, pruned_loss=0.03763, over 4896.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2485, pruned_loss=0.0522, over 955307.01 frames. ], batch size: 37, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:19:56,776 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3100, 1.5914, 2.3133, 1.7488, 1.8459, 3.9224, 1.5142, 1.7347], + device='cuda:4'), covar=tensor([0.1007, 0.1639, 0.1138, 0.0978, 0.1430, 0.0184, 0.1309, 0.1619], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0081, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 03:20:01,717 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-27 03:20:10,863 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.759e+01 1.403e+02 1.743e+02 2.180e+02 4.274e+02, threshold=3.486e+02, percent-clipped=1.0 +2023-03-27 03:20:13,364 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125664.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:20:25,091 INFO [finetune.py:976] (4/7) Epoch 22, batch 5400, loss[loss=0.2021, simple_loss=0.2647, pruned_loss=0.06981, over 4916.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2461, pruned_loss=0.05163, over 954960.45 frames. ], batch size: 37, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:20:29,870 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5242, 2.3809, 1.9204, 0.9466, 2.1358, 2.0421, 1.8827, 2.0998], + device='cuda:4'), covar=tensor([0.0905, 0.0699, 0.1686, 0.2078, 0.1359, 0.2119, 0.2320, 0.1042], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0192, 0.0198, 0.0182, 0.0210, 0.0208, 0.0224, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:20:44,652 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125712.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 03:20:53,963 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125725.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:20:58,061 INFO [finetune.py:976] (4/7) Epoch 22, batch 5450, loss[loss=0.1859, simple_loss=0.2569, pruned_loss=0.0574, over 4820.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.243, pruned_loss=0.05075, over 955491.93 frames. ], batch size: 41, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:20:58,129 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125732.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:21:14,398 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9612, 1.5027, 2.0367, 1.9776, 1.7642, 1.6817, 1.9771, 1.9512], + device='cuda:4'), covar=tensor([0.3834, 0.3613, 0.2926, 0.3453, 0.4471, 0.3731, 0.4063, 0.2799], + device='cuda:4'), in_proj_covar=tensor([0.0260, 0.0245, 0.0265, 0.0288, 0.0285, 0.0263, 0.0295, 0.0248], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:21:15,868 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-27 03:21:17,000 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.470e+02 1.717e+02 2.000e+02 3.602e+02, threshold=3.434e+02, percent-clipped=1.0 +2023-03-27 03:21:27,718 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125777.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:21:31,103 INFO [finetune.py:976] (4/7) Epoch 22, batch 5500, loss[loss=0.1859, simple_loss=0.2504, pruned_loss=0.06071, over 4843.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2394, pruned_loss=0.04935, over 954876.68 frames. ], batch size: 47, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:21:32,433 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=125784.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:21:33,659 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=125786.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:22:04,441 INFO [finetune.py:976] (4/7) Epoch 22, batch 5550, loss[loss=0.2779, simple_loss=0.34, pruned_loss=0.1079, over 4092.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2413, pruned_loss=0.05036, over 952101.60 frames. ], batch size: 65, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:22:04,491 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=125832.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:22:05,403 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-27 03:22:32,016 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.163e+02 1.619e+02 1.949e+02 2.379e+02 4.295e+02, threshold=3.899e+02, percent-clipped=6.0 +2023-03-27 03:22:48,803 INFO [finetune.py:976] (4/7) Epoch 22, batch 5600, loss[loss=0.1898, simple_loss=0.2801, pruned_loss=0.04972, over 4815.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.244, pruned_loss=0.05108, over 949793.44 frames. ], batch size: 40, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:23:03,588 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5065, 2.3130, 1.9220, 2.4150, 2.2594, 1.9999, 2.6475, 2.3952], + device='cuda:4'), covar=tensor([0.1296, 0.2033, 0.2938, 0.2417, 0.2704, 0.1708, 0.2888, 0.1734], + device='cuda:4'), in_proj_covar=tensor([0.0187, 0.0188, 0.0234, 0.0252, 0.0247, 0.0204, 0.0214, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:23:20,374 INFO [finetune.py:976] (4/7) Epoch 22, batch 5650, loss[loss=0.1931, simple_loss=0.2783, pruned_loss=0.05392, over 4871.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2465, pruned_loss=0.05106, over 950732.55 frames. ], batch size: 44, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:23:47,401 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=125959.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:23:48,546 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.919e+01 1.591e+02 1.865e+02 2.221e+02 3.612e+02, threshold=3.730e+02, percent-clipped=0.0 +2023-03-27 03:24:11,014 INFO [finetune.py:976] (4/7) Epoch 22, batch 5700, loss[loss=0.1648, simple_loss=0.2177, pruned_loss=0.05598, over 4050.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2427, pruned_loss=0.05044, over 934423.82 frames. ], batch size: 17, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:24:20,552 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=125998.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:24:37,873 INFO [finetune.py:976] (4/7) Epoch 23, batch 0, loss[loss=0.174, simple_loss=0.2489, pruned_loss=0.0495, over 4916.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2489, pruned_loss=0.0495, over 4916.00 frames. ], batch size: 33, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:24:37,873 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 03:24:43,654 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6957, 1.6006, 2.0878, 2.9511, 1.9472, 2.3192, 1.0561, 2.4623], + device='cuda:4'), covar=tensor([0.1548, 0.1225, 0.1020, 0.0521, 0.0871, 0.1111, 0.1611, 0.0502], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0163, 0.0100, 0.0135, 0.0123, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 03:24:44,653 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1115, 1.8691, 2.1290, 1.2864, 2.0683, 2.0934, 2.1284, 1.7628], + device='cuda:4'), covar=tensor([0.0518, 0.0682, 0.0547, 0.0825, 0.0914, 0.0648, 0.0530, 0.1085], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0134, 0.0138, 0.0119, 0.0123, 0.0137, 0.0137, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:24:53,039 INFO [finetune.py:1010] (4/7) Epoch 23, validation: loss=0.1587, simple_loss=0.2268, pruned_loss=0.04533, over 2265189.00 frames. +2023-03-27 03:24:53,039 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-27 03:24:54,189 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4781, 1.4673, 1.9625, 1.7939, 1.5988, 3.1560, 1.3004, 1.5526], + device='cuda:4'), covar=tensor([0.0897, 0.1638, 0.1074, 0.0841, 0.1450, 0.0276, 0.1429, 0.1639], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0092, 0.0081, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 03:24:59,386 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126012.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 03:25:12,073 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126032.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:25:30,342 INFO [finetune.py:976] (4/7) Epoch 23, batch 50, loss[loss=0.1652, simple_loss=0.2303, pruned_loss=0.05005, over 4123.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2479, pruned_loss=0.05142, over 216550.72 frames. ], batch size: 17, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:25:30,929 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=126059.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:25:31,916 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126060.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:25:32,453 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.266e+01 1.545e+02 1.923e+02 2.325e+02 3.929e+02, threshold=3.846e+02, percent-clipped=1.0 +2023-03-27 03:25:42,851 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126077.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:25:44,665 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126080.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:25:45,301 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=126081.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:26:03,745 INFO [finetune.py:976] (4/7) Epoch 23, batch 100, loss[loss=0.1343, simple_loss=0.2159, pruned_loss=0.02635, over 4779.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.24, pruned_loss=0.0486, over 379359.58 frames. ], batch size: 29, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:26:11,481 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6044, 1.1207, 0.7586, 1.4821, 2.0940, 1.1412, 1.3915, 1.4535], + device='cuda:4'), covar=tensor([0.1552, 0.2253, 0.2046, 0.1245, 0.2067, 0.1895, 0.1623, 0.2085], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0095, 0.0110, 0.0092, 0.0120, 0.0094, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 03:26:14,893 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126125.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:26:19,894 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7207, 2.4802, 2.0227, 1.1279, 2.1670, 2.1187, 1.8705, 2.2223], + device='cuda:4'), covar=tensor([0.0783, 0.0819, 0.1649, 0.2023, 0.1493, 0.2136, 0.2207, 0.1070], + device='cuda:4'), in_proj_covar=tensor([0.0171, 0.0194, 0.0201, 0.0184, 0.0211, 0.0210, 0.0226, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:26:37,115 INFO [finetune.py:976] (4/7) Epoch 23, batch 150, loss[loss=0.1608, simple_loss=0.2327, pruned_loss=0.04448, over 4870.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2392, pruned_loss=0.05035, over 507678.26 frames. ], batch size: 34, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:26:38,296 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.556e+02 1.791e+02 2.255e+02 5.687e+02, threshold=3.583e+02, percent-clipped=3.0 +2023-03-27 03:27:10,674 INFO [finetune.py:976] (4/7) Epoch 23, batch 200, loss[loss=0.186, simple_loss=0.2519, pruned_loss=0.06002, over 4900.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2384, pruned_loss=0.051, over 608368.01 frames. ], batch size: 32, lr: 3.12e-03, grad_scale: 32.0 +2023-03-27 03:28:05,167 INFO [finetune.py:976] (4/7) Epoch 23, batch 250, loss[loss=0.2042, simple_loss=0.2719, pruned_loss=0.06826, over 4870.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.244, pruned_loss=0.05283, over 685090.75 frames. ], batch size: 31, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:28:05,280 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126259.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:28:06,385 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.143e+02 1.625e+02 1.849e+02 2.180e+02 4.181e+02, threshold=3.697e+02, percent-clipped=1.0 +2023-03-27 03:28:37,017 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126307.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:28:40,578 INFO [finetune.py:976] (4/7) Epoch 23, batch 300, loss[loss=0.2174, simple_loss=0.2879, pruned_loss=0.07343, over 4924.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2469, pruned_loss=0.05287, over 743752.53 frames. ], batch size: 36, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:29:00,157 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.32 vs. limit=5.0 +2023-03-27 03:29:20,738 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=126354.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:29:24,671 INFO [finetune.py:976] (4/7) Epoch 23, batch 350, loss[loss=0.1417, simple_loss=0.2194, pruned_loss=0.03201, over 4822.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2484, pruned_loss=0.05267, over 791162.83 frames. ], batch size: 25, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:29:25,831 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.521e+01 1.514e+02 1.805e+02 2.248e+02 3.946e+02, threshold=3.610e+02, percent-clipped=1.0 +2023-03-27 03:29:39,004 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=126380.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:29:39,583 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126381.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:29:45,203 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.62 vs. limit=5.0 +2023-03-27 03:29:56,091 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1741, 1.7810, 2.2084, 2.0996, 1.8451, 1.8351, 2.0642, 2.0280], + device='cuda:4'), covar=tensor([0.4267, 0.4499, 0.3392, 0.4295, 0.5509, 0.4524, 0.4937, 0.3205], + device='cuda:4'), in_proj_covar=tensor([0.0259, 0.0244, 0.0264, 0.0287, 0.0284, 0.0262, 0.0293, 0.0246], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:29:59,416 INFO [finetune.py:976] (4/7) Epoch 23, batch 400, loss[loss=0.1685, simple_loss=0.2426, pruned_loss=0.04723, over 4822.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2483, pruned_loss=0.05222, over 827979.65 frames. ], batch size: 38, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:30:21,835 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126429.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:30:25,161 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-27 03:30:29,125 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7400, 1.6456, 1.4262, 1.6446, 2.0574, 1.8716, 1.5942, 1.4555], + device='cuda:4'), covar=tensor([0.0297, 0.0299, 0.0599, 0.0280, 0.0180, 0.0462, 0.0362, 0.0380], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0106, 0.0142, 0.0111, 0.0098, 0.0111, 0.0101, 0.0111], + device='cuda:4'), out_proj_covar=tensor([7.6345e-05, 8.1182e-05, 1.1140e-04, 8.5128e-05, 7.6349e-05, 8.1906e-05, + 7.5054e-05, 8.4533e-05], device='cuda:4') +2023-03-27 03:30:29,725 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=126441.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:30:35,181 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=126450.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:30:40,549 INFO [finetune.py:976] (4/7) Epoch 23, batch 450, loss[loss=0.1928, simple_loss=0.2672, pruned_loss=0.05924, over 4808.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2475, pruned_loss=0.0515, over 857872.75 frames. ], batch size: 41, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:30:42,258 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.347e+01 1.414e+02 1.639e+02 2.017e+02 3.767e+02, threshold=3.277e+02, percent-clipped=3.0 +2023-03-27 03:30:42,666 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-27 03:30:45,276 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9903, 1.9472, 1.8155, 2.1671, 2.6298, 2.1631, 1.9252, 1.6387], + device='cuda:4'), covar=tensor([0.2274, 0.1956, 0.1932, 0.1738, 0.1566, 0.1196, 0.2190, 0.2003], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0211, 0.0213, 0.0197, 0.0244, 0.0189, 0.0217, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:31:13,831 INFO [finetune.py:976] (4/7) Epoch 23, batch 500, loss[loss=0.1691, simple_loss=0.2414, pruned_loss=0.04844, over 4824.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2448, pruned_loss=0.05056, over 879788.87 frames. ], batch size: 38, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:31:15,655 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=126511.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:31:47,409 INFO [finetune.py:976] (4/7) Epoch 23, batch 550, loss[loss=0.1832, simple_loss=0.2423, pruned_loss=0.06205, over 4781.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.242, pruned_loss=0.05046, over 897762.27 frames. ], batch size: 28, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:31:48,613 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.316e+01 1.588e+02 1.845e+02 2.407e+02 4.330e+02, threshold=3.691e+02, percent-clipped=4.0 +2023-03-27 03:32:13,000 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-27 03:32:21,215 INFO [finetune.py:976] (4/7) Epoch 23, batch 600, loss[loss=0.2064, simple_loss=0.2955, pruned_loss=0.05871, over 4891.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2434, pruned_loss=0.05101, over 912430.05 frames. ], batch size: 43, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:32:31,631 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-27 03:32:44,163 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.61 vs. limit=5.0 +2023-03-27 03:33:02,512 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=126654.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:33:03,819 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7007, 1.6096, 1.3992, 1.3554, 1.7779, 1.5037, 1.8358, 1.7230], + device='cuda:4'), covar=tensor([0.1512, 0.1943, 0.3123, 0.2553, 0.2659, 0.1798, 0.2961, 0.1824], + device='cuda:4'), in_proj_covar=tensor([0.0190, 0.0190, 0.0237, 0.0256, 0.0251, 0.0207, 0.0217, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:33:05,532 INFO [finetune.py:976] (4/7) Epoch 23, batch 650, loss[loss=0.1575, simple_loss=0.2211, pruned_loss=0.04696, over 4048.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2459, pruned_loss=0.05194, over 918326.73 frames. ], batch size: 17, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:33:06,763 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.083e+02 1.579e+02 1.902e+02 2.270e+02 1.001e+03, threshold=3.804e+02, percent-clipped=1.0 +2023-03-27 03:33:11,626 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 03:33:13,527 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-27 03:33:42,521 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=126702.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:33:47,143 INFO [finetune.py:976] (4/7) Epoch 23, batch 700, loss[loss=0.1326, simple_loss=0.2013, pruned_loss=0.03194, over 4035.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2478, pruned_loss=0.05224, over 926894.99 frames. ], batch size: 17, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:34:11,546 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=126736.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:34:29,142 INFO [finetune.py:976] (4/7) Epoch 23, batch 750, loss[loss=0.1869, simple_loss=0.2618, pruned_loss=0.056, over 4921.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2499, pruned_loss=0.05306, over 934512.45 frames. ], batch size: 33, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:34:30,819 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.063e+01 1.442e+02 1.710e+02 2.057e+02 3.398e+02, threshold=3.419e+02, percent-clipped=0.0 +2023-03-27 03:34:43,719 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-27 03:34:44,557 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1917, 1.8228, 2.2353, 3.9536, 2.6233, 2.7903, 0.8833, 3.1975], + device='cuda:4'), covar=tensor([0.1614, 0.1318, 0.1482, 0.0523, 0.0719, 0.1536, 0.1992, 0.0419], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0163, 0.0100, 0.0136, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 03:34:56,260 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7174, 1.4797, 2.1194, 3.2729, 2.1296, 2.3392, 1.1706, 2.6570], + device='cuda:4'), covar=tensor([0.1608, 0.1487, 0.1274, 0.0616, 0.0819, 0.1572, 0.1723, 0.0530], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0163, 0.0100, 0.0136, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 03:35:00,638 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=126806.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:35:02,394 INFO [finetune.py:976] (4/7) Epoch 23, batch 800, loss[loss=0.2574, simple_loss=0.3142, pruned_loss=0.1003, over 4114.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2497, pruned_loss=0.05247, over 938716.17 frames. ], batch size: 66, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:35:07,272 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=126816.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:35:44,510 INFO [finetune.py:976] (4/7) Epoch 23, batch 850, loss[loss=0.2054, simple_loss=0.2667, pruned_loss=0.07205, over 4929.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2479, pruned_loss=0.05225, over 943430.55 frames. ], batch size: 33, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:35:45,682 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.329e+01 1.478e+02 1.768e+02 2.024e+02 3.574e+02, threshold=3.536e+02, percent-clipped=1.0 +2023-03-27 03:35:56,144 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=126877.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:36:18,421 INFO [finetune.py:976] (4/7) Epoch 23, batch 900, loss[loss=0.1443, simple_loss=0.2161, pruned_loss=0.03629, over 4792.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2448, pruned_loss=0.05114, over 946735.09 frames. ], batch size: 26, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:36:52,033 INFO [finetune.py:976] (4/7) Epoch 23, batch 950, loss[loss=0.214, simple_loss=0.2872, pruned_loss=0.07035, over 4812.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2433, pruned_loss=0.05115, over 949684.33 frames. ], batch size: 40, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:36:53,227 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.057e+02 1.486e+02 1.775e+02 2.132e+02 3.351e+02, threshold=3.551e+02, percent-clipped=0.0 +2023-03-27 03:37:26,096 INFO [finetune.py:976] (4/7) Epoch 23, batch 1000, loss[loss=0.2088, simple_loss=0.2742, pruned_loss=0.07172, over 4936.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2453, pruned_loss=0.05192, over 951065.47 frames. ], batch size: 38, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:37:43,510 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=127036.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:37:51,213 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1400, 2.0400, 1.8072, 2.1507, 2.5927, 2.1129, 2.0819, 1.5735], + device='cuda:4'), covar=tensor([0.2022, 0.1932, 0.1880, 0.1615, 0.1749, 0.1185, 0.2085, 0.1817], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0211, 0.0214, 0.0198, 0.0245, 0.0190, 0.0217, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:38:01,449 INFO [finetune.py:976] (4/7) Epoch 23, batch 1050, loss[loss=0.2222, simple_loss=0.2823, pruned_loss=0.08108, over 4928.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2487, pruned_loss=0.05302, over 952091.98 frames. ], batch size: 38, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:38:02,653 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.597e+02 1.830e+02 2.248e+02 5.450e+02, threshold=3.660e+02, percent-clipped=4.0 +2023-03-27 03:38:27,256 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=127084.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:38:42,648 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=127106.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:38:44,898 INFO [finetune.py:976] (4/7) Epoch 23, batch 1100, loss[loss=0.1909, simple_loss=0.2712, pruned_loss=0.05535, over 4851.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2497, pruned_loss=0.05339, over 952045.68 frames. ], batch size: 44, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:39:08,781 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.87 vs. limit=5.0 +2023-03-27 03:39:14,543 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=127154.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:39:21,917 INFO [finetune.py:976] (4/7) Epoch 23, batch 1150, loss[loss=0.1399, simple_loss=0.2099, pruned_loss=0.03499, over 4737.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2495, pruned_loss=0.05317, over 950983.72 frames. ], batch size: 23, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:39:23,614 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.031e+02 1.547e+02 1.781e+02 2.032e+02 3.739e+02, threshold=3.562e+02, percent-clipped=1.0 +2023-03-27 03:39:33,493 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.10 vs. limit=5.0 +2023-03-27 03:39:35,836 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=127172.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:40:04,150 INFO [finetune.py:976] (4/7) Epoch 23, batch 1200, loss[loss=0.1562, simple_loss=0.222, pruned_loss=0.04519, over 4751.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2471, pruned_loss=0.05211, over 951652.52 frames. ], batch size: 23, lr: 3.11e-03, grad_scale: 64.0 +2023-03-27 03:40:10,142 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1304, 2.0528, 2.1718, 0.9825, 2.5252, 2.7419, 2.4159, 1.8611], + device='cuda:4'), covar=tensor([0.0893, 0.0618, 0.0541, 0.0632, 0.0409, 0.0607, 0.0438, 0.0652], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0148, 0.0127, 0.0122, 0.0130, 0.0129, 0.0141, 0.0148], + device='cuda:4'), out_proj_covar=tensor([8.9516e-05, 1.0709e-04, 9.0898e-05, 8.5929e-05, 9.1560e-05, 9.1630e-05, + 1.0076e-04, 1.0597e-04], device='cuda:4') +2023-03-27 03:40:33,021 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127249.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:40:42,285 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8514, 1.2133, 0.7189, 1.8054, 2.2728, 1.7356, 1.4098, 1.6773], + device='cuda:4'), covar=tensor([0.1334, 0.2182, 0.2101, 0.1143, 0.1758, 0.1938, 0.1497, 0.1921], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0094, 0.0110, 0.0092, 0.0119, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 03:40:44,043 INFO [finetune.py:976] (4/7) Epoch 23, batch 1250, loss[loss=0.1379, simple_loss=0.2025, pruned_loss=0.03663, over 4836.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2443, pruned_loss=0.05128, over 951483.43 frames. ], batch size: 30, lr: 3.11e-03, grad_scale: 64.0 +2023-03-27 03:40:45,212 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.489e+02 1.742e+02 2.248e+02 3.707e+02, threshold=3.484e+02, percent-clipped=1.0 +2023-03-27 03:41:21,264 INFO [finetune.py:976] (4/7) Epoch 23, batch 1300, loss[loss=0.1548, simple_loss=0.2216, pruned_loss=0.04396, over 4101.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2416, pruned_loss=0.05043, over 953121.79 frames. ], batch size: 65, lr: 3.11e-03, grad_scale: 64.0 +2023-03-27 03:41:22,025 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127310.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:41:23,225 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5970, 1.5165, 1.4829, 1.5535, 1.1164, 2.9651, 1.1247, 1.4970], + device='cuda:4'), covar=tensor([0.3197, 0.2473, 0.2143, 0.2365, 0.1735, 0.0278, 0.2548, 0.1312], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0116, 0.0120, 0.0123, 0.0113, 0.0095, 0.0094, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 03:41:54,396 INFO [finetune.py:976] (4/7) Epoch 23, batch 1350, loss[loss=0.1321, simple_loss=0.2035, pruned_loss=0.03029, over 4214.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2428, pruned_loss=0.05113, over 954287.44 frames. ], batch size: 18, lr: 3.11e-03, grad_scale: 64.0 +2023-03-27 03:41:55,605 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.500e+02 1.797e+02 2.250e+02 4.549e+02, threshold=3.594e+02, percent-clipped=1.0 +2023-03-27 03:41:56,842 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7867, 1.7525, 1.7042, 1.7991, 1.5986, 4.5146, 1.6626, 2.2512], + device='cuda:4'), covar=tensor([0.3341, 0.2598, 0.2102, 0.2373, 0.1584, 0.0146, 0.2365, 0.1163], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0117, 0.0121, 0.0124, 0.0113, 0.0096, 0.0094, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 03:42:27,768 INFO [finetune.py:976] (4/7) Epoch 23, batch 1400, loss[loss=0.2076, simple_loss=0.2909, pruned_loss=0.06212, over 4841.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2457, pruned_loss=0.0516, over 955950.68 frames. ], batch size: 49, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:42:28,504 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6040, 1.5021, 1.4753, 0.8627, 1.6653, 1.8623, 1.8855, 1.3988], + device='cuda:4'), covar=tensor([0.0893, 0.0595, 0.0545, 0.0539, 0.0409, 0.0653, 0.0333, 0.0713], + device='cuda:4'), in_proj_covar=tensor([0.0124, 0.0149, 0.0128, 0.0122, 0.0131, 0.0130, 0.0142, 0.0149], + device='cuda:4'), out_proj_covar=tensor([9.0077e-05, 1.0739e-04, 9.1468e-05, 8.6158e-05, 9.1887e-05, 9.2330e-05, + 1.0113e-04, 1.0634e-04], device='cuda:4') +2023-03-27 03:42:40,729 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127427.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 03:43:01,002 INFO [finetune.py:976] (4/7) Epoch 23, batch 1450, loss[loss=0.1838, simple_loss=0.2547, pruned_loss=0.05639, over 4792.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2463, pruned_loss=0.05116, over 956379.04 frames. ], batch size: 26, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:43:03,313 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.817e+01 1.596e+02 1.913e+02 2.194e+02 3.811e+02, threshold=3.827e+02, percent-clipped=1.0 +2023-03-27 03:43:09,876 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=127472.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:43:25,082 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127488.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 03:43:46,709 INFO [finetune.py:976] (4/7) Epoch 23, batch 1500, loss[loss=0.1905, simple_loss=0.2636, pruned_loss=0.05873, over 4880.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2467, pruned_loss=0.05123, over 954342.75 frames. ], batch size: 32, lr: 3.11e-03, grad_scale: 32.0 +2023-03-27 03:43:54,392 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=127520.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:44:00,476 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127529.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:44:20,425 INFO [finetune.py:976] (4/7) Epoch 23, batch 1550, loss[loss=0.1526, simple_loss=0.2283, pruned_loss=0.03842, over 4895.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2469, pruned_loss=0.05098, over 954162.12 frames. ], batch size: 32, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:44:22,242 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.462e+02 1.755e+02 2.123e+02 3.197e+02, threshold=3.511e+02, percent-clipped=0.0 +2023-03-27 03:44:45,094 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8062, 1.8833, 1.6560, 1.8230, 1.5476, 4.4916, 1.6004, 2.0779], + device='cuda:4'), covar=tensor([0.3268, 0.2457, 0.2012, 0.2239, 0.1503, 0.0154, 0.2432, 0.1217], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0117, 0.0121, 0.0123, 0.0113, 0.0096, 0.0094, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 03:44:47,995 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127590.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:45:04,790 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=127605.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:45:07,684 INFO [finetune.py:976] (4/7) Epoch 23, batch 1600, loss[loss=0.1489, simple_loss=0.2286, pruned_loss=0.03459, over 4790.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2441, pruned_loss=0.04985, over 955300.43 frames. ], batch size: 29, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:45:40,959 INFO [finetune.py:976] (4/7) Epoch 23, batch 1650, loss[loss=0.205, simple_loss=0.2681, pruned_loss=0.07094, over 4707.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2408, pruned_loss=0.04933, over 954213.98 frames. ], batch size: 59, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:45:43,335 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.567e+02 1.812e+02 2.212e+02 4.212e+02, threshold=3.624e+02, percent-clipped=4.0 +2023-03-27 03:45:43,458 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127662.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:46:26,933 INFO [finetune.py:976] (4/7) Epoch 23, batch 1700, loss[loss=0.1712, simple_loss=0.2265, pruned_loss=0.05793, over 4778.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2403, pruned_loss=0.04948, over 954877.98 frames. ], batch size: 28, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:46:36,929 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127723.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:47:00,095 INFO [finetune.py:976] (4/7) Epoch 23, batch 1750, loss[loss=0.1815, simple_loss=0.2598, pruned_loss=0.05155, over 4899.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2428, pruned_loss=0.05053, over 954429.80 frames. ], batch size: 35, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:47:01,902 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.127e+02 1.511e+02 1.783e+02 2.157e+02 3.427e+02, threshold=3.565e+02, percent-clipped=0.0 +2023-03-27 03:47:03,266 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7220, 2.6110, 2.1243, 1.1127, 2.2420, 2.0376, 1.9931, 2.3526], + device='cuda:4'), covar=tensor([0.0781, 0.0700, 0.1681, 0.1951, 0.1338, 0.2216, 0.2145, 0.0821], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0193, 0.0199, 0.0182, 0.0210, 0.0209, 0.0225, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:47:13,311 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7520, 2.5414, 2.1237, 2.8414, 2.6477, 2.2930, 3.1360, 2.6401], + device='cuda:4'), covar=tensor([0.1177, 0.2007, 0.2747, 0.2371, 0.2284, 0.1532, 0.2801, 0.1736], + device='cuda:4'), in_proj_covar=tensor([0.0188, 0.0187, 0.0234, 0.0253, 0.0248, 0.0204, 0.0214, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:47:13,910 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127779.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:47:16,782 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=127783.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 03:47:33,037 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-27 03:47:33,985 INFO [finetune.py:976] (4/7) Epoch 23, batch 1800, loss[loss=0.1719, simple_loss=0.2489, pruned_loss=0.04741, over 4865.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2456, pruned_loss=0.05116, over 954418.58 frames. ], batch size: 31, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:47:48,290 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-27 03:47:54,753 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127840.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:48:00,649 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0360, 1.9497, 1.5728, 1.8336, 1.8584, 1.8119, 1.8994, 2.5896], + device='cuda:4'), covar=tensor([0.3912, 0.4025, 0.3577, 0.3670, 0.4009, 0.2753, 0.3684, 0.1857], + device='cuda:4'), in_proj_covar=tensor([0.0288, 0.0262, 0.0234, 0.0275, 0.0255, 0.0227, 0.0253, 0.0236], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:48:01,858 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127850.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:48:07,745 INFO [finetune.py:976] (4/7) Epoch 23, batch 1850, loss[loss=0.1979, simple_loss=0.2586, pruned_loss=0.06859, over 4756.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2473, pruned_loss=0.05174, over 955622.35 frames. ], batch size: 54, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:48:09,576 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.256e+01 1.480e+02 1.674e+02 2.095e+02 4.093e+02, threshold=3.347e+02, percent-clipped=1.0 +2023-03-27 03:48:16,137 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=127872.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:48:24,939 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=127885.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:48:40,251 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=127905.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:48:40,451 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-27 03:48:42,624 INFO [finetune.py:976] (4/7) Epoch 23, batch 1900, loss[loss=0.1768, simple_loss=0.259, pruned_loss=0.04732, over 4828.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2487, pruned_loss=0.05232, over 954510.07 frames. ], batch size: 47, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:48:43,932 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127911.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:48:58,799 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=127933.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:49:04,381 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-27 03:49:11,958 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=127953.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:49:16,015 INFO [finetune.py:976] (4/7) Epoch 23, batch 1950, loss[loss=0.1873, simple_loss=0.2547, pruned_loss=0.0599, over 4813.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2473, pruned_loss=0.05159, over 953983.28 frames. ], batch size: 40, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:49:17,848 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.653e+01 1.391e+02 1.764e+02 2.084e+02 3.552e+02, threshold=3.528e+02, percent-clipped=1.0 +2023-03-27 03:49:53,066 INFO [finetune.py:976] (4/7) Epoch 23, batch 2000, loss[loss=0.156, simple_loss=0.2215, pruned_loss=0.04524, over 4821.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2452, pruned_loss=0.05101, over 953889.17 frames. ], batch size: 30, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:50:07,579 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128018.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:50:38,969 INFO [finetune.py:976] (4/7) Epoch 23, batch 2050, loss[loss=0.1574, simple_loss=0.2213, pruned_loss=0.04681, over 4832.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2411, pruned_loss=0.04981, over 954806.21 frames. ], batch size: 33, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:50:41,272 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.050e+02 1.483e+02 1.749e+02 2.101e+02 3.191e+02, threshold=3.498e+02, percent-clipped=0.0 +2023-03-27 03:50:54,710 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128083.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 03:51:13,973 INFO [finetune.py:976] (4/7) Epoch 23, batch 2100, loss[loss=0.1882, simple_loss=0.2563, pruned_loss=0.05998, over 4831.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2417, pruned_loss=0.05042, over 953347.84 frames. ], batch size: 39, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:51:21,161 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0725, 2.1338, 2.7533, 1.5666, 2.2243, 2.5212, 1.8273, 2.5741], + device='cuda:4'), covar=tensor([0.1477, 0.1820, 0.1327, 0.2291, 0.0894, 0.1540, 0.2661, 0.0928], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0207, 0.0193, 0.0190, 0.0174, 0.0215, 0.0217, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:51:40,716 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128131.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 03:51:43,605 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128135.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:52:00,527 INFO [finetune.py:976] (4/7) Epoch 23, batch 2150, loss[loss=0.1849, simple_loss=0.2661, pruned_loss=0.05178, over 4900.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2447, pruned_loss=0.05124, over 953516.31 frames. ], batch size: 36, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:52:02,378 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.558e+02 1.811e+02 2.178e+02 3.611e+02, threshold=3.622e+02, percent-clipped=2.0 +2023-03-27 03:52:17,009 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128185.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:52:18,125 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2806, 2.0991, 1.6576, 0.8477, 1.9356, 1.7658, 1.4683, 1.9604], + device='cuda:4'), covar=tensor([0.0895, 0.0871, 0.1640, 0.1892, 0.1311, 0.2493, 0.2765, 0.0865], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0192, 0.0200, 0.0183, 0.0210, 0.0209, 0.0225, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:52:31,848 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128206.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:52:33,600 INFO [finetune.py:976] (4/7) Epoch 23, batch 2200, loss[loss=0.173, simple_loss=0.2524, pruned_loss=0.04678, over 4807.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2476, pruned_loss=0.05182, over 954400.32 frames. ], batch size: 25, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:52:46,642 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128228.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:52:49,639 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128233.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:53:07,269 INFO [finetune.py:976] (4/7) Epoch 23, batch 2250, loss[loss=0.1953, simple_loss=0.2437, pruned_loss=0.07348, over 4801.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2486, pruned_loss=0.05235, over 956265.47 frames. ], batch size: 25, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:53:09,089 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.465e+01 1.491e+02 1.772e+02 2.217e+02 3.841e+02, threshold=3.544e+02, percent-clipped=1.0 +2023-03-27 03:53:15,694 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-27 03:53:38,817 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-27 03:53:40,846 INFO [finetune.py:976] (4/7) Epoch 23, batch 2300, loss[loss=0.1727, simple_loss=0.2558, pruned_loss=0.04473, over 4902.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2479, pruned_loss=0.05187, over 954347.00 frames. ], batch size: 35, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:53:47,290 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128318.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:53:47,405 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-27 03:53:48,519 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=128320.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:54:13,561 INFO [finetune.py:976] (4/7) Epoch 23, batch 2350, loss[loss=0.1859, simple_loss=0.2553, pruned_loss=0.05829, over 4854.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2454, pruned_loss=0.05116, over 952393.58 frames. ], batch size: 44, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:54:15,447 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6880, 2.2226, 1.8636, 0.8924, 2.1985, 1.9732, 1.5773, 2.0831], + device='cuda:4'), covar=tensor([0.0835, 0.1258, 0.2266, 0.2531, 0.1535, 0.2540, 0.2972, 0.1153], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0192, 0.0199, 0.0182, 0.0209, 0.0208, 0.0224, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:54:15,916 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.510e+02 1.720e+02 2.103e+02 3.385e+02, threshold=3.440e+02, percent-clipped=0.0 +2023-03-27 03:54:17,859 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7656, 1.2062, 0.8014, 1.7070, 2.1042, 1.5101, 1.5366, 1.6004], + device='cuda:4'), covar=tensor([0.1337, 0.1964, 0.1878, 0.1064, 0.1892, 0.1962, 0.1369, 0.1762], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0111, 0.0092, 0.0120, 0.0094, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 03:54:18,452 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128366.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:54:25,478 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1114, 1.9521, 1.8476, 2.0253, 1.8217, 1.8375, 1.9382, 2.5549], + device='cuda:4'), covar=tensor([0.2625, 0.3276, 0.2544, 0.2743, 0.3567, 0.2083, 0.3303, 0.1161], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0263, 0.0236, 0.0277, 0.0257, 0.0228, 0.0255, 0.0237], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:54:29,005 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=128381.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:54:36,173 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4204, 2.9608, 2.7649, 1.3325, 3.0692, 2.2800, 0.8153, 2.0120], + device='cuda:4'), covar=tensor([0.2017, 0.2332, 0.1861, 0.3471, 0.1339, 0.1212, 0.4027, 0.1663], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0178, 0.0159, 0.0129, 0.0161, 0.0124, 0.0149, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 03:54:46,898 INFO [finetune.py:976] (4/7) Epoch 23, batch 2400, loss[loss=0.1655, simple_loss=0.2306, pruned_loss=0.05017, over 4809.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2419, pruned_loss=0.05002, over 954102.38 frames. ], batch size: 25, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:54:49,476 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=128413.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:55:06,799 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128435.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:55:35,542 INFO [finetune.py:976] (4/7) Epoch 23, batch 2450, loss[loss=0.1417, simple_loss=0.2202, pruned_loss=0.03162, over 4908.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2394, pruned_loss=0.04913, over 954717.20 frames. ], batch size: 43, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:55:41,390 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.054e+01 1.492e+02 1.779e+02 2.209e+02 4.084e+02, threshold=3.557e+02, percent-clipped=2.0 +2023-03-27 03:55:49,817 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=128474.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:55:55,749 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128483.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:56:10,254 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128506.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:56:12,479 INFO [finetune.py:976] (4/7) Epoch 23, batch 2500, loss[loss=0.1337, simple_loss=0.2085, pruned_loss=0.02948, over 4776.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2417, pruned_loss=0.05011, over 956195.85 frames. ], batch size: 28, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:56:25,985 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128528.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:56:30,117 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4264, 0.9459, 0.7818, 1.2679, 1.8021, 0.6818, 1.0700, 1.1588], + device='cuda:4'), covar=tensor([0.1571, 0.2399, 0.1814, 0.1292, 0.2040, 0.2155, 0.1714, 0.2153], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0095, 0.0111, 0.0093, 0.0120, 0.0095, 0.0100, 0.0090], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 03:56:48,370 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128554.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:56:55,780 INFO [finetune.py:976] (4/7) Epoch 23, batch 2550, loss[loss=0.1328, simple_loss=0.2184, pruned_loss=0.02356, over 4903.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2454, pruned_loss=0.0516, over 955946.16 frames. ], batch size: 43, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:56:58,590 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.110e+02 1.563e+02 1.837e+02 2.202e+02 4.665e+02, threshold=3.674e+02, percent-clipped=3.0 +2023-03-27 03:57:11,632 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=128576.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:57:33,477 INFO [finetune.py:976] (4/7) Epoch 23, batch 2600, loss[loss=0.1703, simple_loss=0.2575, pruned_loss=0.04158, over 4817.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2472, pruned_loss=0.05198, over 951454.97 frames. ], batch size: 45, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:58:07,052 INFO [finetune.py:976] (4/7) Epoch 23, batch 2650, loss[loss=0.1631, simple_loss=0.238, pruned_loss=0.04416, over 4768.00 frames. ], tot_loss[loss=0.175, simple_loss=0.247, pruned_loss=0.05149, over 951952.28 frames. ], batch size: 26, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:58:08,890 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.089e+02 1.579e+02 1.820e+02 2.183e+02 3.562e+02, threshold=3.640e+02, percent-clipped=0.0 +2023-03-27 03:58:18,915 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128676.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:58:40,902 INFO [finetune.py:976] (4/7) Epoch 23, batch 2700, loss[loss=0.2039, simple_loss=0.2731, pruned_loss=0.06737, over 4886.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2465, pruned_loss=0.05092, over 951822.05 frames. ], batch size: 35, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:59:10,625 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8165, 1.0475, 1.8626, 1.7564, 1.6419, 1.5751, 1.6614, 1.7492], + device='cuda:4'), covar=tensor([0.3304, 0.3543, 0.2630, 0.3092, 0.4029, 0.3289, 0.3842, 0.2559], + device='cuda:4'), in_proj_covar=tensor([0.0258, 0.0243, 0.0263, 0.0288, 0.0286, 0.0262, 0.0294, 0.0247], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:59:14,674 INFO [finetune.py:976] (4/7) Epoch 23, batch 2750, loss[loss=0.1514, simple_loss=0.2267, pruned_loss=0.03807, over 4759.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2438, pruned_loss=0.05005, over 951192.83 frames. ], batch size: 28, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 03:59:16,465 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.314e+01 1.516e+02 1.813e+02 2.146e+02 3.615e+02, threshold=3.627e+02, percent-clipped=0.0 +2023-03-27 03:59:21,322 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=128769.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:59:30,033 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=128781.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 03:59:46,804 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8839, 1.7274, 1.5317, 1.9465, 2.3834, 2.0554, 1.5887, 1.5015], + device='cuda:4'), covar=tensor([0.2136, 0.2067, 0.1928, 0.1603, 0.1655, 0.1196, 0.2417, 0.1906], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0211, 0.0214, 0.0197, 0.0245, 0.0190, 0.0218, 0.0206], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 03:59:48,514 INFO [finetune.py:976] (4/7) Epoch 23, batch 2800, loss[loss=0.166, simple_loss=0.2364, pruned_loss=0.04784, over 4859.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2414, pruned_loss=0.04935, over 954001.55 frames. ], batch size: 31, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 04:00:10,919 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=128842.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:00:22,070 INFO [finetune.py:976] (4/7) Epoch 23, batch 2850, loss[loss=0.2226, simple_loss=0.2864, pruned_loss=0.07943, over 4861.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2406, pruned_loss=0.04987, over 955577.49 frames. ], batch size: 47, lr: 3.10e-03, grad_scale: 32.0 +2023-03-27 04:00:23,884 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.021e+01 1.432e+02 1.754e+02 2.169e+02 4.729e+02, threshold=3.508e+02, percent-clipped=1.0 +2023-03-27 04:00:55,220 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=128890.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 04:01:08,142 INFO [finetune.py:976] (4/7) Epoch 23, batch 2900, loss[loss=0.1787, simple_loss=0.2682, pruned_loss=0.04467, over 4808.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2437, pruned_loss=0.05027, over 954410.66 frames. ], batch size: 51, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:01:36,743 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=128951.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 04:01:41,988 INFO [finetune.py:976] (4/7) Epoch 23, batch 2950, loss[loss=0.1686, simple_loss=0.2418, pruned_loss=0.0477, over 4813.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2457, pruned_loss=0.05014, over 956222.36 frames. ], batch size: 33, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:01:43,784 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.806e+01 1.599e+02 2.012e+02 2.314e+02 4.261e+02, threshold=4.024e+02, percent-clipped=1.0 +2023-03-27 04:01:52,341 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=128976.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:01:59,739 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.3496, 2.9436, 2.7295, 1.3224, 2.8714, 2.3297, 2.2838, 2.6852], + device='cuda:4'), covar=tensor([0.0849, 0.0911, 0.1850, 0.2247, 0.1662, 0.2263, 0.2147, 0.1050], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0191, 0.0199, 0.0182, 0.0210, 0.0208, 0.0224, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:02:31,487 INFO [finetune.py:976] (4/7) Epoch 23, batch 3000, loss[loss=0.2219, simple_loss=0.291, pruned_loss=0.07641, over 4721.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2477, pruned_loss=0.05133, over 955192.49 frames. ], batch size: 59, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:02:31,487 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 04:02:36,902 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2257, 2.0807, 1.6040, 0.5601, 1.7601, 1.8757, 1.7598, 1.9642], + device='cuda:4'), covar=tensor([0.0959, 0.0743, 0.1420, 0.2022, 0.1282, 0.2339, 0.2218, 0.0755], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0191, 0.0199, 0.0182, 0.0210, 0.0209, 0.0225, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:02:42,304 INFO [finetune.py:1010] (4/7) Epoch 23, validation: loss=0.1567, simple_loss=0.225, pruned_loss=0.04424, over 2265189.00 frames. +2023-03-27 04:02:42,304 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6489MB +2023-03-27 04:02:51,934 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=129024.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:03:14,563 INFO [finetune.py:976] (4/7) Epoch 23, batch 3050, loss[loss=0.1746, simple_loss=0.2519, pruned_loss=0.04862, over 4769.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2484, pruned_loss=0.05142, over 952383.96 frames. ], batch size: 26, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:03:16,832 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.606e+02 1.899e+02 2.236e+02 5.313e+02, threshold=3.798e+02, percent-clipped=3.0 +2023-03-27 04:03:22,116 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=129069.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:03:47,839 INFO [finetune.py:976] (4/7) Epoch 23, batch 3100, loss[loss=0.1354, simple_loss=0.2069, pruned_loss=0.03197, over 4018.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2455, pruned_loss=0.05064, over 952512.03 frames. ], batch size: 17, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:03:53,280 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=129117.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:04:06,400 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=129137.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:04:13,562 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4897, 1.5144, 1.9436, 1.8719, 1.6553, 3.3847, 1.3304, 1.6067], + device='cuda:4'), covar=tensor([0.0983, 0.1650, 0.1246, 0.0877, 0.1475, 0.0246, 0.1468, 0.1729], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0092, 0.0082, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 04:04:20,606 INFO [finetune.py:976] (4/7) Epoch 23, batch 3150, loss[loss=0.2004, simple_loss=0.2527, pruned_loss=0.07405, over 4820.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2427, pruned_loss=0.05004, over 951962.63 frames. ], batch size: 39, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:04:22,456 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.200e+02 1.503e+02 1.751e+02 2.296e+02 3.694e+02, threshold=3.502e+02, percent-clipped=0.0 +2023-03-27 04:04:25,891 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5598, 1.5884, 1.9960, 1.8965, 1.8183, 3.5842, 1.4267, 1.6206], + device='cuda:4'), covar=tensor([0.1046, 0.1880, 0.1001, 0.0992, 0.1597, 0.0245, 0.1572, 0.1874], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0092, 0.0082, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 04:04:40,793 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3674, 2.2288, 1.8430, 2.3619, 2.3630, 2.1317, 2.6645, 2.4098], + device='cuda:4'), covar=tensor([0.1376, 0.2113, 0.2954, 0.2468, 0.2266, 0.1521, 0.2854, 0.1793], + device='cuda:4'), in_proj_covar=tensor([0.0189, 0.0189, 0.0234, 0.0254, 0.0248, 0.0204, 0.0214, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:05:01,889 INFO [finetune.py:976] (4/7) Epoch 23, batch 3200, loss[loss=0.1317, simple_loss=0.2, pruned_loss=0.03168, over 4703.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2391, pruned_loss=0.04834, over 953646.91 frames. ], batch size: 23, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:05:08,855 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-27 04:05:17,829 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5298, 1.3018, 1.4603, 0.8154, 1.5238, 1.4493, 1.4888, 1.3064], + device='cuda:4'), covar=tensor([0.0587, 0.0932, 0.0792, 0.0979, 0.0929, 0.0831, 0.0671, 0.1282], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0135, 0.0138, 0.0119, 0.0124, 0.0137, 0.0137, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:05:26,800 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=129246.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 04:05:34,306 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.42 vs. limit=5.0 +2023-03-27 04:05:37,439 INFO [finetune.py:976] (4/7) Epoch 23, batch 3250, loss[loss=0.2282, simple_loss=0.2934, pruned_loss=0.08149, over 4167.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2397, pruned_loss=0.04862, over 952834.33 frames. ], batch size: 65, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:05:39,767 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.902e+01 1.487e+02 1.735e+02 2.015e+02 4.622e+02, threshold=3.470e+02, percent-clipped=1.0 +2023-03-27 04:06:15,798 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0246, 1.7589, 2.3400, 1.4628, 2.1728, 2.3033, 1.6532, 2.3972], + device='cuda:4'), covar=tensor([0.1289, 0.2369, 0.1325, 0.2058, 0.0997, 0.1311, 0.3115, 0.0934], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0208, 0.0194, 0.0191, 0.0175, 0.0216, 0.0217, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:06:17,062 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5972, 2.4072, 1.8349, 0.9597, 2.1016, 2.2001, 1.9882, 2.1620], + device='cuda:4'), covar=tensor([0.0724, 0.0726, 0.1549, 0.2014, 0.1245, 0.1801, 0.2048, 0.0766], + device='cuda:4'), in_proj_covar=tensor([0.0172, 0.0193, 0.0201, 0.0184, 0.0212, 0.0211, 0.0226, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:06:22,329 INFO [finetune.py:976] (4/7) Epoch 23, batch 3300, loss[loss=0.2591, simple_loss=0.3315, pruned_loss=0.09334, over 4864.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2428, pruned_loss=0.04983, over 953442.03 frames. ], batch size: 34, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:06:52,049 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7964, 1.2983, 0.8822, 1.7046, 2.0854, 1.5797, 1.5760, 1.7027], + device='cuda:4'), covar=tensor([0.1337, 0.1842, 0.1938, 0.1049, 0.1777, 0.2017, 0.1321, 0.1644], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0111, 0.0092, 0.0119, 0.0094, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 04:06:56,064 INFO [finetune.py:976] (4/7) Epoch 23, batch 3350, loss[loss=0.1632, simple_loss=0.233, pruned_loss=0.0467, over 4699.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2446, pruned_loss=0.05034, over 949897.67 frames. ], batch size: 23, lr: 3.09e-03, grad_scale: 32.0 +2023-03-27 04:06:57,832 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.264e+01 1.616e+02 1.807e+02 2.144e+02 4.365e+02, threshold=3.613e+02, percent-clipped=1.0 +2023-03-27 04:06:59,813 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6985, 1.6016, 1.5751, 0.8801, 1.7154, 1.7590, 1.7330, 1.4641], + device='cuda:4'), covar=tensor([0.0860, 0.0626, 0.0560, 0.0547, 0.0462, 0.0633, 0.0368, 0.0693], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0149, 0.0127, 0.0123, 0.0131, 0.0130, 0.0141, 0.0149], + device='cuda:4'), out_proj_covar=tensor([8.9553e-05, 1.0763e-04, 9.0870e-05, 8.6689e-05, 9.2280e-05, 9.2281e-05, + 1.0076e-04, 1.0661e-04], device='cuda:4') +2023-03-27 04:07:47,796 INFO [finetune.py:976] (4/7) Epoch 23, batch 3400, loss[loss=0.2051, simple_loss=0.2748, pruned_loss=0.06768, over 4895.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2455, pruned_loss=0.0502, over 950571.71 frames. ], batch size: 36, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:07:50,373 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=129413.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 04:08:07,357 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=129437.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:08:21,220 INFO [finetune.py:976] (4/7) Epoch 23, batch 3450, loss[loss=0.1298, simple_loss=0.2026, pruned_loss=0.02848, over 4800.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2454, pruned_loss=0.04998, over 952491.48 frames. ], batch size: 25, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:08:23,467 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.549e+02 1.981e+02 2.372e+02 4.494e+02, threshold=3.962e+02, percent-clipped=6.0 +2023-03-27 04:08:25,360 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8236, 3.9516, 3.7548, 1.8804, 4.1747, 3.2391, 1.1446, 2.8645], + device='cuda:4'), covar=tensor([0.2260, 0.1764, 0.1433, 0.3318, 0.1077, 0.0915, 0.4261, 0.1364], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0178, 0.0160, 0.0129, 0.0161, 0.0123, 0.0148, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 04:08:31,809 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=129474.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 04:08:39,421 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=129485.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:08:54,911 INFO [finetune.py:976] (4/7) Epoch 23, batch 3500, loss[loss=0.1404, simple_loss=0.2118, pruned_loss=0.03443, over 4846.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2438, pruned_loss=0.05016, over 953696.62 frames. ], batch size: 49, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:09:20,954 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=129546.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 04:09:28,786 INFO [finetune.py:976] (4/7) Epoch 23, batch 3550, loss[loss=0.1497, simple_loss=0.2259, pruned_loss=0.03672, over 4900.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2416, pruned_loss=0.04991, over 955089.06 frames. ], batch size: 32, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:09:30,574 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.427e+01 1.468e+02 1.701e+02 2.000e+02 4.470e+02, threshold=3.402e+02, percent-clipped=1.0 +2023-03-27 04:09:35,397 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5372, 1.6217, 1.9203, 1.8536, 1.6862, 3.5567, 1.4472, 1.6827], + device='cuda:4'), covar=tensor([0.0925, 0.1708, 0.1076, 0.0923, 0.1560, 0.0196, 0.1380, 0.1783], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0077, 0.0092, 0.0081, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 04:09:43,362 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-27 04:09:44,242 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1638, 1.9581, 2.0286, 1.4913, 2.0483, 2.1555, 2.1884, 1.6805], + device='cuda:4'), covar=tensor([0.0487, 0.0638, 0.0714, 0.0795, 0.0782, 0.0640, 0.0561, 0.1191], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0135, 0.0138, 0.0119, 0.0124, 0.0137, 0.0137, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:09:52,430 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=129594.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 04:10:11,048 INFO [finetune.py:976] (4/7) Epoch 23, batch 3600, loss[loss=0.1703, simple_loss=0.2383, pruned_loss=0.0511, over 4905.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2389, pruned_loss=0.0493, over 954720.38 frames. ], batch size: 35, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:10:12,978 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=129612.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:10:33,703 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6527, 2.3171, 2.7668, 2.6672, 2.4235, 2.3855, 2.6193, 2.6104], + device='cuda:4'), covar=tensor([0.3940, 0.3732, 0.2734, 0.3391, 0.4435, 0.3435, 0.4421, 0.2696], + device='cuda:4'), in_proj_covar=tensor([0.0258, 0.0243, 0.0264, 0.0287, 0.0286, 0.0262, 0.0295, 0.0247], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:10:38,922 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 04:10:44,784 INFO [finetune.py:976] (4/7) Epoch 23, batch 3650, loss[loss=0.1478, simple_loss=0.2311, pruned_loss=0.03225, over 4914.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2423, pruned_loss=0.05056, over 955452.89 frames. ], batch size: 36, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:10:46,575 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.022e+02 1.535e+02 1.802e+02 2.202e+02 3.404e+02, threshold=3.605e+02, percent-clipped=1.0 +2023-03-27 04:10:53,467 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=129673.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:11:23,359 INFO [finetune.py:976] (4/7) Epoch 23, batch 3700, loss[loss=0.2031, simple_loss=0.2738, pruned_loss=0.06626, over 4764.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2464, pruned_loss=0.05119, over 954781.03 frames. ], batch size: 28, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:11:53,939 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-27 04:12:00,260 INFO [finetune.py:976] (4/7) Epoch 23, batch 3750, loss[loss=0.1981, simple_loss=0.2609, pruned_loss=0.06764, over 4798.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2483, pruned_loss=0.05225, over 952546.47 frames. ], batch size: 51, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:12:02,070 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.237e+01 1.564e+02 1.874e+02 2.284e+02 3.839e+02, threshold=3.748e+02, percent-clipped=2.0 +2023-03-27 04:12:06,379 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=129769.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 04:12:20,813 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=129791.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:12:23,222 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5849, 2.3867, 1.9002, 0.8222, 2.0906, 1.9666, 1.8459, 2.0950], + device='cuda:4'), covar=tensor([0.0866, 0.0845, 0.1698, 0.2143, 0.1364, 0.2324, 0.2242, 0.1014], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0192, 0.0200, 0.0182, 0.0211, 0.0209, 0.0224, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:12:35,164 INFO [finetune.py:976] (4/7) Epoch 23, batch 3800, loss[loss=0.1622, simple_loss=0.2405, pruned_loss=0.04196, over 4912.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2507, pruned_loss=0.05298, over 955356.88 frames. ], batch size: 33, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:13:07,450 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.9079, 3.4551, 3.6277, 3.8260, 3.7010, 3.4486, 3.9929, 1.3630], + device='cuda:4'), covar=tensor([0.0919, 0.0885, 0.1007, 0.0967, 0.1429, 0.1589, 0.0787, 0.5282], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0247, 0.0279, 0.0294, 0.0338, 0.0286, 0.0305, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:13:16,846 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=129852.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:13:18,385 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-27 04:13:21,830 INFO [finetune.py:976] (4/7) Epoch 23, batch 3850, loss[loss=0.1967, simple_loss=0.2549, pruned_loss=0.06931, over 4820.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2483, pruned_loss=0.05211, over 956461.19 frames. ], batch size: 33, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:13:24,149 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.288e+01 1.597e+02 1.881e+02 2.160e+02 3.613e+02, threshold=3.763e+02, percent-clipped=0.0 +2023-03-27 04:13:32,767 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2588, 3.7039, 3.8580, 4.0878, 4.0384, 3.7266, 4.3163, 1.4039], + device='cuda:4'), covar=tensor([0.0721, 0.0847, 0.0896, 0.1023, 0.1099, 0.1565, 0.0692, 0.5642], + device='cuda:4'), in_proj_covar=tensor([0.0346, 0.0247, 0.0279, 0.0293, 0.0338, 0.0285, 0.0305, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:13:55,058 INFO [finetune.py:976] (4/7) Epoch 23, batch 3900, loss[loss=0.1425, simple_loss=0.2216, pruned_loss=0.03173, over 4813.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2455, pruned_loss=0.05093, over 956825.58 frames. ], batch size: 38, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:14:02,804 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4802, 1.3422, 1.6396, 2.4368, 1.6420, 2.1850, 1.0362, 2.0940], + device='cuda:4'), covar=tensor([0.1509, 0.1386, 0.1042, 0.0702, 0.0834, 0.1234, 0.1358, 0.0583], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0163, 0.0099, 0.0136, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 04:14:27,723 INFO [finetune.py:976] (4/7) Epoch 23, batch 3950, loss[loss=0.1319, simple_loss=0.2052, pruned_loss=0.02935, over 4904.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2417, pruned_loss=0.05019, over 957794.32 frames. ], batch size: 35, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:14:29,947 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.481e+02 1.820e+02 2.101e+02 4.779e+02, threshold=3.640e+02, percent-clipped=1.0 +2023-03-27 04:14:34,548 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=129968.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:14:45,377 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.8508, 3.4092, 3.5538, 3.6871, 3.5898, 3.4120, 3.9266, 1.2347], + device='cuda:4'), covar=tensor([0.1020, 0.0957, 0.1062, 0.1340, 0.1590, 0.1747, 0.1023, 0.6140], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0247, 0.0280, 0.0294, 0.0339, 0.0286, 0.0306, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:14:49,651 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6708, 1.5933, 1.9797, 1.9504, 1.7124, 2.9517, 1.3926, 1.6451], + device='cuda:4'), covar=tensor([0.0836, 0.1521, 0.1283, 0.0743, 0.1354, 0.0291, 0.1329, 0.1470], + device='cuda:4'), in_proj_covar=tensor([0.0074, 0.0081, 0.0073, 0.0076, 0.0091, 0.0081, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 04:14:50,991 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.82 vs. limit=5.0 +2023-03-27 04:14:56,355 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130000.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:14:58,076 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130002.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:15:02,811 INFO [finetune.py:976] (4/7) Epoch 23, batch 4000, loss[loss=0.1828, simple_loss=0.2631, pruned_loss=0.05126, over 4842.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2419, pruned_loss=0.05071, over 954880.81 frames. ], batch size: 49, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:15:31,794 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1488, 1.9594, 1.9990, 1.5316, 1.9682, 2.0997, 2.1660, 1.6755], + device='cuda:4'), covar=tensor([0.0465, 0.0546, 0.0676, 0.0788, 0.0788, 0.0666, 0.0540, 0.1055], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0135, 0.0139, 0.0120, 0.0125, 0.0138, 0.0138, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:15:45,415 INFO [finetune.py:976] (4/7) Epoch 23, batch 4050, loss[loss=0.1742, simple_loss=0.2541, pruned_loss=0.04717, over 4895.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2433, pruned_loss=0.05118, over 953545.45 frames. ], batch size: 37, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:15:47,201 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130061.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:15:47,689 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.063e+02 1.665e+02 1.960e+02 2.479e+02 5.275e+02, threshold=3.921e+02, percent-clipped=4.0 +2023-03-27 04:15:48,460 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130063.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 04:15:53,610 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130069.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 04:16:19,204 INFO [finetune.py:976] (4/7) Epoch 23, batch 4100, loss[loss=0.2691, simple_loss=0.3329, pruned_loss=0.1026, over 4141.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2468, pruned_loss=0.05198, over 953573.86 frames. ], batch size: 65, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:16:19,923 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1289, 1.3327, 1.3603, 0.7309, 1.3446, 1.5885, 1.6048, 1.3199], + device='cuda:4'), covar=tensor([0.0983, 0.0768, 0.0580, 0.0587, 0.0601, 0.0673, 0.0390, 0.0801], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0149, 0.0127, 0.0123, 0.0131, 0.0129, 0.0141, 0.0149], + device='cuda:4'), out_proj_covar=tensor([8.9532e-05, 1.0740e-04, 9.1046e-05, 8.6277e-05, 9.2303e-05, 9.2273e-05, + 1.0087e-04, 1.0639e-04], device='cuda:4') +2023-03-27 04:16:26,590 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130117.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 04:16:54,993 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130147.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:17:02,638 INFO [finetune.py:976] (4/7) Epoch 23, batch 4150, loss[loss=0.1782, simple_loss=0.2605, pruned_loss=0.04797, over 4797.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2483, pruned_loss=0.05226, over 953901.88 frames. ], batch size: 45, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:17:04,907 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.036e+02 1.505e+02 1.863e+02 2.291e+02 4.324e+02, threshold=3.726e+02, percent-clipped=3.0 +2023-03-27 04:17:17,647 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-27 04:17:36,689 INFO [finetune.py:976] (4/7) Epoch 23, batch 4200, loss[loss=0.171, simple_loss=0.2472, pruned_loss=0.04742, over 4815.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2471, pruned_loss=0.05128, over 954802.59 frames. ], batch size: 38, lr: 3.09e-03, grad_scale: 64.0 +2023-03-27 04:18:24,042 INFO [finetune.py:976] (4/7) Epoch 23, batch 4250, loss[loss=0.1434, simple_loss=0.2105, pruned_loss=0.03817, over 4866.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2448, pruned_loss=0.0506, over 955576.29 frames. ], batch size: 34, lr: 3.08e-03, grad_scale: 64.0 +2023-03-27 04:18:25,853 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.027e+02 1.516e+02 1.759e+02 2.094e+02 3.793e+02, threshold=3.518e+02, percent-clipped=1.0 +2023-03-27 04:18:30,107 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130268.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:18:57,494 INFO [finetune.py:976] (4/7) Epoch 23, batch 4300, loss[loss=0.1773, simple_loss=0.2418, pruned_loss=0.05642, over 4909.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2423, pruned_loss=0.05, over 954390.21 frames. ], batch size: 37, lr: 3.08e-03, grad_scale: 64.0 +2023-03-27 04:19:02,812 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130316.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:19:09,424 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.0493, 4.4453, 4.3695, 2.1253, 4.4920, 3.5227, 0.8972, 3.0739], + device='cuda:4'), covar=tensor([0.2050, 0.1822, 0.1268, 0.3318, 0.0761, 0.0823, 0.4522, 0.1468], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0178, 0.0160, 0.0128, 0.0161, 0.0123, 0.0148, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 04:19:29,231 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130356.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:19:30,482 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130358.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 04:19:31,027 INFO [finetune.py:976] (4/7) Epoch 23, batch 4350, loss[loss=0.1681, simple_loss=0.237, pruned_loss=0.04954, over 4812.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2399, pruned_loss=0.04971, over 954919.41 frames. ], batch size: 25, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:19:33,425 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.417e+02 1.746e+02 2.112e+02 4.412e+02, threshold=3.492e+02, percent-clipped=1.0 +2023-03-27 04:19:48,215 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130384.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:19:54,680 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-27 04:20:04,340 INFO [finetune.py:976] (4/7) Epoch 23, batch 4400, loss[loss=0.1858, simple_loss=0.2673, pruned_loss=0.05222, over 4811.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2415, pruned_loss=0.05033, over 955057.65 frames. ], batch size: 51, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:20:19,495 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130432.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:20:24,169 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130438.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:20:30,584 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130445.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:20:31,768 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130447.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:20:39,201 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8938, 1.7988, 1.5850, 2.0840, 2.4146, 2.0645, 1.7827, 1.5067], + device='cuda:4'), covar=tensor([0.2098, 0.1941, 0.1903, 0.1503, 0.1609, 0.1125, 0.2178, 0.1913], + device='cuda:4'), in_proj_covar=tensor([0.0245, 0.0209, 0.0212, 0.0196, 0.0243, 0.0189, 0.0215, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:20:42,276 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1524, 1.8562, 2.2311, 2.1805, 1.9259, 1.8866, 2.1388, 2.0782], + device='cuda:4'), covar=tensor([0.4055, 0.4273, 0.3148, 0.4021, 0.5040, 0.4244, 0.4825, 0.3090], + device='cuda:4'), in_proj_covar=tensor([0.0257, 0.0242, 0.0263, 0.0286, 0.0285, 0.0261, 0.0293, 0.0247], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:20:46,863 INFO [finetune.py:976] (4/7) Epoch 23, batch 4450, loss[loss=0.1677, simple_loss=0.24, pruned_loss=0.04768, over 4748.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2446, pruned_loss=0.0508, over 957298.23 frames. ], batch size: 27, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:20:49,236 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.124e+02 1.491e+02 1.813e+02 2.246e+02 3.707e+02, threshold=3.626e+02, percent-clipped=3.0 +2023-03-27 04:21:10,566 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130493.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:21:11,732 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130495.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:21:14,692 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130499.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:21:20,677 INFO [finetune.py:976] (4/7) Epoch 23, batch 4500, loss[loss=0.1641, simple_loss=0.2401, pruned_loss=0.04405, over 4233.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2461, pruned_loss=0.05109, over 956690.61 frames. ], batch size: 65, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:21:53,320 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130548.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:22:04,078 INFO [finetune.py:976] (4/7) Epoch 23, batch 4550, loss[loss=0.1874, simple_loss=0.2621, pruned_loss=0.05639, over 4789.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2472, pruned_loss=0.05069, over 955790.80 frames. ], batch size: 51, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:22:06,503 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.768e+01 1.546e+02 1.777e+02 2.233e+02 3.779e+02, threshold=3.553e+02, percent-clipped=2.0 +2023-03-27 04:22:07,817 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=130565.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:22:37,456 INFO [finetune.py:976] (4/7) Epoch 23, batch 4600, loss[loss=0.1578, simple_loss=0.2203, pruned_loss=0.04761, over 4767.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.247, pruned_loss=0.05066, over 956171.69 frames. ], batch size: 27, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:22:37,566 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130609.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 04:22:47,785 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=130626.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:23:10,861 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130656.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:23:17,055 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=130658.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 04:23:17,582 INFO [finetune.py:976] (4/7) Epoch 23, batch 4650, loss[loss=0.1602, simple_loss=0.2316, pruned_loss=0.04441, over 4719.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2452, pruned_loss=0.05058, over 955809.84 frames. ], batch size: 54, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:23:19,982 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.906e+01 1.515e+02 1.768e+02 2.232e+02 6.495e+02, threshold=3.536e+02, percent-clipped=3.0 +2023-03-27 04:23:54,690 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130704.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:23:55,884 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=130706.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:23:58,153 INFO [finetune.py:976] (4/7) Epoch 23, batch 4700, loss[loss=0.1437, simple_loss=0.2212, pruned_loss=0.03312, over 4761.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2436, pruned_loss=0.05055, over 956806.04 frames. ], batch size: 26, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:24:18,053 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130740.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:24:31,369 INFO [finetune.py:976] (4/7) Epoch 23, batch 4750, loss[loss=0.1843, simple_loss=0.2609, pruned_loss=0.05389, over 4811.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2411, pruned_loss=0.05005, over 954789.29 frames. ], batch size: 38, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:24:34,233 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.253e+02 1.528e+02 1.803e+02 2.150e+02 3.686e+02, threshold=3.606e+02, percent-clipped=2.0 +2023-03-27 04:24:49,884 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130788.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:24:52,208 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5784, 1.6447, 2.2593, 1.8455, 1.8079, 4.2200, 1.5238, 1.6979], + device='cuda:4'), covar=tensor([0.0985, 0.1732, 0.1318, 0.1042, 0.1584, 0.0250, 0.1513, 0.1854], + device='cuda:4'), in_proj_covar=tensor([0.0074, 0.0081, 0.0073, 0.0076, 0.0091, 0.0081, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 04:24:54,009 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130794.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:25:04,662 INFO [finetune.py:976] (4/7) Epoch 23, batch 4800, loss[loss=0.2169, simple_loss=0.2857, pruned_loss=0.07402, over 4835.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2441, pruned_loss=0.05121, over 955410.56 frames. ], batch size: 49, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:25:24,290 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.48 vs. limit=5.0 +2023-03-27 04:25:37,286 INFO [finetune.py:976] (4/7) Epoch 23, batch 4850, loss[loss=0.1643, simple_loss=0.2348, pruned_loss=0.04691, over 4751.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2481, pruned_loss=0.05242, over 954403.02 frames. ], batch size: 59, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:25:40,093 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.613e+02 1.947e+02 2.336e+02 6.046e+02, threshold=3.894e+02, percent-clipped=4.0 +2023-03-27 04:26:15,664 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130904.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 04:26:19,114 INFO [finetune.py:976] (4/7) Epoch 23, batch 4900, loss[loss=0.1908, simple_loss=0.2643, pruned_loss=0.05861, over 4793.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.2486, pruned_loss=0.05222, over 951859.50 frames. ], batch size: 51, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:26:25,709 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-27 04:26:28,433 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=130921.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:26:52,302 INFO [finetune.py:976] (4/7) Epoch 23, batch 4950, loss[loss=0.1786, simple_loss=0.2602, pruned_loss=0.0485, over 4812.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2496, pruned_loss=0.0527, over 951465.27 frames. ], batch size: 38, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:26:57,595 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.290e+01 1.586e+02 1.789e+02 2.374e+02 3.586e+02, threshold=3.578e+02, percent-clipped=0.0 +2023-03-27 04:27:20,094 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5718, 2.4130, 2.0354, 1.0238, 2.1001, 1.9667, 1.8403, 2.1672], + device='cuda:4'), covar=tensor([0.0824, 0.0807, 0.1734, 0.2110, 0.1492, 0.2202, 0.2163, 0.1089], + device='cuda:4'), in_proj_covar=tensor([0.0171, 0.0192, 0.0201, 0.0182, 0.0210, 0.0209, 0.0225, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:27:36,350 INFO [finetune.py:976] (4/7) Epoch 23, batch 5000, loss[loss=0.1671, simple_loss=0.2274, pruned_loss=0.05338, over 4833.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2471, pruned_loss=0.05192, over 952088.07 frames. ], batch size: 25, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:27:57,576 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131040.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:28:07,649 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1021, 2.0603, 1.7592, 1.9597, 1.9167, 1.9616, 1.9715, 2.6831], + device='cuda:4'), covar=tensor([0.3729, 0.4178, 0.3190, 0.3907, 0.4101, 0.2381, 0.3742, 0.1615], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0261, 0.0232, 0.0274, 0.0254, 0.0225, 0.0252, 0.0234], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:28:09,930 INFO [finetune.py:976] (4/7) Epoch 23, batch 5050, loss[loss=0.1606, simple_loss=0.225, pruned_loss=0.04814, over 4870.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2439, pruned_loss=0.05097, over 951761.63 frames. ], batch size: 49, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:28:12,371 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 1.381e+02 1.770e+02 2.059e+02 4.416e+02, threshold=3.539e+02, percent-clipped=4.0 +2023-03-27 04:28:41,480 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131088.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:28:41,508 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131088.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:28:45,134 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131094.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:28:53,722 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-27 04:28:57,922 INFO [finetune.py:976] (4/7) Epoch 23, batch 5100, loss[loss=0.1541, simple_loss=0.2178, pruned_loss=0.04521, over 4784.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.24, pruned_loss=0.0494, over 951667.16 frames. ], batch size: 26, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:29:01,955 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-27 04:29:05,252 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-27 04:29:17,263 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131136.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:29:19,148 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5491, 1.5429, 1.2955, 1.6151, 1.9049, 1.7683, 1.5741, 1.3309], + device='cuda:4'), covar=tensor([0.0375, 0.0382, 0.0605, 0.0316, 0.0250, 0.0544, 0.0364, 0.0522], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0107, 0.0144, 0.0111, 0.0100, 0.0111, 0.0101, 0.0112], + device='cuda:4'), out_proj_covar=tensor([7.7289e-05, 8.1733e-05, 1.1278e-04, 8.5361e-05, 7.7707e-05, 8.2375e-05, + 7.5356e-05, 8.5504e-05], device='cuda:4') +2023-03-27 04:29:20,884 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131142.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:29:31,085 INFO [finetune.py:976] (4/7) Epoch 23, batch 5150, loss[loss=0.2085, simple_loss=0.2696, pruned_loss=0.07372, over 4873.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2422, pruned_loss=0.05095, over 949050.33 frames. ], batch size: 34, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:29:34,466 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.214e+01 1.572e+02 1.903e+02 2.241e+02 4.010e+02, threshold=3.805e+02, percent-clipped=1.0 +2023-03-27 04:29:41,799 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131174.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:29:58,216 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-27 04:30:01,315 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131204.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 04:30:04,269 INFO [finetune.py:976] (4/7) Epoch 23, batch 5200, loss[loss=0.2003, simple_loss=0.2736, pruned_loss=0.06347, over 4816.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2446, pruned_loss=0.05157, over 948638.90 frames. ], batch size: 45, lr: 3.08e-03, grad_scale: 32.0 +2023-03-27 04:30:11,444 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3709, 1.4757, 1.4822, 0.9979, 1.5387, 1.7622, 1.7888, 1.3491], + device='cuda:4'), covar=tensor([0.0828, 0.0555, 0.0512, 0.0444, 0.0415, 0.0534, 0.0298, 0.0642], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0149, 0.0127, 0.0123, 0.0131, 0.0130, 0.0141, 0.0149], + device='cuda:4'), out_proj_covar=tensor([8.9536e-05, 1.0731e-04, 9.0613e-05, 8.6726e-05, 9.2235e-05, 9.2595e-05, + 1.0097e-04, 1.0652e-04], device='cuda:4') +2023-03-27 04:30:11,467 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0506, 1.9294, 1.6343, 1.7228, 1.8049, 1.7989, 1.8665, 2.5638], + device='cuda:4'), covar=tensor([0.3597, 0.4076, 0.3077, 0.3719, 0.3767, 0.2402, 0.3697, 0.1633], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0261, 0.0233, 0.0274, 0.0254, 0.0225, 0.0253, 0.0235], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:30:12,596 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131221.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:30:22,993 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=131235.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 04:30:33,243 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131252.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:30:37,394 INFO [finetune.py:976] (4/7) Epoch 23, batch 5250, loss[loss=0.1783, simple_loss=0.2514, pruned_loss=0.05261, over 4819.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.247, pruned_loss=0.05235, over 947530.19 frames. ], batch size: 33, lr: 3.08e-03, grad_scale: 16.0 +2023-03-27 04:30:40,888 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.531e+02 1.792e+02 2.239e+02 3.281e+02, threshold=3.585e+02, percent-clipped=0.0 +2023-03-27 04:30:44,382 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131269.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:31:21,563 INFO [finetune.py:976] (4/7) Epoch 23, batch 5300, loss[loss=0.1811, simple_loss=0.2601, pruned_loss=0.05104, over 4897.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2485, pruned_loss=0.05279, over 948535.88 frames. ], batch size: 32, lr: 3.08e-03, grad_scale: 16.0 +2023-03-27 04:31:22,912 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5384, 1.3880, 1.2763, 1.4491, 1.7808, 1.8027, 1.4678, 1.2825], + device='cuda:4'), covar=tensor([0.0380, 0.0397, 0.0653, 0.0322, 0.0257, 0.0439, 0.0428, 0.0500], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0107, 0.0145, 0.0112, 0.0100, 0.0112, 0.0102, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.7610e-05, 8.1814e-05, 1.1332e-04, 8.5527e-05, 7.7735e-05, 8.2631e-05, + 7.5849e-05, 8.5857e-05], device='cuda:4') +2023-03-27 04:31:37,936 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-27 04:31:54,362 INFO [finetune.py:976] (4/7) Epoch 23, batch 5350, loss[loss=0.139, simple_loss=0.2047, pruned_loss=0.03661, over 4252.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2491, pruned_loss=0.05276, over 947784.99 frames. ], batch size: 18, lr: 3.08e-03, grad_scale: 16.0 +2023-03-27 04:31:57,385 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.087e+02 1.529e+02 1.830e+02 2.196e+02 3.219e+02, threshold=3.659e+02, percent-clipped=0.0 +2023-03-27 04:32:38,097 INFO [finetune.py:976] (4/7) Epoch 23, batch 5400, loss[loss=0.176, simple_loss=0.2545, pruned_loss=0.04881, over 4788.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2474, pruned_loss=0.05243, over 947618.11 frames. ], batch size: 51, lr: 3.08e-03, grad_scale: 16.0 +2023-03-27 04:32:38,213 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131409.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:32:42,422 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131416.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:33:11,756 INFO [finetune.py:976] (4/7) Epoch 23, batch 5450, loss[loss=0.1402, simple_loss=0.2186, pruned_loss=0.03089, over 4845.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2451, pruned_loss=0.0521, over 950528.14 frames. ], batch size: 30, lr: 3.08e-03, grad_scale: 16.0 +2023-03-27 04:33:11,846 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.3467, 3.8079, 3.9583, 4.1753, 4.1258, 3.8572, 4.4469, 1.5015], + device='cuda:4'), covar=tensor([0.0905, 0.0891, 0.1069, 0.1170, 0.1343, 0.1853, 0.0740, 0.5998], + device='cuda:4'), in_proj_covar=tensor([0.0349, 0.0247, 0.0280, 0.0295, 0.0339, 0.0287, 0.0305, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:33:14,785 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.514e+02 1.875e+02 2.409e+02 5.439e+02, threshold=3.749e+02, percent-clipped=4.0 +2023-03-27 04:33:18,556 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=131470.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:33:23,311 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=131477.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:33:33,330 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131492.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:33:51,831 INFO [finetune.py:976] (4/7) Epoch 23, batch 5500, loss[loss=0.1937, simple_loss=0.2774, pruned_loss=0.05496, over 4906.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2409, pruned_loss=0.05059, over 950349.51 frames. ], batch size: 43, lr: 3.08e-03, grad_scale: 16.0 +2023-03-27 04:34:12,700 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=131530.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 04:34:29,174 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=131553.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:34:33,175 INFO [finetune.py:976] (4/7) Epoch 23, batch 5550, loss[loss=0.1407, simple_loss=0.2089, pruned_loss=0.03625, over 4222.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2419, pruned_loss=0.05072, over 952374.07 frames. ], batch size: 18, lr: 3.08e-03, grad_scale: 16.0 +2023-03-27 04:34:36,709 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.823e+01 1.422e+02 1.728e+02 2.186e+02 5.215e+02, threshold=3.457e+02, percent-clipped=2.0 +2023-03-27 04:34:46,354 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-03-27 04:35:04,722 INFO [finetune.py:976] (4/7) Epoch 23, batch 5600, loss[loss=0.1586, simple_loss=0.2352, pruned_loss=0.04096, over 4822.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2459, pruned_loss=0.05207, over 953086.70 frames. ], batch size: 25, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:35:34,625 INFO [finetune.py:976] (4/7) Epoch 23, batch 5650, loss[loss=0.1771, simple_loss=0.2637, pruned_loss=0.04526, over 4826.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.247, pruned_loss=0.05163, over 954073.38 frames. ], batch size: 33, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:35:37,861 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.955e+01 1.494e+02 1.801e+02 2.339e+02 4.576e+02, threshold=3.601e+02, percent-clipped=4.0 +2023-03-27 04:35:39,126 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5778, 1.4332, 1.3837, 1.5307, 1.7944, 1.7184, 1.5608, 1.3019], + device='cuda:4'), covar=tensor([0.0327, 0.0342, 0.0617, 0.0298, 0.0220, 0.0447, 0.0304, 0.0433], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0106, 0.0143, 0.0110, 0.0099, 0.0111, 0.0101, 0.0112], + device='cuda:4'), out_proj_covar=tensor([7.6967e-05, 8.0976e-05, 1.1196e-04, 8.4464e-05, 7.6900e-05, 8.2251e-05, + 7.4904e-05, 8.4943e-05], device='cuda:4') +2023-03-27 04:35:43,939 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.2954, 1.2478, 1.2181, 0.7479, 1.2347, 1.4183, 1.4483, 1.1738], + device='cuda:4'), covar=tensor([0.0924, 0.0609, 0.0505, 0.0517, 0.0608, 0.0504, 0.0365, 0.0757], + device='cuda:4'), in_proj_covar=tensor([0.0121, 0.0147, 0.0125, 0.0122, 0.0130, 0.0129, 0.0140, 0.0147], + device='cuda:4'), out_proj_covar=tensor([8.8448e-05, 1.0627e-04, 8.9468e-05, 8.5615e-05, 9.1301e-05, 9.1540e-05, + 9.9759e-05, 1.0532e-04], device='cuda:4') +2023-03-27 04:36:04,503 INFO [finetune.py:976] (4/7) Epoch 23, batch 5700, loss[loss=0.1405, simple_loss=0.1954, pruned_loss=0.04285, over 4161.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2453, pruned_loss=0.0514, over 939640.92 frames. ], batch size: 18, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:36:26,696 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2214, 1.2916, 1.4548, 1.0948, 1.2111, 1.4568, 1.2766, 1.5183], + device='cuda:4'), covar=tensor([0.1053, 0.1984, 0.1276, 0.1292, 0.1054, 0.1129, 0.2688, 0.0819], + device='cuda:4'), in_proj_covar=tensor([0.0190, 0.0205, 0.0190, 0.0189, 0.0172, 0.0212, 0.0215, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:36:40,049 INFO [finetune.py:976] (4/7) Epoch 24, batch 0, loss[loss=0.1657, simple_loss=0.2301, pruned_loss=0.05061, over 4777.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2301, pruned_loss=0.05061, over 4777.00 frames. ], batch size: 26, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:36:40,049 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 04:36:43,072 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2298, 2.0101, 1.9231, 1.8475, 1.9464, 2.0882, 1.9948, 2.6437], + device='cuda:4'), covar=tensor([0.3732, 0.4535, 0.3242, 0.3806, 0.4181, 0.2331, 0.3703, 0.1918], + device='cuda:4'), in_proj_covar=tensor([0.0290, 0.0263, 0.0234, 0.0275, 0.0256, 0.0226, 0.0253, 0.0235], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:36:49,369 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1244, 1.9363, 1.7603, 1.7933, 1.8925, 1.9132, 1.8696, 2.5810], + device='cuda:4'), covar=tensor([0.4169, 0.4890, 0.3455, 0.4122, 0.4208, 0.2628, 0.4347, 0.1930], + device='cuda:4'), in_proj_covar=tensor([0.0290, 0.0263, 0.0234, 0.0275, 0.0256, 0.0226, 0.0253, 0.0235], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:36:50,762 INFO [finetune.py:1010] (4/7) Epoch 24, validation: loss=0.1594, simple_loss=0.227, pruned_loss=0.04592, over 2265189.00 frames. +2023-03-27 04:36:50,762 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6557MB +2023-03-27 04:36:54,265 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0606, 1.2535, 1.4881, 1.2023, 1.4091, 2.4033, 1.2496, 1.3903], + device='cuda:4'), covar=tensor([0.0997, 0.1903, 0.1014, 0.0996, 0.1744, 0.0421, 0.1568, 0.1892], + device='cuda:4'), in_proj_covar=tensor([0.0073, 0.0081, 0.0072, 0.0076, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 04:37:07,466 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 7.206e+01 1.398e+02 1.674e+02 2.004e+02 3.219e+02, threshold=3.348e+02, percent-clipped=0.0 +2023-03-27 04:37:08,164 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=131765.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:37:12,919 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=131772.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:37:25,291 INFO [finetune.py:976] (4/7) Epoch 24, batch 50, loss[loss=0.1817, simple_loss=0.2514, pruned_loss=0.056, over 4820.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2455, pruned_loss=0.05207, over 215614.39 frames. ], batch size: 30, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:38:02,597 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=131830.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:38:07,276 INFO [finetune.py:976] (4/7) Epoch 24, batch 100, loss[loss=0.1612, simple_loss=0.227, pruned_loss=0.04768, over 4844.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2384, pruned_loss=0.04858, over 379918.13 frames. ], batch size: 47, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:38:15,493 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=131848.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:38:17,981 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8913, 1.5642, 1.8630, 1.3895, 1.8443, 1.9103, 1.8795, 1.3266], + device='cuda:4'), covar=tensor([0.0521, 0.0900, 0.0648, 0.0824, 0.0853, 0.0638, 0.0626, 0.1656], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0135, 0.0138, 0.0119, 0.0125, 0.0137, 0.0137, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:38:20,990 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=131857.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:38:25,098 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.662e+01 1.465e+02 1.761e+02 2.142e+02 3.724e+02, threshold=3.523e+02, percent-clipped=1.0 +2023-03-27 04:38:34,600 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=131878.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:38:40,502 INFO [finetune.py:976] (4/7) Epoch 24, batch 150, loss[loss=0.1446, simple_loss=0.2179, pruned_loss=0.03571, over 4761.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2343, pruned_loss=0.04736, over 507654.91 frames. ], batch size: 27, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:39:10,819 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=131918.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:39:29,543 INFO [finetune.py:976] (4/7) Epoch 24, batch 200, loss[loss=0.1656, simple_loss=0.226, pruned_loss=0.0526, over 4678.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2381, pruned_loss=0.0502, over 607998.88 frames. ], batch size: 23, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:39:43,438 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3222, 2.3242, 1.9095, 2.3229, 2.1351, 2.1263, 2.0966, 3.0921], + device='cuda:4'), covar=tensor([0.3529, 0.4885, 0.3320, 0.4343, 0.4337, 0.2403, 0.4336, 0.1649], + device='cuda:4'), in_proj_covar=tensor([0.0290, 0.0263, 0.0234, 0.0275, 0.0256, 0.0227, 0.0254, 0.0236], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:39:51,183 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.027e+02 1.517e+02 1.799e+02 2.123e+02 6.232e+02, threshold=3.598e+02, percent-clipped=3.0 +2023-03-27 04:40:06,649 INFO [finetune.py:976] (4/7) Epoch 24, batch 250, loss[loss=0.2188, simple_loss=0.2974, pruned_loss=0.07009, over 4813.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2424, pruned_loss=0.05134, over 682893.53 frames. ], batch size: 41, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:40:36,975 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=132031.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:40:41,436 INFO [finetune.py:976] (4/7) Epoch 24, batch 300, loss[loss=0.1605, simple_loss=0.2402, pruned_loss=0.04045, over 4895.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2439, pruned_loss=0.05121, over 742240.87 frames. ], batch size: 43, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:40:53,252 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=132054.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:40:56,290 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6631, 2.4705, 2.2819, 2.5777, 2.3285, 2.3990, 2.3181, 3.0549], + device='cuda:4'), covar=tensor([0.3316, 0.4040, 0.2977, 0.3266, 0.3708, 0.2335, 0.3495, 0.1621], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0263, 0.0234, 0.0275, 0.0256, 0.0226, 0.0253, 0.0236], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:40:59,164 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.635e+02 1.887e+02 2.261e+02 6.512e+02, threshold=3.774e+02, percent-clipped=2.0 +2023-03-27 04:40:59,882 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132065.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:41:04,186 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132072.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:41:14,120 INFO [finetune.py:976] (4/7) Epoch 24, batch 350, loss[loss=0.1972, simple_loss=0.2622, pruned_loss=0.06606, over 4735.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2463, pruned_loss=0.05179, over 790444.93 frames. ], batch size: 54, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:41:17,775 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=132092.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:41:33,450 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132113.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:41:34,741 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=132115.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 04:41:40,376 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1724, 1.3344, 1.3898, 0.6987, 1.3319, 1.5643, 1.6108, 1.2739], + device='cuda:4'), covar=tensor([0.0902, 0.0573, 0.0538, 0.0487, 0.0487, 0.0586, 0.0313, 0.0751], + device='cuda:4'), in_proj_covar=tensor([0.0121, 0.0147, 0.0124, 0.0120, 0.0129, 0.0128, 0.0139, 0.0146], + device='cuda:4'), out_proj_covar=tensor([8.7834e-05, 1.0559e-04, 8.8833e-05, 8.4564e-05, 9.0815e-05, 9.0859e-05, + 9.8914e-05, 1.0472e-04], device='cuda:4') +2023-03-27 04:41:42,166 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132120.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:41:56,060 INFO [finetune.py:976] (4/7) Epoch 24, batch 400, loss[loss=0.2146, simple_loss=0.2875, pruned_loss=0.0708, over 4818.00 frames. ], tot_loss[loss=0.1769, simple_loss=0.2488, pruned_loss=0.05247, over 829008.17 frames. ], batch size: 38, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:42:03,867 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132148.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:42:11,268 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1588, 1.8647, 2.4491, 4.2524, 2.9680, 2.7830, 1.0697, 3.5055], + device='cuda:4'), covar=tensor([0.1621, 0.1441, 0.1453, 0.0616, 0.0698, 0.1548, 0.1949, 0.0398], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0162, 0.0100, 0.0136, 0.0124, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 04:42:15,409 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.594e+02 1.883e+02 2.349e+02 4.739e+02, threshold=3.766e+02, percent-clipped=1.0 +2023-03-27 04:42:29,844 INFO [finetune.py:976] (4/7) Epoch 24, batch 450, loss[loss=0.152, simple_loss=0.2352, pruned_loss=0.03435, over 4762.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.247, pruned_loss=0.0516, over 858415.65 frames. ], batch size: 28, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:42:36,329 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132196.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:42:38,165 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8880, 2.0252, 1.7585, 1.8271, 2.4269, 2.4719, 2.1129, 1.9622], + device='cuda:4'), covar=tensor([0.0409, 0.0336, 0.0548, 0.0297, 0.0251, 0.0364, 0.0277, 0.0372], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0105, 0.0142, 0.0110, 0.0098, 0.0110, 0.0100, 0.0110], + device='cuda:4'), out_proj_covar=tensor([7.6328e-05, 8.0439e-05, 1.1112e-04, 8.3876e-05, 7.6321e-05, 8.1177e-05, + 7.4037e-05, 8.3915e-05], device='cuda:4') +2023-03-27 04:42:55,015 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=132213.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:43:13,242 INFO [finetune.py:976] (4/7) Epoch 24, batch 500, loss[loss=0.1616, simple_loss=0.2378, pruned_loss=0.04269, over 4747.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2446, pruned_loss=0.05116, over 880800.39 frames. ], batch size: 27, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:43:32,461 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 7.902e+01 1.519e+02 1.809e+02 2.135e+02 3.897e+02, threshold=3.617e+02, percent-clipped=1.0 +2023-03-27 04:43:46,924 INFO [finetune.py:976] (4/7) Epoch 24, batch 550, loss[loss=0.1427, simple_loss=0.2083, pruned_loss=0.03859, over 4833.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2418, pruned_loss=0.05072, over 895672.28 frames. ], batch size: 25, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:44:30,205 INFO [finetune.py:976] (4/7) Epoch 24, batch 600, loss[loss=0.1848, simple_loss=0.2532, pruned_loss=0.05822, over 4917.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2429, pruned_loss=0.05112, over 910782.74 frames. ], batch size: 37, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:44:58,205 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.737e+01 1.576e+02 1.859e+02 2.330e+02 3.343e+02, threshold=3.718e+02, percent-clipped=0.0 +2023-03-27 04:45:07,800 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3060, 2.9174, 2.7830, 1.2811, 3.0600, 2.3688, 0.6003, 1.9132], + device='cuda:4'), covar=tensor([0.2338, 0.2882, 0.1924, 0.3692, 0.1421, 0.1120, 0.4540, 0.1895], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0179, 0.0161, 0.0129, 0.0160, 0.0123, 0.0149, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 04:45:12,586 INFO [finetune.py:976] (4/7) Epoch 24, batch 650, loss[loss=0.157, simple_loss=0.2339, pruned_loss=0.04006, over 4920.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2459, pruned_loss=0.0522, over 919392.55 frames. ], batch size: 36, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:45:12,660 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=132387.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:45:28,118 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=132410.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 04:45:46,140 INFO [finetune.py:976] (4/7) Epoch 24, batch 700, loss[loss=0.1463, simple_loss=0.2352, pruned_loss=0.02867, over 4785.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2475, pruned_loss=0.05239, over 928687.96 frames. ], batch size: 29, lr: 3.07e-03, grad_scale: 16.0 +2023-03-27 04:46:03,852 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.085e+02 1.592e+02 1.911e+02 2.262e+02 4.191e+02, threshold=3.822e+02, percent-clipped=2.0 +2023-03-27 04:46:19,326 INFO [finetune.py:976] (4/7) Epoch 24, batch 750, loss[loss=0.1497, simple_loss=0.2269, pruned_loss=0.03629, over 4807.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2486, pruned_loss=0.05246, over 936186.20 frames. ], batch size: 25, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:46:36,503 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132513.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:46:58,682 INFO [finetune.py:976] (4/7) Epoch 24, batch 800, loss[loss=0.1745, simple_loss=0.2379, pruned_loss=0.05561, over 4703.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2481, pruned_loss=0.05186, over 940246.15 frames. ], batch size: 59, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:47:17,622 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132561.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:47:19,958 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.775e+01 1.480e+02 1.729e+02 2.075e+02 4.531e+02, threshold=3.459e+02, percent-clipped=1.0 +2023-03-27 04:47:35,955 INFO [finetune.py:976] (4/7) Epoch 24, batch 850, loss[loss=0.1781, simple_loss=0.2388, pruned_loss=0.0587, over 4823.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2455, pruned_loss=0.05116, over 943479.16 frames. ], batch size: 38, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:48:18,619 INFO [finetune.py:976] (4/7) Epoch 24, batch 900, loss[loss=0.1955, simple_loss=0.2409, pruned_loss=0.07509, over 4058.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2426, pruned_loss=0.04971, over 946158.12 frames. ], batch size: 65, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:48:33,136 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=132660.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:48:35,407 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.417e+02 1.718e+02 2.002e+02 3.598e+02, threshold=3.436e+02, percent-clipped=1.0 +2023-03-27 04:48:52,530 INFO [finetune.py:976] (4/7) Epoch 24, batch 950, loss[loss=0.1576, simple_loss=0.2213, pruned_loss=0.04695, over 4353.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2409, pruned_loss=0.0492, over 949421.44 frames. ], batch size: 65, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:48:52,614 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132687.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:49:07,061 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=132710.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 04:49:08,296 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7002, 1.5739, 2.0730, 1.3430, 1.8531, 2.0396, 1.4705, 2.1629], + device='cuda:4'), covar=tensor([0.1295, 0.2154, 0.1438, 0.1910, 0.0903, 0.1365, 0.2943, 0.0860], + device='cuda:4'), in_proj_covar=tensor([0.0191, 0.0207, 0.0191, 0.0191, 0.0172, 0.0214, 0.0216, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:49:14,300 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=132721.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:49:26,440 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132735.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:49:27,539 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7395, 3.7221, 3.5557, 1.7229, 3.8168, 2.9982, 0.9026, 2.4715], + device='cuda:4'), covar=tensor([0.2524, 0.1906, 0.1543, 0.3387, 0.0955, 0.0910, 0.4143, 0.1569], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0179, 0.0162, 0.0129, 0.0161, 0.0124, 0.0148, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 04:49:28,061 INFO [finetune.py:976] (4/7) Epoch 24, batch 1000, loss[loss=0.2061, simple_loss=0.2837, pruned_loss=0.06419, over 4798.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2433, pruned_loss=0.05033, over 950209.29 frames. ], batch size: 51, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:49:38,673 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=132746.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:49:50,821 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=132758.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:49:54,881 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.132e+02 1.613e+02 1.803e+02 2.355e+02 4.590e+02, threshold=3.605e+02, percent-clipped=3.0 +2023-03-27 04:50:17,567 INFO [finetune.py:976] (4/7) Epoch 24, batch 1050, loss[loss=0.1974, simple_loss=0.2739, pruned_loss=0.06047, over 4816.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2461, pruned_loss=0.05057, over 951883.86 frames. ], batch size: 38, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:50:31,467 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=132807.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:50:51,432 INFO [finetune.py:976] (4/7) Epoch 24, batch 1100, loss[loss=0.2023, simple_loss=0.2629, pruned_loss=0.07087, over 4898.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2475, pruned_loss=0.05108, over 952676.60 frames. ], batch size: 32, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:51:08,748 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.623e+02 1.902e+02 2.304e+02 4.937e+02, threshold=3.804e+02, percent-clipped=2.0 +2023-03-27 04:51:24,185 INFO [finetune.py:976] (4/7) Epoch 24, batch 1150, loss[loss=0.1556, simple_loss=0.2319, pruned_loss=0.03959, over 4886.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2484, pruned_loss=0.0514, over 954464.51 frames. ], batch size: 35, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:51:57,324 INFO [finetune.py:976] (4/7) Epoch 24, batch 1200, loss[loss=0.2162, simple_loss=0.2746, pruned_loss=0.0789, over 4179.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2465, pruned_loss=0.05098, over 954407.60 frames. ], batch size: 65, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:52:24,195 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1902, 3.6707, 3.8238, 3.9944, 3.9749, 3.6900, 4.2615, 1.4064], + device='cuda:4'), covar=tensor([0.0722, 0.0898, 0.0868, 0.0967, 0.1082, 0.1436, 0.0727, 0.5521], + device='cuda:4'), in_proj_covar=tensor([0.0348, 0.0248, 0.0282, 0.0294, 0.0340, 0.0288, 0.0307, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:52:24,716 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.055e+01 1.470e+02 1.716e+02 2.148e+02 3.548e+02, threshold=3.432e+02, percent-clipped=0.0 +2023-03-27 04:52:28,525 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1396, 2.0878, 1.6962, 2.0734, 2.1591, 1.8731, 2.4129, 2.1357], + device='cuda:4'), covar=tensor([0.1259, 0.1835, 0.2785, 0.2260, 0.2303, 0.1540, 0.2616, 0.1629], + device='cuda:4'), in_proj_covar=tensor([0.0186, 0.0188, 0.0233, 0.0251, 0.0246, 0.0204, 0.0212, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:52:40,278 INFO [finetune.py:976] (4/7) Epoch 24, batch 1250, loss[loss=0.1899, simple_loss=0.2559, pruned_loss=0.06192, over 4816.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2442, pruned_loss=0.05038, over 954961.71 frames. ], batch size: 41, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:52:41,321 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.72 vs. limit=5.0 +2023-03-27 04:52:59,660 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133016.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:53:04,045 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2863, 2.1651, 1.7380, 2.1869, 2.1422, 1.9201, 2.5353, 2.2885], + device='cuda:4'), covar=tensor([0.1255, 0.2110, 0.2643, 0.2419, 0.2350, 0.1596, 0.2948, 0.1497], + device='cuda:4'), in_proj_covar=tensor([0.0186, 0.0187, 0.0233, 0.0250, 0.0246, 0.0203, 0.0211, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:53:15,461 INFO [finetune.py:976] (4/7) Epoch 24, batch 1300, loss[loss=0.1389, simple_loss=0.2089, pruned_loss=0.03441, over 4787.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2408, pruned_loss=0.04933, over 954098.47 frames. ], batch size: 26, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:53:42,214 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.345e+01 1.496e+02 1.852e+02 2.141e+02 4.041e+02, threshold=3.705e+02, percent-clipped=2.0 +2023-03-27 04:53:57,209 INFO [finetune.py:976] (4/7) Epoch 24, batch 1350, loss[loss=0.1888, simple_loss=0.2462, pruned_loss=0.06575, over 4894.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2399, pruned_loss=0.04901, over 955659.29 frames. ], batch size: 35, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:54:07,945 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133102.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:54:18,371 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.19 vs. limit=5.0 +2023-03-27 04:54:31,060 INFO [finetune.py:976] (4/7) Epoch 24, batch 1400, loss[loss=0.1725, simple_loss=0.2425, pruned_loss=0.05127, over 4909.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2447, pruned_loss=0.05073, over 956148.01 frames. ], batch size: 36, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:54:35,885 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4549, 1.3590, 1.3442, 1.3634, 0.9767, 2.9489, 1.0635, 1.4093], + device='cuda:4'), covar=tensor([0.3778, 0.2944, 0.2467, 0.2836, 0.2143, 0.0331, 0.2872, 0.1484], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0115, 0.0121, 0.0123, 0.0113, 0.0096, 0.0094, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 04:54:59,479 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.355e+01 1.624e+02 1.914e+02 2.315e+02 3.947e+02, threshold=3.828e+02, percent-clipped=1.0 +2023-03-27 04:55:19,623 INFO [finetune.py:976] (4/7) Epoch 24, batch 1450, loss[loss=0.1356, simple_loss=0.2074, pruned_loss=0.03193, over 4775.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2463, pruned_loss=0.05108, over 957488.21 frames. ], batch size: 26, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:55:35,367 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9900, 1.7104, 2.2889, 1.5961, 2.0844, 2.3059, 1.6109, 2.3703], + device='cuda:4'), covar=tensor([0.1061, 0.1888, 0.1157, 0.1658, 0.0766, 0.1155, 0.2667, 0.0733], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0208, 0.0192, 0.0190, 0.0173, 0.0214, 0.0216, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:55:56,690 INFO [finetune.py:976] (4/7) Epoch 24, batch 1500, loss[loss=0.1842, simple_loss=0.2576, pruned_loss=0.0554, over 4820.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2462, pruned_loss=0.05058, over 955792.08 frames. ], batch size: 30, lr: 3.06e-03, grad_scale: 16.0 +2023-03-27 04:55:56,775 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.4408, 3.8841, 4.0490, 4.3199, 4.1952, 3.9062, 4.5315, 1.4574], + device='cuda:4'), covar=tensor([0.0695, 0.0806, 0.0884, 0.0776, 0.1109, 0.1578, 0.0685, 0.5692], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0247, 0.0281, 0.0293, 0.0338, 0.0286, 0.0306, 0.0300], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:55:56,959 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.74 vs. limit=5.0 +2023-03-27 04:56:15,020 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.176e+02 1.579e+02 1.864e+02 2.355e+02 5.095e+02, threshold=3.727e+02, percent-clipped=2.0 +2023-03-27 04:56:24,028 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1781, 2.0097, 1.6299, 2.0668, 2.0922, 1.7938, 2.3745, 2.1614], + device='cuda:4'), covar=tensor([0.1369, 0.2213, 0.3214, 0.2804, 0.2727, 0.1835, 0.3328, 0.1725], + device='cuda:4'), in_proj_covar=tensor([0.0187, 0.0188, 0.0235, 0.0251, 0.0247, 0.0204, 0.0213, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:56:30,467 INFO [finetune.py:976] (4/7) Epoch 24, batch 1550, loss[loss=0.1743, simple_loss=0.2455, pruned_loss=0.05156, over 4824.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2464, pruned_loss=0.05022, over 954707.57 frames. ], batch size: 38, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 04:56:50,655 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=133316.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:57:04,269 INFO [finetune.py:976] (4/7) Epoch 24, batch 1600, loss[loss=0.1785, simple_loss=0.2548, pruned_loss=0.0511, over 4903.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2454, pruned_loss=0.05053, over 955753.35 frames. ], batch size: 43, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 04:57:28,473 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.451e+02 1.796e+02 2.043e+02 3.402e+02, threshold=3.593e+02, percent-clipped=0.0 +2023-03-27 04:57:28,563 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=133364.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:57:36,396 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0688, 0.9751, 1.0122, 0.5232, 0.8746, 1.1207, 1.1817, 0.9736], + device='cuda:4'), covar=tensor([0.0740, 0.0535, 0.0515, 0.0444, 0.0515, 0.0530, 0.0364, 0.0587], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0148, 0.0127, 0.0122, 0.0131, 0.0130, 0.0141, 0.0148], + device='cuda:4'), out_proj_covar=tensor([8.8879e-05, 1.0682e-04, 9.0526e-05, 8.5786e-05, 9.2008e-05, 9.2299e-05, + 1.0044e-04, 1.0582e-04], device='cuda:4') +2023-03-27 04:57:38,827 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=133374.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 04:57:39,498 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.62 vs. limit=5.0 +2023-03-27 04:57:40,052 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=133376.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:57:46,650 INFO [finetune.py:976] (4/7) Epoch 24, batch 1650, loss[loss=0.139, simple_loss=0.2133, pruned_loss=0.03237, over 4874.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2429, pruned_loss=0.04949, over 955657.49 frames. ], batch size: 31, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 04:57:52,071 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6372, 2.8286, 2.7217, 1.9548, 2.6450, 2.9848, 3.0413, 2.5501], + device='cuda:4'), covar=tensor([0.0619, 0.0528, 0.0678, 0.0872, 0.0740, 0.0694, 0.0544, 0.1004], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0136, 0.0139, 0.0119, 0.0127, 0.0138, 0.0138, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:57:56,871 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=133402.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:58:19,419 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=133435.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 04:58:20,500 INFO [finetune.py:976] (4/7) Epoch 24, batch 1700, loss[loss=0.2004, simple_loss=0.2754, pruned_loss=0.06267, over 4908.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2407, pruned_loss=0.04863, over 956654.29 frames. ], batch size: 43, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 04:58:20,619 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=133437.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:58:21,797 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5909, 1.5795, 1.9517, 3.4710, 2.3040, 2.3774, 1.3091, 2.8420], + device='cuda:4'), covar=tensor([0.1869, 0.1594, 0.1543, 0.0642, 0.0859, 0.1477, 0.1826, 0.0543], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0162, 0.0100, 0.0136, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 04:58:31,234 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=133450.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:58:48,616 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.499e+01 1.457e+02 1.770e+02 2.219e+02 3.253e+02, threshold=3.541e+02, percent-clipped=0.0 +2023-03-27 04:58:54,594 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.3060, 4.6650, 4.8186, 5.1296, 5.0210, 4.7824, 5.4226, 1.4998], + device='cuda:4'), covar=tensor([0.0768, 0.0999, 0.0829, 0.0824, 0.1322, 0.1496, 0.0631, 0.6408], + device='cuda:4'), in_proj_covar=tensor([0.0348, 0.0248, 0.0282, 0.0293, 0.0338, 0.0287, 0.0307, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 04:58:55,845 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=133474.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:59:04,125 INFO [finetune.py:976] (4/7) Epoch 24, batch 1750, loss[loss=0.1782, simple_loss=0.2503, pruned_loss=0.05303, over 4119.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2428, pruned_loss=0.0495, over 954970.31 frames. ], batch size: 65, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 04:59:24,973 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2709, 1.7737, 0.8392, 2.0592, 2.5356, 1.7894, 2.0714, 2.0275], + device='cuda:4'), covar=tensor([0.1331, 0.1802, 0.1923, 0.1065, 0.1810, 0.1828, 0.1242, 0.1861], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0093, 0.0108, 0.0091, 0.0118, 0.0093, 0.0097, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 04:59:36,819 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=133535.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 04:59:37,913 INFO [finetune.py:976] (4/7) Epoch 24, batch 1800, loss[loss=0.1365, simple_loss=0.2074, pruned_loss=0.03281, over 4703.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2458, pruned_loss=0.0507, over 953558.92 frames. ], batch size: 23, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 04:59:57,741 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.575e+02 1.839e+02 2.282e+02 3.463e+02, threshold=3.677e+02, percent-clipped=0.0 +2023-03-27 05:00:09,629 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9843, 1.8378, 1.6339, 1.7112, 1.7607, 1.7459, 1.8246, 2.4627], + device='cuda:4'), covar=tensor([0.3229, 0.3364, 0.2755, 0.3018, 0.3517, 0.2172, 0.2950, 0.1328], + device='cuda:4'), in_proj_covar=tensor([0.0291, 0.0263, 0.0235, 0.0277, 0.0258, 0.0228, 0.0255, 0.0236], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:00:23,456 INFO [finetune.py:976] (4/7) Epoch 24, batch 1850, loss[loss=0.1988, simple_loss=0.2668, pruned_loss=0.06541, over 4730.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2484, pruned_loss=0.05236, over 952862.09 frames. ], batch size: 54, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 05:00:52,117 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-27 05:00:56,574 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.3018, 2.9445, 3.1187, 3.2304, 3.0936, 2.9193, 3.3595, 0.9168], + device='cuda:4'), covar=tensor([0.0975, 0.0949, 0.1015, 0.1027, 0.1530, 0.1715, 0.1120, 0.5289], + device='cuda:4'), in_proj_covar=tensor([0.0346, 0.0247, 0.0280, 0.0292, 0.0335, 0.0286, 0.0305, 0.0300], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:00:59,962 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6694, 3.7086, 3.4540, 1.7151, 3.8133, 2.9122, 0.7371, 2.5847], + device='cuda:4'), covar=tensor([0.2214, 0.2021, 0.1683, 0.3555, 0.1022, 0.1046, 0.4785, 0.1473], + device='cuda:4'), in_proj_covar=tensor([0.0154, 0.0180, 0.0163, 0.0130, 0.0162, 0.0125, 0.0150, 0.0125], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 05:01:04,055 INFO [finetune.py:976] (4/7) Epoch 24, batch 1900, loss[loss=0.2186, simple_loss=0.2867, pruned_loss=0.07519, over 4800.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2506, pruned_loss=0.05333, over 952344.32 frames. ], batch size: 41, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 05:01:14,206 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=133652.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:01:21,803 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.050e+02 1.540e+02 1.881e+02 2.277e+02 3.366e+02, threshold=3.762e+02, percent-clipped=0.0 +2023-03-27 05:01:28,955 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4469, 1.7794, 1.4434, 1.5899, 1.9732, 1.8917, 1.7580, 1.6887], + device='cuda:4'), covar=tensor([0.0595, 0.0335, 0.0578, 0.0319, 0.0308, 0.0580, 0.0374, 0.0411], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0107, 0.0145, 0.0112, 0.0100, 0.0112, 0.0102, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.7950e-05, 8.1812e-05, 1.1346e-04, 8.5645e-05, 7.8006e-05, 8.3128e-05, + 7.6071e-05, 8.5900e-05], device='cuda:4') +2023-03-27 05:01:35,358 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9953, 2.1957, 1.7819, 1.9585, 2.6403, 2.7096, 2.1145, 2.0152], + device='cuda:4'), covar=tensor([0.0423, 0.0330, 0.0545, 0.0331, 0.0225, 0.0387, 0.0378, 0.0401], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0107, 0.0145, 0.0112, 0.0100, 0.0112, 0.0102, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.7816e-05, 8.1684e-05, 1.1325e-04, 8.5518e-05, 7.7891e-05, 8.2988e-05, + 7.5920e-05, 8.5748e-05], device='cuda:4') +2023-03-27 05:01:37,659 INFO [finetune.py:976] (4/7) Epoch 24, batch 1950, loss[loss=0.2149, simple_loss=0.276, pruned_loss=0.07696, over 4818.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2476, pruned_loss=0.05171, over 952700.93 frames. ], batch size: 39, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 05:01:55,038 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=133713.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:02:06,303 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133730.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 05:02:07,531 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133732.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:02:11,411 INFO [finetune.py:976] (4/7) Epoch 24, batch 2000, loss[loss=0.1557, simple_loss=0.233, pruned_loss=0.03916, over 4866.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2443, pruned_loss=0.0509, over 953456.25 frames. ], batch size: 34, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 05:02:28,713 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.381e+01 1.373e+02 1.735e+02 2.243e+02 3.912e+02, threshold=3.469e+02, percent-clipped=2.0 +2023-03-27 05:02:54,161 INFO [finetune.py:976] (4/7) Epoch 24, batch 2050, loss[loss=0.1965, simple_loss=0.2534, pruned_loss=0.0698, over 4823.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2407, pruned_loss=0.05, over 954694.70 frames. ], batch size: 40, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 05:03:23,703 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=133830.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:03:27,930 INFO [finetune.py:976] (4/7) Epoch 24, batch 2100, loss[loss=0.1758, simple_loss=0.2402, pruned_loss=0.05568, over 4379.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2406, pruned_loss=0.04991, over 953881.86 frames. ], batch size: 19, lr: 3.06e-03, grad_scale: 32.0 +2023-03-27 05:03:34,474 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8023, 1.7412, 1.6742, 1.7278, 1.2307, 3.5692, 1.4906, 2.0806], + device='cuda:4'), covar=tensor([0.3185, 0.2288, 0.1965, 0.2345, 0.1673, 0.0200, 0.2323, 0.1060], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0123, 0.0113, 0.0097, 0.0095, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 05:03:47,594 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.827e+01 1.541e+02 1.860e+02 2.293e+02 6.118e+02, threshold=3.720e+02, percent-clipped=3.0 +2023-03-27 05:04:07,686 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4418, 1.3166, 1.4328, 0.7688, 1.4787, 1.4823, 1.4552, 1.2853], + device='cuda:4'), covar=tensor([0.0616, 0.0858, 0.0760, 0.1017, 0.0958, 0.0775, 0.0668, 0.1343], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0138, 0.0141, 0.0121, 0.0127, 0.0139, 0.0139, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:04:10,126 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4896, 1.0578, 0.7226, 1.3349, 1.9591, 1.0468, 1.1786, 1.3543], + device='cuda:4'), covar=tensor([0.2044, 0.2922, 0.2527, 0.1727, 0.2383, 0.2688, 0.2250, 0.2937], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0093, 0.0110, 0.0092, 0.0119, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 05:04:11,227 INFO [finetune.py:976] (4/7) Epoch 24, batch 2150, loss[loss=0.1897, simple_loss=0.2684, pruned_loss=0.05547, over 4738.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2449, pruned_loss=0.05161, over 954719.83 frames. ], batch size: 59, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:04:35,076 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.3176, 2.9503, 3.0826, 3.2300, 3.0887, 2.9234, 3.3623, 0.9524], + device='cuda:4'), covar=tensor([0.1009, 0.0992, 0.1068, 0.1029, 0.1572, 0.1884, 0.1119, 0.5501], + device='cuda:4'), in_proj_covar=tensor([0.0346, 0.0247, 0.0281, 0.0293, 0.0337, 0.0286, 0.0306, 0.0301], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:04:35,130 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4752, 1.3773, 1.2155, 1.5259, 1.6143, 1.5357, 0.9986, 1.2349], + device='cuda:4'), covar=tensor([0.2407, 0.2255, 0.2186, 0.1793, 0.1687, 0.1408, 0.2651, 0.2104], + device='cuda:4'), in_proj_covar=tensor([0.0244, 0.0209, 0.0213, 0.0196, 0.0243, 0.0190, 0.0216, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:04:36,217 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5346, 1.3859, 1.2344, 1.5562, 1.6326, 1.5708, 1.0029, 1.2529], + device='cuda:4'), covar=tensor([0.2118, 0.2069, 0.1906, 0.1690, 0.1539, 0.1269, 0.2563, 0.1929], + device='cuda:4'), in_proj_covar=tensor([0.0244, 0.0209, 0.0213, 0.0196, 0.0243, 0.0190, 0.0216, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:04:38,024 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2988, 1.3677, 1.5542, 1.0426, 1.3016, 1.5068, 1.3220, 1.6781], + device='cuda:4'), covar=tensor([0.1098, 0.2054, 0.1268, 0.1464, 0.0914, 0.1153, 0.3041, 0.0761], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0209, 0.0192, 0.0191, 0.0174, 0.0215, 0.0218, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:04:44,944 INFO [finetune.py:976] (4/7) Epoch 24, batch 2200, loss[loss=0.206, simple_loss=0.2728, pruned_loss=0.06961, over 4835.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2471, pruned_loss=0.05199, over 954604.41 frames. ], batch size: 47, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:05:02,720 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.027e+02 1.464e+02 1.824e+02 2.239e+02 3.694e+02, threshold=3.648e+02, percent-clipped=0.0 +2023-03-27 05:05:24,729 INFO [finetune.py:976] (4/7) Epoch 24, batch 2250, loss[loss=0.1621, simple_loss=0.2388, pruned_loss=0.04272, over 4829.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2482, pruned_loss=0.05206, over 954694.18 frames. ], batch size: 33, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:05:35,456 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=133997.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:05:49,827 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=134008.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:05:50,970 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4333, 1.3695, 1.8233, 1.6690, 1.5921, 3.1711, 1.4141, 1.5045], + device='cuda:4'), covar=tensor([0.0980, 0.1861, 0.1059, 0.0994, 0.1627, 0.0263, 0.1496, 0.1898], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0073, 0.0077, 0.0092, 0.0080, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 05:06:08,019 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134030.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 05:06:09,734 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134032.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:06:12,682 INFO [finetune.py:976] (4/7) Epoch 24, batch 2300, loss[loss=0.1734, simple_loss=0.239, pruned_loss=0.05392, over 4717.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2469, pruned_loss=0.05107, over 953471.72 frames. ], batch size: 59, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:06:27,066 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=134058.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:06:31,055 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.533e+02 1.723e+02 2.089e+02 4.293e+02, threshold=3.445e+02, percent-clipped=2.0 +2023-03-27 05:06:40,132 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134078.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 05:06:41,359 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134080.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:06:42,998 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.81 vs. limit=5.0 +2023-03-27 05:06:46,520 INFO [finetune.py:976] (4/7) Epoch 24, batch 2350, loss[loss=0.1925, simple_loss=0.2652, pruned_loss=0.05993, over 4829.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.244, pruned_loss=0.05007, over 953387.31 frames. ], batch size: 38, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:07:14,950 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134130.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:07:19,178 INFO [finetune.py:976] (4/7) Epoch 24, batch 2400, loss[loss=0.1475, simple_loss=0.2229, pruned_loss=0.03608, over 4843.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2429, pruned_loss=0.04999, over 953795.40 frames. ], batch size: 30, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:07:38,331 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.553e+01 1.434e+02 1.789e+02 2.166e+02 3.942e+02, threshold=3.577e+02, percent-clipped=1.0 +2023-03-27 05:07:49,929 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134178.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:07:55,475 INFO [finetune.py:976] (4/7) Epoch 24, batch 2450, loss[loss=0.1816, simple_loss=0.2489, pruned_loss=0.05713, over 4895.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2395, pruned_loss=0.04885, over 955109.02 frames. ], batch size: 35, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:08:36,907 INFO [finetune.py:976] (4/7) Epoch 24, batch 2500, loss[loss=0.191, simple_loss=0.2524, pruned_loss=0.06475, over 4380.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2432, pruned_loss=0.05075, over 952773.39 frames. ], batch size: 19, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:08:55,289 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-27 05:08:55,715 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.396e+01 1.499e+02 1.865e+02 2.171e+02 5.575e+02, threshold=3.730e+02, percent-clipped=1.0 +2023-03-27 05:09:20,384 INFO [finetune.py:976] (4/7) Epoch 24, batch 2550, loss[loss=0.166, simple_loss=0.2396, pruned_loss=0.04615, over 4806.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.247, pruned_loss=0.05141, over 954096.57 frames. ], batch size: 39, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:09:32,222 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=134304.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:09:34,657 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134308.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:09:35,286 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=134309.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:09:47,615 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9522, 1.3747, 2.3796, 3.7302, 2.3921, 2.7970, 0.9736, 3.2165], + device='cuda:4'), covar=tensor([0.2013, 0.2119, 0.1599, 0.0819, 0.0973, 0.1656, 0.2294, 0.0535], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0162, 0.0100, 0.0136, 0.0124, 0.0099], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 05:09:54,096 INFO [finetune.py:976] (4/7) Epoch 24, batch 2600, loss[loss=0.1522, simple_loss=0.2086, pruned_loss=0.04791, over 4228.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2475, pruned_loss=0.05155, over 953374.94 frames. ], batch size: 18, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:10:04,830 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=134353.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:10:07,200 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134356.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:10:09,170 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.40 vs. limit=5.0 +2023-03-27 05:10:12,059 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.471e+02 1.806e+02 2.184e+02 4.519e+02, threshold=3.612e+02, percent-clipped=2.0 +2023-03-27 05:10:13,293 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=134365.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:10:16,828 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=134370.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:10:21,966 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-27 05:10:29,856 INFO [finetune.py:976] (4/7) Epoch 24, batch 2650, loss[loss=0.1655, simple_loss=0.2384, pruned_loss=0.04629, over 4792.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2493, pruned_loss=0.05181, over 954514.36 frames. ], batch size: 51, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:10:54,502 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.12 vs. limit=5.0 +2023-03-27 05:11:21,026 INFO [finetune.py:976] (4/7) Epoch 24, batch 2700, loss[loss=0.1725, simple_loss=0.2401, pruned_loss=0.05247, over 4895.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2467, pruned_loss=0.05051, over 954610.39 frames. ], batch size: 32, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:11:39,167 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.916e+01 1.416e+02 1.758e+02 2.159e+02 3.599e+02, threshold=3.516e+02, percent-clipped=0.0 +2023-03-27 05:11:54,602 INFO [finetune.py:976] (4/7) Epoch 24, batch 2750, loss[loss=0.1919, simple_loss=0.2463, pruned_loss=0.06879, over 4813.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2434, pruned_loss=0.04976, over 954407.90 frames. ], batch size: 25, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:12:27,874 INFO [finetune.py:976] (4/7) Epoch 24, batch 2800, loss[loss=0.1888, simple_loss=0.2417, pruned_loss=0.06801, over 4934.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2394, pruned_loss=0.04836, over 953867.15 frames. ], batch size: 33, lr: 3.05e-03, grad_scale: 32.0 +2023-03-27 05:12:42,138 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9873, 2.7262, 2.6173, 1.3736, 2.7022, 2.2248, 2.1105, 2.5202], + device='cuda:4'), covar=tensor([0.1161, 0.0850, 0.1653, 0.2075, 0.1746, 0.2219, 0.2113, 0.1313], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0190, 0.0199, 0.0180, 0.0210, 0.0208, 0.0223, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:12:46,116 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.301e+01 1.456e+02 1.823e+02 2.196e+02 4.309e+02, threshold=3.645e+02, percent-clipped=3.0 +2023-03-27 05:12:46,208 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.3226, 2.9387, 3.0952, 3.2589, 3.1197, 2.8966, 3.3569, 1.0439], + device='cuda:4'), covar=tensor([0.1243, 0.1127, 0.1201, 0.1292, 0.1746, 0.1958, 0.1187, 0.5875], + device='cuda:4'), in_proj_covar=tensor([0.0347, 0.0246, 0.0281, 0.0293, 0.0336, 0.0286, 0.0307, 0.0301], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:13:01,599 INFO [finetune.py:976] (4/7) Epoch 24, batch 2850, loss[loss=0.1728, simple_loss=0.242, pruned_loss=0.05187, over 4868.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2379, pruned_loss=0.04817, over 952881.14 frames. ], batch size: 31, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:13:10,389 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.7422, 4.0612, 4.3082, 4.5205, 4.4935, 4.1712, 4.8434, 1.6499], + device='cuda:4'), covar=tensor([0.0792, 0.0851, 0.0865, 0.1000, 0.1288, 0.1687, 0.0628, 0.5709], + device='cuda:4'), in_proj_covar=tensor([0.0348, 0.0247, 0.0282, 0.0293, 0.0337, 0.0287, 0.0307, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:13:29,389 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6249, 1.6118, 1.2926, 1.5359, 1.8790, 1.8685, 1.5438, 1.4184], + device='cuda:4'), covar=tensor([0.0320, 0.0286, 0.0634, 0.0297, 0.0225, 0.0443, 0.0325, 0.0418], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0107, 0.0145, 0.0112, 0.0101, 0.0114, 0.0103, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.7921e-05, 8.1857e-05, 1.1371e-04, 8.6043e-05, 7.8426e-05, 8.4255e-05, + 7.6537e-05, 8.6072e-05], device='cuda:4') +2023-03-27 05:13:45,376 INFO [finetune.py:976] (4/7) Epoch 24, batch 2900, loss[loss=0.2237, simple_loss=0.3084, pruned_loss=0.06947, over 4863.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2419, pruned_loss=0.04996, over 953049.29 frames. ], batch size: 44, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:13:49,790 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5929, 1.5540, 1.3842, 1.7289, 2.0023, 1.7193, 1.2844, 1.3483], + device='cuda:4'), covar=tensor([0.2233, 0.2049, 0.2026, 0.1658, 0.1713, 0.1307, 0.2524, 0.1939], + device='cuda:4'), in_proj_covar=tensor([0.0244, 0.0210, 0.0213, 0.0196, 0.0243, 0.0190, 0.0215, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:13:50,343 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7437, 4.1819, 3.9296, 2.1957, 4.2539, 3.1833, 0.7169, 3.0149], + device='cuda:4'), covar=tensor([0.2965, 0.1685, 0.1462, 0.3341, 0.0879, 0.1019, 0.4794, 0.1452], + device='cuda:4'), in_proj_covar=tensor([0.0154, 0.0180, 0.0162, 0.0130, 0.0161, 0.0125, 0.0149, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 05:13:56,311 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134653.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:14:00,531 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=134660.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:14:03,946 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.559e+02 1.783e+02 2.063e+02 3.902e+02, threshold=3.566e+02, percent-clipped=1.0 +2023-03-27 05:14:04,031 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=134665.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:14:20,895 INFO [finetune.py:976] (4/7) Epoch 24, batch 2950, loss[loss=0.1787, simple_loss=0.2508, pruned_loss=0.05333, over 4882.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2436, pruned_loss=0.04986, over 953348.14 frames. ], batch size: 35, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:14:36,266 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3041, 1.4156, 1.6526, 1.5632, 1.5683, 2.9921, 1.4465, 1.5209], + device='cuda:4'), covar=tensor([0.1008, 0.1740, 0.1039, 0.0949, 0.1570, 0.0262, 0.1394, 0.1776], + device='cuda:4'), in_proj_covar=tensor([0.0074, 0.0081, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 05:14:37,443 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=134701.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:15:02,625 INFO [finetune.py:976] (4/7) Epoch 24, batch 3000, loss[loss=0.1846, simple_loss=0.2598, pruned_loss=0.05469, over 4858.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2457, pruned_loss=0.05068, over 955645.82 frames. ], batch size: 49, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:15:02,625 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 05:15:13,334 INFO [finetune.py:1010] (4/7) Epoch 24, validation: loss=0.1561, simple_loss=0.2251, pruned_loss=0.0436, over 2265189.00 frames. +2023-03-27 05:15:13,334 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6557MB +2023-03-27 05:15:31,260 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.188e+02 1.527e+02 1.858e+02 2.241e+02 5.364e+02, threshold=3.716e+02, percent-clipped=3.0 +2023-03-27 05:15:39,223 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.2292, 1.2311, 1.3186, 0.6409, 1.2095, 1.5379, 1.5243, 1.3003], + device='cuda:4'), covar=tensor([0.0798, 0.0579, 0.0489, 0.0447, 0.0481, 0.0474, 0.0297, 0.0608], + device='cuda:4'), in_proj_covar=tensor([0.0121, 0.0147, 0.0126, 0.0121, 0.0130, 0.0129, 0.0140, 0.0147], + device='cuda:4'), out_proj_covar=tensor([8.8225e-05, 1.0598e-04, 8.9793e-05, 8.5033e-05, 9.1365e-05, 9.1677e-05, + 9.9864e-05, 1.0524e-04], device='cuda:4') +2023-03-27 05:15:48,110 INFO [finetune.py:976] (4/7) Epoch 24, batch 3050, loss[loss=0.1283, simple_loss=0.1984, pruned_loss=0.02905, over 4703.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2474, pruned_loss=0.05122, over 955463.44 frames. ], batch size: 23, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:16:05,363 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 05:16:39,366 INFO [finetune.py:976] (4/7) Epoch 24, batch 3100, loss[loss=0.1832, simple_loss=0.2533, pruned_loss=0.05658, over 4919.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2454, pruned_loss=0.05055, over 954403.51 frames. ], batch size: 33, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:16:56,493 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-27 05:16:56,672 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=134862.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:16:58,377 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.835e+01 1.385e+02 1.668e+02 2.115e+02 5.080e+02, threshold=3.336e+02, percent-clipped=1.0 +2023-03-27 05:17:06,143 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4507, 1.0498, 0.6695, 1.2745, 1.9400, 0.7702, 1.1739, 1.2827], + device='cuda:4'), covar=tensor([0.1904, 0.2646, 0.2142, 0.1577, 0.2149, 0.2381, 0.1884, 0.2638], + device='cuda:4'), in_proj_covar=tensor([0.0088, 0.0091, 0.0108, 0.0090, 0.0117, 0.0091, 0.0096, 0.0087], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 05:17:12,755 INFO [finetune.py:976] (4/7) Epoch 24, batch 3150, loss[loss=0.1342, simple_loss=0.2079, pruned_loss=0.0303, over 4758.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2431, pruned_loss=0.05003, over 953809.67 frames. ], batch size: 26, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:17:27,481 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1232, 1.8946, 2.5080, 1.6104, 2.2987, 2.4862, 1.7970, 2.5700], + device='cuda:4'), covar=tensor([0.1344, 0.2108, 0.1549, 0.2146, 0.1019, 0.1177, 0.2833, 0.0861], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0207, 0.0192, 0.0190, 0.0173, 0.0214, 0.0216, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:17:37,234 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=134923.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:17:39,599 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9115, 1.3482, 1.9673, 1.9298, 1.7531, 1.6702, 1.8531, 1.8698], + device='cuda:4'), covar=tensor([0.3793, 0.3928, 0.3212, 0.3594, 0.4712, 0.3751, 0.4379, 0.2768], + device='cuda:4'), in_proj_covar=tensor([0.0260, 0.0245, 0.0266, 0.0291, 0.0290, 0.0266, 0.0297, 0.0249], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:17:46,591 INFO [finetune.py:976] (4/7) Epoch 24, batch 3200, loss[loss=0.1693, simple_loss=0.2408, pruned_loss=0.04891, over 4776.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2397, pruned_loss=0.04898, over 954239.70 frames. ], batch size: 28, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:18:03,011 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134960.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:18:05,933 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.665e+01 1.513e+02 1.793e+02 2.252e+02 3.579e+02, threshold=3.586e+02, percent-clipped=1.0 +2023-03-27 05:18:06,034 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=134965.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:18:22,483 INFO [finetune.py:976] (4/7) Epoch 24, batch 3250, loss[loss=0.1829, simple_loss=0.2409, pruned_loss=0.06241, over 4657.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2402, pruned_loss=0.04899, over 954421.69 frames. ], batch size: 23, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:18:45,538 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=135008.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:18:48,605 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=135013.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:18:59,447 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8662, 1.8762, 1.7248, 1.9075, 1.5524, 4.6938, 1.7991, 2.1858], + device='cuda:4'), covar=tensor([0.3169, 0.2497, 0.2105, 0.2244, 0.1637, 0.0113, 0.2408, 0.1229], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0123, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 05:19:04,070 INFO [finetune.py:976] (4/7) Epoch 24, batch 3300, loss[loss=0.2118, simple_loss=0.2817, pruned_loss=0.07095, over 4847.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2437, pruned_loss=0.05017, over 954925.09 frames. ], batch size: 47, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:19:13,239 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.29 vs. limit=5.0 +2023-03-27 05:19:23,523 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.563e+02 1.899e+02 2.275e+02 5.700e+02, threshold=3.799e+02, percent-clipped=2.0 +2023-03-27 05:19:44,198 INFO [finetune.py:976] (4/7) Epoch 24, batch 3350, loss[loss=0.1686, simple_loss=0.2502, pruned_loss=0.0435, over 4809.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2467, pruned_loss=0.05099, over 954518.76 frames. ], batch size: 45, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:20:21,440 INFO [finetune.py:976] (4/7) Epoch 24, batch 3400, loss[loss=0.1824, simple_loss=0.2514, pruned_loss=0.05663, over 4915.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2481, pruned_loss=0.05163, over 954662.53 frames. ], batch size: 38, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:20:40,369 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.077e+02 1.584e+02 1.828e+02 2.150e+02 3.792e+02, threshold=3.656e+02, percent-clipped=0.0 +2023-03-27 05:20:44,624 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=135171.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:20:53,780 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8857, 2.3728, 3.1479, 2.0131, 2.7829, 3.2593, 2.3821, 3.1179], + device='cuda:4'), covar=tensor([0.1283, 0.1957, 0.1401, 0.2135, 0.0966, 0.1256, 0.2364, 0.0866], + device='cuda:4'), in_proj_covar=tensor([0.0191, 0.0206, 0.0190, 0.0189, 0.0172, 0.0212, 0.0214, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:20:54,278 INFO [finetune.py:976] (4/7) Epoch 24, batch 3450, loss[loss=0.1395, simple_loss=0.2066, pruned_loss=0.03619, over 4726.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2469, pruned_loss=0.05087, over 954224.44 frames. ], batch size: 23, lr: 3.05e-03, grad_scale: 16.0 +2023-03-27 05:21:03,892 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.44 vs. limit=5.0 +2023-03-27 05:21:27,773 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=135218.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:21:39,449 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1127, 2.0077, 2.1437, 1.3860, 2.0446, 2.2018, 2.2255, 1.7345], + device='cuda:4'), covar=tensor([0.0487, 0.0576, 0.0631, 0.0871, 0.1260, 0.0560, 0.0492, 0.1119], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0136, 0.0139, 0.0119, 0.0126, 0.0137, 0.0137, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:21:40,720 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=135232.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:21:47,102 INFO [finetune.py:976] (4/7) Epoch 24, batch 3500, loss[loss=0.2293, simple_loss=0.2881, pruned_loss=0.08522, over 4248.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2441, pruned_loss=0.05032, over 951191.81 frames. ], batch size: 66, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:21:54,355 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.55 vs. limit=2.0 +2023-03-27 05:22:06,079 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.678e+01 1.506e+02 1.714e+02 2.011e+02 3.544e+02, threshold=3.428e+02, percent-clipped=0.0 +2023-03-27 05:22:16,397 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.65 vs. limit=2.0 +2023-03-27 05:22:20,455 INFO [finetune.py:976] (4/7) Epoch 24, batch 3550, loss[loss=0.1376, simple_loss=0.2105, pruned_loss=0.0324, over 4762.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2413, pruned_loss=0.04919, over 953061.10 frames. ], batch size: 26, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:22:54,388 INFO [finetune.py:976] (4/7) Epoch 24, batch 3600, loss[loss=0.1488, simple_loss=0.2172, pruned_loss=0.04022, over 4914.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2401, pruned_loss=0.04972, over 952650.71 frames. ], batch size: 37, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:23:03,990 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8373, 1.3041, 1.8995, 1.8420, 1.6632, 1.5931, 1.7833, 1.7864], + device='cuda:4'), covar=tensor([0.3516, 0.3573, 0.2909, 0.3372, 0.4190, 0.3604, 0.4093, 0.2736], + device='cuda:4'), in_proj_covar=tensor([0.0260, 0.0244, 0.0264, 0.0290, 0.0289, 0.0265, 0.0296, 0.0248], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:23:11,479 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-03-27 05:23:12,797 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.699e+01 1.474e+02 1.759e+02 2.084e+02 3.295e+02, threshold=3.517e+02, percent-clipped=0.0 +2023-03-27 05:23:28,233 INFO [finetune.py:976] (4/7) Epoch 24, batch 3650, loss[loss=0.1197, simple_loss=0.1956, pruned_loss=0.02187, over 4746.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2407, pruned_loss=0.0497, over 952817.44 frames. ], batch size: 23, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:24:09,543 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1058, 1.8382, 2.4136, 1.6619, 2.2423, 2.3990, 1.7255, 2.5110], + device='cuda:4'), covar=tensor([0.1372, 0.2062, 0.1455, 0.1877, 0.1023, 0.1297, 0.2686, 0.0746], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0209, 0.0194, 0.0192, 0.0175, 0.0216, 0.0218, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:24:11,249 INFO [finetune.py:976] (4/7) Epoch 24, batch 3700, loss[loss=0.1662, simple_loss=0.2359, pruned_loss=0.04825, over 4680.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2447, pruned_loss=0.05127, over 953447.11 frames. ], batch size: 23, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:24:28,522 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.118e+01 1.614e+02 1.999e+02 2.429e+02 5.138e+02, threshold=3.998e+02, percent-clipped=6.0 +2023-03-27 05:24:35,157 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3451, 1.3907, 1.3905, 0.7492, 1.4668, 1.6538, 1.6981, 1.3192], + device='cuda:4'), covar=tensor([0.1051, 0.0550, 0.0509, 0.0505, 0.0461, 0.0581, 0.0303, 0.0664], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0147, 0.0126, 0.0121, 0.0130, 0.0129, 0.0140, 0.0148], + device='cuda:4'), out_proj_covar=tensor([8.8528e-05, 1.0611e-04, 8.9813e-05, 8.4981e-05, 9.1552e-05, 9.1598e-05, + 1.0005e-04, 1.0564e-04], device='cuda:4') +2023-03-27 05:24:39,785 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.2396, 4.5674, 4.7725, 5.0882, 4.9411, 4.7759, 5.4079, 1.7195], + device='cuda:4'), covar=tensor([0.0745, 0.0710, 0.0755, 0.0870, 0.1342, 0.1663, 0.0478, 0.5642], + device='cuda:4'), in_proj_covar=tensor([0.0348, 0.0247, 0.0283, 0.0294, 0.0338, 0.0287, 0.0308, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:24:43,336 INFO [finetune.py:976] (4/7) Epoch 24, batch 3750, loss[loss=0.1627, simple_loss=0.2437, pruned_loss=0.04088, over 4736.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2474, pruned_loss=0.05245, over 952452.03 frames. ], batch size: 54, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:24:46,886 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-27 05:25:12,913 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=135518.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:25:18,844 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=135527.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:25:26,762 INFO [finetune.py:976] (4/7) Epoch 24, batch 3800, loss[loss=0.1479, simple_loss=0.2289, pruned_loss=0.03348, over 4756.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2475, pruned_loss=0.05181, over 954008.19 frames. ], batch size: 27, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:25:44,706 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.524e+02 1.815e+02 2.221e+02 4.659e+02, threshold=3.630e+02, percent-clipped=3.0 +2023-03-27 05:25:45,391 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=135566.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:25:52,388 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-27 05:26:00,455 INFO [finetune.py:976] (4/7) Epoch 24, batch 3850, loss[loss=0.1584, simple_loss=0.222, pruned_loss=0.04743, over 4883.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2449, pruned_loss=0.0505, over 953282.96 frames. ], batch size: 32, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:26:20,881 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6836, 1.5238, 1.0362, 0.2942, 1.3033, 1.4831, 1.4125, 1.4756], + device='cuda:4'), covar=tensor([0.0848, 0.0734, 0.1330, 0.1756, 0.1222, 0.2142, 0.2125, 0.0788], + device='cuda:4'), in_proj_covar=tensor([0.0168, 0.0189, 0.0197, 0.0178, 0.0207, 0.0206, 0.0220, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:26:38,633 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=135630.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:26:45,793 INFO [finetune.py:976] (4/7) Epoch 24, batch 3900, loss[loss=0.1404, simple_loss=0.2145, pruned_loss=0.03316, over 4698.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.243, pruned_loss=0.05002, over 953157.20 frames. ], batch size: 23, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:27:10,710 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.672e+01 1.400e+02 1.667e+02 1.961e+02 4.314e+02, threshold=3.334e+02, percent-clipped=1.0 +2023-03-27 05:27:26,037 INFO [finetune.py:976] (4/7) Epoch 24, batch 3950, loss[loss=0.1761, simple_loss=0.2443, pruned_loss=0.05397, over 4819.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2403, pruned_loss=0.04896, over 954094.79 frames. ], batch size: 41, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:27:29,095 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=135691.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:27:43,667 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9070, 1.6807, 2.0743, 1.4308, 1.9420, 2.1818, 1.5833, 2.2993], + device='cuda:4'), covar=tensor([0.1271, 0.2138, 0.1319, 0.1775, 0.0982, 0.1200, 0.2768, 0.0787], + device='cuda:4'), in_proj_covar=tensor([0.0191, 0.0207, 0.0191, 0.0190, 0.0173, 0.0213, 0.0216, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:27:58,424 INFO [finetune.py:976] (4/7) Epoch 24, batch 4000, loss[loss=0.1256, simple_loss=0.2042, pruned_loss=0.02349, over 4685.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2388, pruned_loss=0.04833, over 952244.49 frames. ], batch size: 23, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:28:16,421 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.536e+01 1.548e+02 1.897e+02 2.315e+02 3.943e+02, threshold=3.793e+02, percent-clipped=5.0 +2023-03-27 05:28:31,233 INFO [finetune.py:976] (4/7) Epoch 24, batch 4050, loss[loss=0.1856, simple_loss=0.2664, pruned_loss=0.05233, over 4870.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2428, pruned_loss=0.04975, over 952059.33 frames. ], batch size: 34, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:28:36,007 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6828, 1.4449, 1.0264, 0.3160, 1.2581, 1.4623, 1.3731, 1.4213], + device='cuda:4'), covar=tensor([0.0930, 0.0842, 0.1531, 0.2026, 0.1374, 0.2479, 0.2397, 0.0901], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0189, 0.0198, 0.0179, 0.0209, 0.0208, 0.0222, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:28:38,934 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.9341, 0.9025, 0.8325, 0.9942, 1.0411, 1.0252, 0.8916, 0.8689], + device='cuda:4'), covar=tensor([0.0426, 0.0256, 0.0595, 0.0274, 0.0255, 0.0401, 0.0335, 0.0336], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0107, 0.0146, 0.0112, 0.0101, 0.0114, 0.0103, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.7655e-05, 8.1767e-05, 1.1436e-04, 8.5986e-05, 7.8522e-05, 8.4847e-05, + 7.6266e-05, 8.6120e-05], device='cuda:4') +2023-03-27 05:28:49,470 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9467, 1.6670, 2.2489, 1.4846, 2.0453, 2.2405, 1.5958, 2.3782], + device='cuda:4'), covar=tensor([0.1365, 0.2297, 0.1444, 0.2056, 0.0886, 0.1384, 0.2757, 0.0751], + device='cuda:4'), in_proj_covar=tensor([0.0191, 0.0206, 0.0191, 0.0189, 0.0173, 0.0212, 0.0215, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:28:59,159 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=135827.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:29:05,387 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.4777, 1.4006, 1.4765, 0.7454, 1.5131, 1.7783, 1.7376, 1.3426], + device='cuda:4'), covar=tensor([0.0854, 0.0593, 0.0478, 0.0515, 0.0410, 0.0498, 0.0277, 0.0621], + device='cuda:4'), in_proj_covar=tensor([0.0121, 0.0146, 0.0124, 0.0119, 0.0129, 0.0127, 0.0138, 0.0146], + device='cuda:4'), out_proj_covar=tensor([8.7707e-05, 1.0508e-04, 8.8607e-05, 8.3924e-05, 9.0415e-05, 9.0493e-05, + 9.8766e-05, 1.0411e-04], device='cuda:4') +2023-03-27 05:29:09,954 INFO [finetune.py:976] (4/7) Epoch 24, batch 4100, loss[loss=0.1936, simple_loss=0.2486, pruned_loss=0.06933, over 4832.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2457, pruned_loss=0.05061, over 952239.93 frames. ], batch size: 30, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:29:26,948 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.95 vs. limit=5.0 +2023-03-27 05:29:28,035 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-27 05:29:30,371 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7946, 1.3283, 1.9152, 1.8552, 1.6848, 1.5952, 1.7973, 1.7771], + device='cuda:4'), covar=tensor([0.4029, 0.4007, 0.3321, 0.3663, 0.4632, 0.3933, 0.4682, 0.3170], + device='cuda:4'), in_proj_covar=tensor([0.0260, 0.0244, 0.0265, 0.0290, 0.0289, 0.0266, 0.0296, 0.0249], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:29:32,641 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.978e+01 1.562e+02 1.866e+02 2.353e+02 4.250e+02, threshold=3.731e+02, percent-clipped=2.0 +2023-03-27 05:29:39,221 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=135875.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:29:46,950 INFO [finetune.py:976] (4/7) Epoch 24, batch 4150, loss[loss=0.1681, simple_loss=0.2421, pruned_loss=0.04705, over 4808.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2469, pruned_loss=0.0511, over 952710.39 frames. ], batch size: 33, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:29:52,789 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 05:30:00,934 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1149, 2.0632, 2.2058, 0.9412, 2.5603, 2.7536, 2.3687, 1.9896], + device='cuda:4'), covar=tensor([0.0941, 0.0769, 0.0473, 0.0737, 0.0427, 0.0563, 0.0410, 0.0677], + device='cuda:4'), in_proj_covar=tensor([0.0121, 0.0147, 0.0125, 0.0120, 0.0130, 0.0128, 0.0139, 0.0147], + device='cuda:4'), out_proj_covar=tensor([8.8068e-05, 1.0564e-04, 8.9188e-05, 8.4334e-05, 9.1070e-05, 9.0898e-05, + 9.9351e-05, 1.0483e-04], device='cuda:4') +2023-03-27 05:30:19,806 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2971, 1.8648, 2.2797, 2.3095, 2.0152, 2.0284, 2.2640, 2.0951], + device='cuda:4'), covar=tensor([0.4273, 0.4284, 0.3428, 0.3904, 0.5173, 0.4112, 0.4765, 0.3273], + device='cuda:4'), in_proj_covar=tensor([0.0260, 0.0244, 0.0264, 0.0290, 0.0289, 0.0266, 0.0295, 0.0248], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:30:30,443 INFO [finetune.py:976] (4/7) Epoch 24, batch 4200, loss[loss=0.158, simple_loss=0.2438, pruned_loss=0.0361, over 4862.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2467, pruned_loss=0.05035, over 952073.02 frames. ], batch size: 34, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:30:49,318 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.053e+01 1.587e+02 1.796e+02 2.438e+02 3.967e+02, threshold=3.591e+02, percent-clipped=1.0 +2023-03-27 05:31:00,628 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=135982.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:31:03,039 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=135986.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:31:03,577 INFO [finetune.py:976] (4/7) Epoch 24, batch 4250, loss[loss=0.1518, simple_loss=0.2284, pruned_loss=0.03764, over 4826.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2442, pruned_loss=0.04987, over 955214.87 frames. ], batch size: 38, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:31:08,463 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=135994.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:31:45,377 INFO [finetune.py:976] (4/7) Epoch 24, batch 4300, loss[loss=0.1716, simple_loss=0.2407, pruned_loss=0.05122, over 4765.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.241, pruned_loss=0.04911, over 955463.98 frames. ], batch size: 27, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:31:49,197 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136043.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:32:02,583 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136055.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:32:10,847 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-27 05:32:14,137 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.579e+01 1.491e+02 1.827e+02 2.181e+02 5.621e+02, threshold=3.653e+02, percent-clipped=1.0 +2023-03-27 05:32:31,268 INFO [finetune.py:976] (4/7) Epoch 24, batch 4350, loss[loss=0.1686, simple_loss=0.2579, pruned_loss=0.03968, over 4812.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2389, pruned_loss=0.04832, over 956573.48 frames. ], batch size: 39, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:33:04,534 INFO [finetune.py:976] (4/7) Epoch 24, batch 4400, loss[loss=0.2173, simple_loss=0.2837, pruned_loss=0.07549, over 4808.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2402, pruned_loss=0.0491, over 954509.80 frames. ], batch size: 51, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:33:05,276 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5554, 1.1075, 0.7929, 1.3893, 2.0475, 0.7412, 1.2817, 1.3844], + device='cuda:4'), covar=tensor([0.1539, 0.2187, 0.1700, 0.1292, 0.1906, 0.1911, 0.1562, 0.1997], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0111, 0.0092, 0.0119, 0.0093, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 05:33:08,172 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=136142.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:33:17,141 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.48 vs. limit=5.0 +2023-03-27 05:33:23,889 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.065e+02 1.540e+02 1.819e+02 2.170e+02 3.954e+02, threshold=3.638e+02, percent-clipped=3.0 +2023-03-27 05:33:37,769 INFO [finetune.py:976] (4/7) Epoch 24, batch 4450, loss[loss=0.1826, simple_loss=0.2577, pruned_loss=0.05371, over 4849.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.244, pruned_loss=0.05031, over 954322.16 frames. ], batch size: 47, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:33:48,791 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136203.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:34:13,602 INFO [finetune.py:976] (4/7) Epoch 24, batch 4500, loss[loss=0.2324, simple_loss=0.2909, pruned_loss=0.08698, over 4792.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2453, pruned_loss=0.05083, over 952978.69 frames. ], batch size: 51, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:34:13,783 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-27 05:34:39,506 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.886e+01 1.509e+02 1.852e+02 2.239e+02 3.856e+02, threshold=3.704e+02, percent-clipped=1.0 +2023-03-27 05:34:54,252 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136286.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:34:54,760 INFO [finetune.py:976] (4/7) Epoch 24, batch 4550, loss[loss=0.2464, simple_loss=0.2964, pruned_loss=0.0982, over 4261.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2475, pruned_loss=0.05205, over 953203.30 frames. ], batch size: 65, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:35:28,229 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=136334.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:35:28,899 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=136335.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:35:30,006 INFO [finetune.py:976] (4/7) Epoch 24, batch 4600, loss[loss=0.219, simple_loss=0.2795, pruned_loss=0.07924, over 4820.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2479, pruned_loss=0.05197, over 952878.98 frames. ], batch size: 33, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:35:35,219 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136338.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:35:45,768 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136350.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:35:56,270 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.624e+02 1.856e+02 2.259e+02 4.732e+02, threshold=3.713e+02, percent-clipped=2.0 +2023-03-27 05:36:11,519 INFO [finetune.py:976] (4/7) Epoch 24, batch 4650, loss[loss=0.1695, simple_loss=0.2341, pruned_loss=0.05247, over 4808.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2453, pruned_loss=0.05132, over 951845.46 frames. ], batch size: 25, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:36:17,140 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136396.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:36:34,529 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3312, 2.2090, 1.8997, 1.1198, 2.0683, 1.9009, 1.7652, 2.0968], + device='cuda:4'), covar=tensor([0.0779, 0.0681, 0.1416, 0.1790, 0.1120, 0.1687, 0.1908, 0.0833], + device='cuda:4'), in_proj_covar=tensor([0.0172, 0.0193, 0.0202, 0.0183, 0.0213, 0.0211, 0.0226, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:36:45,427 INFO [finetune.py:976] (4/7) Epoch 24, batch 4700, loss[loss=0.122, simple_loss=0.1883, pruned_loss=0.02779, over 4841.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2416, pruned_loss=0.04977, over 952663.45 frames. ], batch size: 49, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:37:13,731 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.054e+02 1.479e+02 1.754e+02 2.064e+02 3.231e+02, threshold=3.507e+02, percent-clipped=0.0 +2023-03-27 05:37:38,206 INFO [finetune.py:976] (4/7) Epoch 24, batch 4750, loss[loss=0.1512, simple_loss=0.2319, pruned_loss=0.0352, over 4744.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2402, pruned_loss=0.04949, over 953288.39 frames. ], batch size: 27, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:37:39,010 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-27 05:37:44,924 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136498.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:38:02,856 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0023, 1.4599, 2.0648, 1.9798, 1.7652, 1.7386, 1.9251, 1.9062], + device='cuda:4'), covar=tensor([0.3994, 0.3990, 0.3225, 0.3641, 0.4980, 0.3737, 0.4286, 0.3108], + device='cuda:4'), in_proj_covar=tensor([0.0261, 0.0244, 0.0265, 0.0290, 0.0289, 0.0266, 0.0296, 0.0248], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:38:10,346 INFO [finetune.py:976] (4/7) Epoch 24, batch 4800, loss[loss=0.179, simple_loss=0.2566, pruned_loss=0.05069, over 4815.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2411, pruned_loss=0.04982, over 953077.68 frames. ], batch size: 45, lr: 3.04e-03, grad_scale: 16.0 +2023-03-27 05:38:17,391 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=2.01 vs. limit=2.0 +2023-03-27 05:38:28,978 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.582e+02 2.021e+02 2.347e+02 5.093e+02, threshold=4.042e+02, percent-clipped=3.0 +2023-03-27 05:38:44,073 INFO [finetune.py:976] (4/7) Epoch 24, batch 4850, loss[loss=0.1712, simple_loss=0.2442, pruned_loss=0.04914, over 4779.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2443, pruned_loss=0.05065, over 953555.24 frames. ], batch size: 25, lr: 3.04e-03, grad_scale: 32.0 +2023-03-27 05:38:50,650 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.9636, 0.9232, 0.9489, 1.1038, 1.1331, 1.0803, 0.9739, 0.9341], + device='cuda:4'), covar=tensor([0.0398, 0.0296, 0.0635, 0.0277, 0.0272, 0.0438, 0.0330, 0.0393], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0106, 0.0145, 0.0112, 0.0101, 0.0114, 0.0103, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.7608e-05, 8.1432e-05, 1.1362e-04, 8.5385e-05, 7.8224e-05, 8.4256e-05, + 7.6197e-05, 8.5803e-05], device='cuda:4') +2023-03-27 05:38:58,736 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7507, 1.6998, 1.6368, 1.6985, 1.4270, 4.1723, 1.5902, 1.9818], + device='cuda:4'), covar=tensor([0.3035, 0.2376, 0.1947, 0.2172, 0.1497, 0.0135, 0.2378, 0.1174], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0117, 0.0122, 0.0124, 0.0114, 0.0097, 0.0095, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 05:39:05,382 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.17 vs. limit=5.0 +2023-03-27 05:39:17,546 INFO [finetune.py:976] (4/7) Epoch 24, batch 4900, loss[loss=0.1666, simple_loss=0.2423, pruned_loss=0.04548, over 4784.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2461, pruned_loss=0.05112, over 952327.15 frames. ], batch size: 51, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:39:18,278 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136638.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:39:22,266 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=136643.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:39:26,546 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136650.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:39:42,308 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.600e+02 1.925e+02 2.438e+02 3.559e+02, threshold=3.849e+02, percent-clipped=0.0 +2023-03-27 05:39:55,030 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-27 05:39:57,218 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-27 05:39:59,909 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=136686.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:40:00,476 INFO [finetune.py:976] (4/7) Epoch 24, batch 4950, loss[loss=0.1572, simple_loss=0.2324, pruned_loss=0.041, over 4884.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2475, pruned_loss=0.05185, over 953138.56 frames. ], batch size: 32, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:40:03,951 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136691.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:40:05,836 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6788, 1.4886, 1.0534, 0.2868, 1.3305, 1.4925, 1.4632, 1.4366], + device='cuda:4'), covar=tensor([0.1010, 0.0949, 0.1500, 0.2186, 0.1468, 0.2389, 0.2544, 0.0978], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0190, 0.0198, 0.0180, 0.0209, 0.0208, 0.0222, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:40:08,770 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=136698.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:40:12,483 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136704.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:40:33,840 INFO [finetune.py:976] (4/7) Epoch 24, batch 5000, loss[loss=0.1572, simple_loss=0.2266, pruned_loss=0.04396, over 4790.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2467, pruned_loss=0.05166, over 954482.62 frames. ], batch size: 29, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:40:54,396 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2069, 3.6569, 3.8416, 3.9986, 4.0140, 3.7742, 4.3042, 1.5016], + device='cuda:4'), covar=tensor([0.0819, 0.0837, 0.0974, 0.1113, 0.1129, 0.1543, 0.0729, 0.5512], + device='cuda:4'), in_proj_covar=tensor([0.0346, 0.0245, 0.0282, 0.0292, 0.0337, 0.0284, 0.0305, 0.0300], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:41:02,608 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.076e+02 1.520e+02 1.782e+02 2.173e+02 3.913e+02, threshold=3.563e+02, percent-clipped=1.0 +2023-03-27 05:41:17,086 INFO [finetune.py:976] (4/7) Epoch 24, batch 5050, loss[loss=0.1826, simple_loss=0.2527, pruned_loss=0.05624, over 4830.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2452, pruned_loss=0.05201, over 952134.55 frames. ], batch size: 40, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:41:25,202 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136798.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:41:30,673 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=136806.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:41:49,360 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0771, 1.0119, 0.9711, 1.1694, 1.1980, 1.1635, 1.0320, 0.9549], + device='cuda:4'), covar=tensor([0.0404, 0.0326, 0.0683, 0.0329, 0.0313, 0.0432, 0.0389, 0.0443], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0107, 0.0146, 0.0112, 0.0101, 0.0114, 0.0103, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.8200e-05, 8.1941e-05, 1.1409e-04, 8.5861e-05, 7.8592e-05, 8.4392e-05, + 7.6503e-05, 8.6301e-05], device='cuda:4') +2023-03-27 05:41:49,839 INFO [finetune.py:976] (4/7) Epoch 24, batch 5100, loss[loss=0.1643, simple_loss=0.232, pruned_loss=0.04834, over 4831.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2411, pruned_loss=0.0502, over 952049.53 frames. ], batch size: 30, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:41:56,305 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=136846.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:41:56,726 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-27 05:42:11,832 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.036e+02 1.576e+02 1.921e+02 2.257e+02 4.191e+02, threshold=3.841e+02, percent-clipped=1.0 +2023-03-27 05:42:13,196 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=136867.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:42:35,177 INFO [finetune.py:976] (4/7) Epoch 24, batch 5150, loss[loss=0.1784, simple_loss=0.2586, pruned_loss=0.04907, over 4813.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2416, pruned_loss=0.05064, over 952995.40 frames. ], batch size: 51, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:43:16,539 INFO [finetune.py:976] (4/7) Epoch 24, batch 5200, loss[loss=0.1797, simple_loss=0.2584, pruned_loss=0.0505, over 4913.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2441, pruned_loss=0.05127, over 951902.18 frames. ], batch size: 36, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:43:21,974 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5025, 1.4123, 1.3894, 1.4287, 1.0434, 3.1189, 1.1964, 1.5377], + device='cuda:4'), covar=tensor([0.3333, 0.2665, 0.2226, 0.2462, 0.1899, 0.0225, 0.2908, 0.1424], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0123, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 05:43:35,514 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.133e+02 1.647e+02 1.940e+02 2.397e+02 3.428e+02, threshold=3.879e+02, percent-clipped=0.0 +2023-03-27 05:43:48,851 INFO [finetune.py:976] (4/7) Epoch 24, batch 5250, loss[loss=0.1552, simple_loss=0.2323, pruned_loss=0.03905, over 4821.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2444, pruned_loss=0.05075, over 952165.03 frames. ], batch size: 38, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:43:51,884 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=136991.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:43:57,217 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=136999.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:44:22,624 INFO [finetune.py:976] (4/7) Epoch 24, batch 5300, loss[loss=0.1678, simple_loss=0.2483, pruned_loss=0.04368, over 4831.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2459, pruned_loss=0.05102, over 952101.16 frames. ], batch size: 47, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:44:23,948 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=137039.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:44:42,412 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.600e+02 1.832e+02 2.198e+02 3.821e+02, threshold=3.665e+02, percent-clipped=0.0 +2023-03-27 05:45:04,673 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7198, 1.2024, 0.8570, 1.4518, 2.1256, 1.0706, 1.3921, 1.4684], + device='cuda:4'), covar=tensor([0.1600, 0.2201, 0.1981, 0.1364, 0.1970, 0.1998, 0.1623, 0.2174], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0093, 0.0110, 0.0092, 0.0119, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 05:45:05,284 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6646, 1.1849, 0.7925, 1.4502, 2.0304, 1.4066, 1.3726, 1.5423], + device='cuda:4'), covar=tensor([0.1466, 0.2045, 0.1970, 0.1237, 0.1967, 0.1931, 0.1495, 0.1860], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0093, 0.0110, 0.0092, 0.0119, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 05:45:05,791 INFO [finetune.py:976] (4/7) Epoch 24, batch 5350, loss[loss=0.1652, simple_loss=0.2402, pruned_loss=0.04507, over 4833.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2468, pruned_loss=0.05104, over 952291.51 frames. ], batch size: 49, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:45:38,840 INFO [finetune.py:976] (4/7) Epoch 24, batch 5400, loss[loss=0.1831, simple_loss=0.2556, pruned_loss=0.05536, over 4821.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2442, pruned_loss=0.04973, over 952703.24 frames. ], batch size: 30, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:45:57,158 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=137162.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:45:58,905 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.720e+01 1.449e+02 1.770e+02 2.209e+02 4.288e+02, threshold=3.541e+02, percent-clipped=1.0 +2023-03-27 05:46:22,823 INFO [finetune.py:976] (4/7) Epoch 24, batch 5450, loss[loss=0.1338, simple_loss=0.2092, pruned_loss=0.02918, over 4824.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2413, pruned_loss=0.04948, over 953538.41 frames. ], batch size: 38, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:46:31,921 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7653, 1.6618, 1.4465, 1.8360, 2.3072, 1.8614, 1.6402, 1.4364], + device='cuda:4'), covar=tensor([0.2336, 0.2184, 0.2222, 0.1809, 0.1607, 0.1349, 0.2489, 0.2037], + device='cuda:4'), in_proj_covar=tensor([0.0242, 0.0210, 0.0213, 0.0194, 0.0242, 0.0188, 0.0214, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:46:37,305 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8244, 1.6722, 2.1258, 1.3571, 1.9203, 2.1826, 1.6049, 2.2867], + device='cuda:4'), covar=tensor([0.1376, 0.2216, 0.1383, 0.1917, 0.0959, 0.1218, 0.3168, 0.0839], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0206, 0.0191, 0.0189, 0.0174, 0.0213, 0.0216, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:46:51,331 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-27 05:46:55,994 INFO [finetune.py:976] (4/7) Epoch 24, batch 5500, loss[loss=0.1626, simple_loss=0.2226, pruned_loss=0.05129, over 4126.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2392, pruned_loss=0.04894, over 956082.69 frames. ], batch size: 65, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:47:13,426 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.031e+02 1.493e+02 1.897e+02 2.213e+02 3.719e+02, threshold=3.794e+02, percent-clipped=2.0 +2023-03-27 05:47:36,857 INFO [finetune.py:976] (4/7) Epoch 24, batch 5550, loss[loss=0.1555, simple_loss=0.2304, pruned_loss=0.04035, over 4832.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.241, pruned_loss=0.04936, over 956918.86 frames. ], batch size: 30, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:47:47,644 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=137299.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:48:13,725 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5443, 2.3679, 2.0520, 1.0991, 2.3256, 1.9025, 1.7958, 2.2361], + device='cuda:4'), covar=tensor([0.0935, 0.0820, 0.1710, 0.2061, 0.1359, 0.2398, 0.2201, 0.0993], + device='cuda:4'), in_proj_covar=tensor([0.0169, 0.0190, 0.0199, 0.0180, 0.0209, 0.0208, 0.0222, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:48:20,577 INFO [finetune.py:976] (4/7) Epoch 24, batch 5600, loss[loss=0.1648, simple_loss=0.238, pruned_loss=0.04586, over 4827.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2433, pruned_loss=0.04988, over 955713.91 frames. ], batch size: 33, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:48:26,398 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=137347.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:48:26,458 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1592, 2.0598, 2.2110, 1.3670, 2.1078, 2.2535, 2.1472, 1.8745], + device='cuda:4'), covar=tensor([0.0553, 0.0707, 0.0640, 0.0918, 0.0761, 0.0728, 0.0616, 0.1109], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0136, 0.0138, 0.0118, 0.0125, 0.0137, 0.0138, 0.0159], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:48:37,702 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.121e+02 1.584e+02 1.843e+02 2.259e+02 3.753e+02, threshold=3.686e+02, percent-clipped=0.0 +2023-03-27 05:48:51,115 INFO [finetune.py:976] (4/7) Epoch 24, batch 5650, loss[loss=0.1902, simple_loss=0.2764, pruned_loss=0.05199, over 4814.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2466, pruned_loss=0.05056, over 957409.35 frames. ], batch size: 38, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:48:57,134 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0188, 1.9355, 1.6147, 1.8400, 1.9869, 1.7086, 2.1828, 2.0534], + device='cuda:4'), covar=tensor([0.1332, 0.1906, 0.2857, 0.2422, 0.2457, 0.1621, 0.2922, 0.1577], + device='cuda:4'), in_proj_covar=tensor([0.0190, 0.0190, 0.0237, 0.0255, 0.0251, 0.0207, 0.0216, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:49:08,663 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1326, 1.2121, 1.3544, 0.9214, 1.1244, 1.3082, 1.1718, 1.4346], + device='cuda:4'), covar=tensor([0.1067, 0.1907, 0.1116, 0.1264, 0.0873, 0.1002, 0.2742, 0.0770], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0206, 0.0192, 0.0189, 0.0174, 0.0213, 0.0216, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:49:20,946 INFO [finetune.py:976] (4/7) Epoch 24, batch 5700, loss[loss=0.1494, simple_loss=0.2098, pruned_loss=0.04455, over 4225.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2429, pruned_loss=0.04966, over 940332.37 frames. ], batch size: 18, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:49:35,758 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=137462.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:49:52,033 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.393e+01 1.414e+02 1.686e+02 2.138e+02 3.465e+02, threshold=3.373e+02, percent-clipped=0.0 +2023-03-27 05:49:52,049 INFO [finetune.py:976] (4/7) Epoch 25, batch 0, loss[loss=0.1653, simple_loss=0.2451, pruned_loss=0.04275, over 4818.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2451, pruned_loss=0.04275, over 4818.00 frames. ], batch size: 47, lr: 3.03e-03, grad_scale: 32.0 +2023-03-27 05:49:52,049 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 05:49:58,504 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6678, 3.6660, 3.4486, 1.5547, 3.6566, 2.8652, 0.7447, 2.5791], + device='cuda:4'), covar=tensor([0.1911, 0.1641, 0.1599, 0.3277, 0.1042, 0.0979, 0.3561, 0.1434], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0177, 0.0160, 0.0129, 0.0161, 0.0123, 0.0147, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 05:50:06,682 INFO [finetune.py:1010] (4/7) Epoch 25, validation: loss=0.1587, simple_loss=0.2267, pruned_loss=0.04536, over 2265189.00 frames. +2023-03-27 05:50:06,683 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6557MB +2023-03-27 05:50:46,518 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=137510.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:50:50,024 INFO [finetune.py:976] (4/7) Epoch 25, batch 50, loss[loss=0.1381, simple_loss=0.2096, pruned_loss=0.03332, over 4729.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2457, pruned_loss=0.0519, over 216456.42 frames. ], batch size: 54, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:50:52,336 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7264, 1.6353, 1.6133, 1.6989, 1.2424, 3.3892, 1.3796, 1.7257], + device='cuda:4'), covar=tensor([0.3181, 0.2446, 0.2083, 0.2218, 0.1671, 0.0208, 0.2427, 0.1262], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0123, 0.0113, 0.0095, 0.0094, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 05:50:54,349 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.69 vs. limit=5.0 +2023-03-27 05:51:06,087 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2448, 1.7570, 0.9762, 1.9452, 2.3850, 1.7032, 2.0765, 1.8423], + device='cuda:4'), covar=tensor([0.1217, 0.1765, 0.1832, 0.1043, 0.1680, 0.1664, 0.1159, 0.1812], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0093, 0.0109, 0.0092, 0.0119, 0.0092, 0.0097, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 05:51:09,928 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4468, 1.3002, 1.9665, 3.2976, 2.0584, 2.4386, 0.9704, 2.8164], + device='cuda:4'), covar=tensor([0.2149, 0.1990, 0.1631, 0.0873, 0.1064, 0.1323, 0.2167, 0.0617], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0116, 0.0133, 0.0163, 0.0101, 0.0136, 0.0124, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 05:51:12,271 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1648, 1.3734, 1.3906, 0.6686, 1.3706, 1.5916, 1.6330, 1.3101], + device='cuda:4'), covar=tensor([0.0862, 0.0554, 0.0605, 0.0492, 0.0551, 0.0588, 0.0382, 0.0751], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0150, 0.0127, 0.0123, 0.0131, 0.0130, 0.0142, 0.0149], + device='cuda:4'), out_proj_covar=tensor([8.9609e-05, 1.0756e-04, 9.0843e-05, 8.6264e-05, 9.2124e-05, 9.2654e-05, + 1.0144e-04, 1.0627e-04], device='cuda:4') +2023-03-27 05:51:25,238 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.123e+02 1.580e+02 1.851e+02 2.170e+02 4.183e+02, threshold=3.702e+02, percent-clipped=2.0 +2023-03-27 05:51:25,254 INFO [finetune.py:976] (4/7) Epoch 25, batch 100, loss[loss=0.2028, simple_loss=0.2767, pruned_loss=0.06441, over 4710.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2426, pruned_loss=0.05087, over 379941.69 frames. ], batch size: 54, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:51:59,265 INFO [finetune.py:976] (4/7) Epoch 25, batch 150, loss[loss=0.1481, simple_loss=0.2204, pruned_loss=0.03785, over 4753.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2392, pruned_loss=0.04938, over 508018.11 frames. ], batch size: 27, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:52:09,247 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-27 05:52:33,560 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.113e+02 1.544e+02 1.791e+02 2.141e+02 4.771e+02, threshold=3.582e+02, percent-clipped=2.0 +2023-03-27 05:52:33,576 INFO [finetune.py:976] (4/7) Epoch 25, batch 200, loss[loss=0.1748, simple_loss=0.2479, pruned_loss=0.05083, over 4824.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2368, pruned_loss=0.04828, over 607865.81 frames. ], batch size: 40, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:52:49,621 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-27 05:53:26,997 INFO [finetune.py:976] (4/7) Epoch 25, batch 250, loss[loss=0.206, simple_loss=0.2752, pruned_loss=0.06843, over 4914.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2403, pruned_loss=0.04977, over 685339.64 frames. ], batch size: 42, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:53:31,853 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1053, 2.2346, 1.8618, 2.2039, 2.0485, 2.1044, 2.0554, 2.9065], + device='cuda:4'), covar=tensor([0.3755, 0.4394, 0.3310, 0.4092, 0.4535, 0.2484, 0.4406, 0.1531], + device='cuda:4'), in_proj_covar=tensor([0.0291, 0.0265, 0.0236, 0.0277, 0.0260, 0.0229, 0.0257, 0.0239], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:54:00,392 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.518e+01 1.603e+02 1.998e+02 2.287e+02 4.515e+02, threshold=3.995e+02, percent-clipped=2.0 +2023-03-27 05:54:00,408 INFO [finetune.py:976] (4/7) Epoch 25, batch 300, loss[loss=0.1226, simple_loss=0.1951, pruned_loss=0.02504, over 4722.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2418, pruned_loss=0.04995, over 744607.84 frames. ], batch size: 23, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:54:02,806 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9399, 1.6767, 2.2893, 1.4265, 1.9561, 2.3099, 1.5014, 2.3163], + device='cuda:4'), covar=tensor([0.1276, 0.2236, 0.1431, 0.2180, 0.1015, 0.1253, 0.3032, 0.0841], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0206, 0.0192, 0.0190, 0.0174, 0.0213, 0.0215, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 05:54:04,628 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3699, 2.2037, 2.8357, 4.4105, 3.1395, 2.8539, 1.0957, 3.6620], + device='cuda:4'), covar=tensor([0.1573, 0.1241, 0.1241, 0.0477, 0.0652, 0.1106, 0.1928, 0.0349], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0117, 0.0134, 0.0165, 0.0102, 0.0138, 0.0125, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 05:54:15,832 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.57 vs. limit=2.0 +2023-03-27 05:54:22,455 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6387, 1.6178, 1.9711, 3.4353, 2.2765, 2.3379, 0.9231, 2.8102], + device='cuda:4'), covar=tensor([0.1874, 0.1449, 0.1492, 0.0578, 0.0825, 0.1294, 0.1984, 0.0507], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0117, 0.0135, 0.0165, 0.0103, 0.0138, 0.0126, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 05:54:33,832 INFO [finetune.py:976] (4/7) Epoch 25, batch 350, loss[loss=0.2097, simple_loss=0.2824, pruned_loss=0.06855, over 4910.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2457, pruned_loss=0.05108, over 792258.92 frames. ], batch size: 37, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:55:07,124 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.080e+02 1.541e+02 1.822e+02 2.129e+02 2.910e+02, threshold=3.644e+02, percent-clipped=0.0 +2023-03-27 05:55:07,140 INFO [finetune.py:976] (4/7) Epoch 25, batch 400, loss[loss=0.1695, simple_loss=0.2406, pruned_loss=0.04921, over 4831.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2465, pruned_loss=0.05073, over 827457.73 frames. ], batch size: 30, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:55:31,434 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=137892.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:55:33,200 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=137895.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:55:41,373 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0841, 1.3426, 1.2696, 1.2476, 1.4819, 2.4446, 1.3441, 1.4429], + device='cuda:4'), covar=tensor([0.1030, 0.1870, 0.1072, 0.0976, 0.1662, 0.0364, 0.1484, 0.1780], + device='cuda:4'), in_proj_covar=tensor([0.0074, 0.0082, 0.0073, 0.0076, 0.0090, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 05:55:53,569 INFO [finetune.py:976] (4/7) Epoch 25, batch 450, loss[loss=0.1687, simple_loss=0.2437, pruned_loss=0.04685, over 4830.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2446, pruned_loss=0.04979, over 856455.85 frames. ], batch size: 40, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:56:17,503 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-27 05:56:19,219 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=137953.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:56:20,998 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=137956.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:56:26,866 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.118e+02 1.489e+02 1.801e+02 2.267e+02 5.324e+02, threshold=3.602e+02, percent-clipped=3.0 +2023-03-27 05:56:26,882 INFO [finetune.py:976] (4/7) Epoch 25, batch 500, loss[loss=0.2157, simple_loss=0.273, pruned_loss=0.07917, over 4813.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2414, pruned_loss=0.04869, over 878822.99 frames. ], batch size: 38, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:56:44,650 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8464, 1.4676, 0.8897, 1.6993, 2.2544, 1.4024, 1.7535, 1.7403], + device='cuda:4'), covar=tensor([0.1391, 0.1964, 0.1814, 0.1122, 0.1721, 0.1677, 0.1292, 0.1807], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0110, 0.0093, 0.0119, 0.0092, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 05:57:01,548 INFO [finetune.py:976] (4/7) Epoch 25, batch 550, loss[loss=0.2127, simple_loss=0.2702, pruned_loss=0.07755, over 4826.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2395, pruned_loss=0.0485, over 894847.86 frames. ], batch size: 40, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:57:34,661 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.805e+01 1.427e+02 1.740e+02 1.994e+02 3.808e+02, threshold=3.480e+02, percent-clipped=1.0 +2023-03-27 05:57:34,677 INFO [finetune.py:976] (4/7) Epoch 25, batch 600, loss[loss=0.234, simple_loss=0.299, pruned_loss=0.08448, over 4726.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2391, pruned_loss=0.04865, over 908109.15 frames. ], batch size: 54, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:58:07,363 INFO [finetune.py:976] (4/7) Epoch 25, batch 650, loss[loss=0.2047, simple_loss=0.268, pruned_loss=0.07067, over 4818.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2437, pruned_loss=0.05036, over 917681.33 frames. ], batch size: 51, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:58:59,111 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.613e+02 1.882e+02 2.325e+02 5.163e+02, threshold=3.765e+02, percent-clipped=4.0 +2023-03-27 05:58:59,127 INFO [finetune.py:976] (4/7) Epoch 25, batch 700, loss[loss=0.1882, simple_loss=0.2527, pruned_loss=0.06187, over 4816.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2462, pruned_loss=0.0509, over 926328.50 frames. ], batch size: 39, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:59:26,315 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-27 05:59:32,528 INFO [finetune.py:976] (4/7) Epoch 25, batch 750, loss[loss=0.1814, simple_loss=0.2429, pruned_loss=0.05995, over 4866.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2466, pruned_loss=0.05102, over 934058.60 frames. ], batch size: 44, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 05:59:53,524 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=138248.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 05:59:55,847 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=138251.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:00:05,193 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.386e+01 1.513e+02 1.803e+02 2.270e+02 6.862e+02, threshold=3.605e+02, percent-clipped=3.0 +2023-03-27 06:00:05,209 INFO [finetune.py:976] (4/7) Epoch 25, batch 800, loss[loss=0.1963, simple_loss=0.262, pruned_loss=0.06533, over 4900.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2471, pruned_loss=0.05069, over 937971.38 frames. ], batch size: 43, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 06:00:08,354 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138270.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:00:19,831 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.52 vs. limit=2.0 +2023-03-27 06:00:38,347 INFO [finetune.py:976] (4/7) Epoch 25, batch 850, loss[loss=0.1688, simple_loss=0.2381, pruned_loss=0.04982, over 4826.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.245, pruned_loss=0.05014, over 942397.71 frames. ], batch size: 30, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 06:00:44,521 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138322.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:00:45,803 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1387, 2.1316, 1.8766, 2.2498, 2.0196, 2.1071, 2.0856, 2.9695], + device='cuda:4'), covar=tensor([0.3747, 0.5014, 0.3299, 0.4292, 0.4573, 0.2428, 0.4269, 0.1512], + device='cuda:4'), in_proj_covar=tensor([0.0291, 0.0264, 0.0235, 0.0276, 0.0259, 0.0229, 0.0257, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:00:54,539 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138331.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:01:24,851 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.388e+02 1.706e+02 2.091e+02 3.563e+02, threshold=3.412e+02, percent-clipped=0.0 +2023-03-27 06:01:24,867 INFO [finetune.py:976] (4/7) Epoch 25, batch 900, loss[loss=0.147, simple_loss=0.2163, pruned_loss=0.03888, over 4760.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2424, pruned_loss=0.0492, over 946567.81 frames. ], batch size: 54, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 06:01:34,649 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6739, 1.6429, 1.5730, 1.7105, 1.1491, 3.3397, 1.2927, 1.7726], + device='cuda:4'), covar=tensor([0.3298, 0.2409, 0.2032, 0.2242, 0.1693, 0.0207, 0.2441, 0.1186], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0124, 0.0113, 0.0096, 0.0094, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 06:01:36,371 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138383.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:01:55,428 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5638, 1.4206, 1.3396, 1.5740, 1.6783, 1.4766, 1.1097, 1.3435], + device='cuda:4'), covar=tensor([0.1792, 0.1830, 0.1655, 0.1331, 0.1466, 0.1211, 0.2637, 0.1648], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0213, 0.0216, 0.0198, 0.0245, 0.0192, 0.0218, 0.0206], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:01:57,678 INFO [finetune.py:976] (4/7) Epoch 25, batch 950, loss[loss=0.1483, simple_loss=0.221, pruned_loss=0.03779, over 4818.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.242, pruned_loss=0.04971, over 950062.36 frames. ], batch size: 38, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 06:02:03,938 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6667, 1.6796, 1.3446, 1.5477, 1.9667, 1.9056, 1.6169, 1.4227], + device='cuda:4'), covar=tensor([0.0302, 0.0331, 0.0623, 0.0349, 0.0217, 0.0424, 0.0388, 0.0433], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0107, 0.0146, 0.0112, 0.0101, 0.0113, 0.0102, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.7886e-05, 8.1737e-05, 1.1419e-04, 8.5897e-05, 7.8502e-05, 8.3867e-05, + 7.5927e-05, 8.6010e-05], device='cuda:4') +2023-03-27 06:02:05,998 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.74 vs. limit=5.0 +2023-03-27 06:02:12,340 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1782, 1.3060, 1.3427, 0.6992, 1.3177, 1.5831, 1.6262, 1.2595], + device='cuda:4'), covar=tensor([0.1005, 0.0643, 0.0530, 0.0556, 0.0542, 0.0641, 0.0331, 0.0728], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0149, 0.0127, 0.0122, 0.0131, 0.0129, 0.0141, 0.0148], + device='cuda:4'), out_proj_covar=tensor([8.9067e-05, 1.0678e-04, 9.0275e-05, 8.5682e-05, 9.1593e-05, 9.1962e-05, + 1.0047e-04, 1.0553e-04], device='cuda:4') +2023-03-27 06:02:30,848 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.004e+02 1.447e+02 1.781e+02 2.298e+02 4.571e+02, threshold=3.563e+02, percent-clipped=3.0 +2023-03-27 06:02:30,863 INFO [finetune.py:976] (4/7) Epoch 25, batch 1000, loss[loss=0.1248, simple_loss=0.1978, pruned_loss=0.02595, over 4697.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2439, pruned_loss=0.05083, over 950593.35 frames. ], batch size: 23, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 06:02:56,199 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4508, 2.1682, 2.8695, 1.7576, 2.5197, 2.8572, 2.0131, 2.8261], + device='cuda:4'), covar=tensor([0.1312, 0.2204, 0.1497, 0.2248, 0.0915, 0.1375, 0.2787, 0.0931], + device='cuda:4'), in_proj_covar=tensor([0.0191, 0.0207, 0.0191, 0.0189, 0.0174, 0.0213, 0.0216, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:03:00,328 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.2809, 2.9847, 3.0854, 3.2088, 3.0685, 2.9418, 3.3570, 0.9761], + device='cuda:4'), covar=tensor([0.1163, 0.1001, 0.1087, 0.1290, 0.1653, 0.1670, 0.1103, 0.5708], + device='cuda:4'), in_proj_covar=tensor([0.0345, 0.0245, 0.0280, 0.0292, 0.0335, 0.0284, 0.0304, 0.0299], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:03:03,759 INFO [finetune.py:976] (4/7) Epoch 25, batch 1050, loss[loss=0.1782, simple_loss=0.249, pruned_loss=0.05374, over 4770.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2466, pruned_loss=0.05121, over 948585.72 frames. ], batch size: 28, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 06:03:15,707 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8887, 1.8830, 1.7122, 1.8882, 1.6329, 4.6022, 1.8626, 2.1062], + device='cuda:4'), covar=tensor([0.3095, 0.2384, 0.2098, 0.2266, 0.1437, 0.0116, 0.2187, 0.1307], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0123, 0.0112, 0.0096, 0.0094, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 06:03:25,371 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=138548.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:03:27,133 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=138551.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:03:39,079 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.570e+02 1.874e+02 2.177e+02 7.699e+02, threshold=3.747e+02, percent-clipped=3.0 +2023-03-27 06:03:39,095 INFO [finetune.py:976] (4/7) Epoch 25, batch 1100, loss[loss=0.1635, simple_loss=0.2432, pruned_loss=0.04191, over 4792.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2474, pruned_loss=0.05148, over 949399.81 frames. ], batch size: 45, lr: 3.02e-03, grad_scale: 32.0 +2023-03-27 06:04:15,225 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=2.03 vs. limit=2.0 +2023-03-27 06:04:17,868 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=138596.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:04:19,634 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=138599.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:04:29,871 INFO [finetune.py:976] (4/7) Epoch 25, batch 1150, loss[loss=0.218, simple_loss=0.2799, pruned_loss=0.07799, over 4789.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2482, pruned_loss=0.05152, over 951633.40 frames. ], batch size: 29, lr: 3.02e-03, grad_scale: 64.0 +2023-03-27 06:04:39,029 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=138626.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:04:52,181 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5792, 1.5741, 1.4673, 1.6898, 1.2453, 3.5742, 1.3746, 1.7465], + device='cuda:4'), covar=tensor([0.3228, 0.2502, 0.2116, 0.2307, 0.1699, 0.0221, 0.2464, 0.1263], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0124, 0.0113, 0.0096, 0.0094, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 06:04:58,858 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5706, 1.6059, 1.3464, 1.6179, 1.9081, 1.7954, 1.5797, 1.4016], + device='cuda:4'), covar=tensor([0.0342, 0.0330, 0.0586, 0.0287, 0.0203, 0.0483, 0.0329, 0.0419], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0108, 0.0147, 0.0113, 0.0102, 0.0115, 0.0103, 0.0114], + device='cuda:4'), out_proj_covar=tensor([7.8626e-05, 8.2718e-05, 1.1515e-04, 8.6753e-05, 7.9317e-05, 8.4797e-05, + 7.6653e-05, 8.6964e-05], device='cuda:4') +2023-03-27 06:05:03,419 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.832e+01 1.520e+02 1.733e+02 2.222e+02 3.582e+02, threshold=3.466e+02, percent-clipped=0.0 +2023-03-27 06:05:03,435 INFO [finetune.py:976] (4/7) Epoch 25, batch 1200, loss[loss=0.1565, simple_loss=0.2249, pruned_loss=0.04409, over 4821.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2467, pruned_loss=0.05109, over 953299.97 frames. ], batch size: 40, lr: 3.02e-03, grad_scale: 64.0 +2023-03-27 06:05:13,800 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=138678.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:05:15,610 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138681.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:05:18,700 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138686.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:05:25,916 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138697.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:05:28,978 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0027, 0.9756, 0.8984, 1.0824, 1.1789, 1.0884, 0.9949, 0.9091], + device='cuda:4'), covar=tensor([0.0439, 0.0337, 0.0752, 0.0376, 0.0306, 0.0521, 0.0359, 0.0437], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0108, 0.0147, 0.0113, 0.0102, 0.0115, 0.0103, 0.0114], + device='cuda:4'), out_proj_covar=tensor([7.8745e-05, 8.2774e-05, 1.1523e-04, 8.6788e-05, 7.9350e-05, 8.4945e-05, + 7.6567e-05, 8.6954e-05], device='cuda:4') +2023-03-27 06:05:37,221 INFO [finetune.py:976] (4/7) Epoch 25, batch 1250, loss[loss=0.1676, simple_loss=0.2357, pruned_loss=0.04975, over 4822.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2445, pruned_loss=0.05049, over 954852.05 frames. ], batch size: 38, lr: 3.02e-03, grad_scale: 64.0 +2023-03-27 06:05:55,590 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138742.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:05:59,141 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138747.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:06:05,731 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138758.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:06:11,261 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4012, 2.3010, 1.8518, 2.3729, 2.2386, 2.0238, 2.6201, 2.3869], + device='cuda:4'), covar=tensor([0.1267, 0.1859, 0.2777, 0.2370, 0.2393, 0.1601, 0.2832, 0.1632], + device='cuda:4'), in_proj_covar=tensor([0.0187, 0.0189, 0.0233, 0.0252, 0.0248, 0.0204, 0.0214, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:06:12,323 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.434e+02 1.700e+02 2.124e+02 3.591e+02, threshold=3.400e+02, percent-clipped=1.0 +2023-03-27 06:06:12,339 INFO [finetune.py:976] (4/7) Epoch 25, batch 1300, loss[loss=0.1848, simple_loss=0.2444, pruned_loss=0.06266, over 4840.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2406, pruned_loss=0.04913, over 955761.35 frames. ], batch size: 30, lr: 3.02e-03, grad_scale: 64.0 +2023-03-27 06:06:30,149 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9769, 1.1957, 0.8386, 1.7373, 2.4685, 1.8631, 1.4900, 1.6595], + device='cuda:4'), covar=tensor([0.1628, 0.2403, 0.2097, 0.1385, 0.1803, 0.1973, 0.1651, 0.2214], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0093, 0.0109, 0.0092, 0.0118, 0.0092, 0.0097, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 06:06:31,734 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-27 06:06:57,380 INFO [finetune.py:976] (4/7) Epoch 25, batch 1350, loss[loss=0.121, simple_loss=0.1917, pruned_loss=0.02516, over 4216.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2405, pruned_loss=0.04957, over 955801.65 frames. ], batch size: 18, lr: 3.02e-03, grad_scale: 64.0 +2023-03-27 06:07:31,277 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.573e+02 1.999e+02 2.321e+02 4.595e+02, threshold=3.999e+02, percent-clipped=3.0 +2023-03-27 06:07:31,293 INFO [finetune.py:976] (4/7) Epoch 25, batch 1400, loss[loss=0.1679, simple_loss=0.2502, pruned_loss=0.04282, over 4905.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2404, pruned_loss=0.04836, over 955415.00 frames. ], batch size: 43, lr: 3.02e-03, grad_scale: 64.0 +2023-03-27 06:08:01,086 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=138909.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:08:04,579 INFO [finetune.py:976] (4/7) Epoch 25, batch 1450, loss[loss=0.1475, simple_loss=0.2218, pruned_loss=0.03662, over 4866.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2431, pruned_loss=0.04891, over 956268.64 frames. ], batch size: 31, lr: 3.01e-03, grad_scale: 64.0 +2023-03-27 06:08:06,363 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6488, 1.2379, 2.1041, 3.3606, 2.1992, 2.5122, 1.0441, 2.9811], + device='cuda:4'), covar=tensor([0.1944, 0.1956, 0.1542, 0.0862, 0.0980, 0.1525, 0.2100, 0.0522], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0116, 0.0134, 0.0165, 0.0102, 0.0137, 0.0125, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 06:08:11,751 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=138926.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:08:38,070 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.768e+01 1.472e+02 1.793e+02 2.201e+02 3.947e+02, threshold=3.587e+02, percent-clipped=0.0 +2023-03-27 06:08:38,086 INFO [finetune.py:976] (4/7) Epoch 25, batch 1500, loss[loss=0.1916, simple_loss=0.2538, pruned_loss=0.06467, over 4820.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2457, pruned_loss=0.05043, over 955145.80 frames. ], batch size: 33, lr: 3.01e-03, grad_scale: 64.0 +2023-03-27 06:08:41,065 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1757, 2.1638, 2.2115, 1.6369, 2.1951, 2.4065, 2.2377, 1.7629], + device='cuda:4'), covar=tensor([0.0687, 0.0733, 0.0751, 0.0924, 0.0647, 0.0760, 0.0701, 0.1340], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0138, 0.0141, 0.0120, 0.0127, 0.0140, 0.0140, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:08:41,676 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=138970.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 06:08:44,019 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=138974.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:08:46,500 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=138978.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:09:23,190 INFO [finetune.py:976] (4/7) Epoch 25, batch 1550, loss[loss=0.1903, simple_loss=0.2641, pruned_loss=0.05828, over 4926.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2451, pruned_loss=0.04999, over 956469.34 frames. ], batch size: 33, lr: 3.01e-03, grad_scale: 64.0 +2023-03-27 06:09:34,899 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139026.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:09:46,457 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-27 06:09:47,267 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139037.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:09:50,278 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139042.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:09:53,870 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3337, 1.3975, 1.7347, 1.7027, 1.5381, 3.1161, 1.3213, 1.5434], + device='cuda:4'), covar=tensor([0.1003, 0.1785, 0.1096, 0.0861, 0.1536, 0.0254, 0.1509, 0.1677], + device='cuda:4'), in_proj_covar=tensor([0.0074, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 06:09:57,456 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139053.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:10:00,587 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8311, 1.7545, 1.5574, 1.9465, 2.4353, 2.0455, 1.8198, 1.5023], + device='cuda:4'), covar=tensor([0.2238, 0.1981, 0.1909, 0.1609, 0.1509, 0.1129, 0.2111, 0.2007], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0212, 0.0215, 0.0198, 0.0244, 0.0191, 0.0217, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:10:05,107 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.522e+01 1.375e+02 1.683e+02 2.077e+02 3.862e+02, threshold=3.366e+02, percent-clipped=3.0 +2023-03-27 06:10:05,123 INFO [finetune.py:976] (4/7) Epoch 25, batch 1600, loss[loss=0.1555, simple_loss=0.2233, pruned_loss=0.0439, over 4801.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2429, pruned_loss=0.04951, over 956427.87 frames. ], batch size: 29, lr: 3.01e-03, grad_scale: 64.0 +2023-03-27 06:10:11,798 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5240, 1.4244, 1.2888, 1.4742, 1.7802, 1.6507, 1.5076, 1.2956], + device='cuda:4'), covar=tensor([0.0335, 0.0304, 0.0633, 0.0299, 0.0208, 0.0435, 0.0303, 0.0421], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0107, 0.0146, 0.0112, 0.0101, 0.0114, 0.0103, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.8105e-05, 8.1949e-05, 1.1416e-04, 8.5830e-05, 7.8619e-05, 8.4122e-05, + 7.6072e-05, 8.5849e-05], device='cuda:4') +2023-03-27 06:10:25,265 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9669, 1.4696, 2.0490, 1.9695, 1.7760, 1.7152, 1.9289, 1.9192], + device='cuda:4'), covar=tensor([0.3933, 0.4237, 0.3207, 0.3942, 0.5196, 0.4004, 0.4843, 0.3277], + device='cuda:4'), in_proj_covar=tensor([0.0262, 0.0245, 0.0265, 0.0291, 0.0291, 0.0268, 0.0297, 0.0249], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:10:38,957 INFO [finetune.py:976] (4/7) Epoch 25, batch 1650, loss[loss=0.1633, simple_loss=0.2315, pruned_loss=0.04757, over 4695.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2415, pruned_loss=0.04888, over 956873.96 frames. ], batch size: 23, lr: 3.01e-03, grad_scale: 64.0 +2023-03-27 06:10:46,062 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2454, 3.6982, 3.8862, 4.0359, 4.0405, 3.8420, 4.3441, 1.5338], + device='cuda:4'), covar=tensor([0.0808, 0.0886, 0.0998, 0.1109, 0.1137, 0.1572, 0.0754, 0.5654], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0248, 0.0282, 0.0296, 0.0338, 0.0286, 0.0306, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:11:06,090 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9850, 1.7347, 2.3288, 1.6327, 2.1516, 2.3316, 1.6394, 2.4313], + device='cuda:4'), covar=tensor([0.1406, 0.2287, 0.1356, 0.1893, 0.0927, 0.1488, 0.3029, 0.0863], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0208, 0.0193, 0.0191, 0.0175, 0.0214, 0.0218, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:11:12,574 INFO [finetune.py:976] (4/7) Epoch 25, batch 1700, loss[loss=0.1473, simple_loss=0.2129, pruned_loss=0.04084, over 4344.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2405, pruned_loss=0.04896, over 957080.24 frames. ], batch size: 19, lr: 3.01e-03, grad_scale: 32.0 +2023-03-27 06:11:13,176 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.831e+01 1.435e+02 1.759e+02 2.193e+02 3.727e+02, threshold=3.518e+02, percent-clipped=3.0 +2023-03-27 06:11:19,407 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.71 vs. limit=5.0 +2023-03-27 06:11:43,662 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-27 06:11:56,395 INFO [finetune.py:976] (4/7) Epoch 25, batch 1750, loss[loss=0.2042, simple_loss=0.2788, pruned_loss=0.06484, over 4264.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2431, pruned_loss=0.04994, over 955235.41 frames. ], batch size: 65, lr: 3.01e-03, grad_scale: 32.0 +2023-03-27 06:12:11,570 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8784, 1.7680, 1.4723, 1.4491, 1.8723, 1.6071, 1.7903, 1.8763], + device='cuda:4'), covar=tensor([0.1483, 0.2118, 0.3212, 0.2745, 0.2593, 0.1729, 0.2801, 0.1866], + device='cuda:4'), in_proj_covar=tensor([0.0187, 0.0190, 0.0234, 0.0253, 0.0249, 0.0205, 0.0214, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:12:20,689 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139238.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:12:39,404 INFO [finetune.py:976] (4/7) Epoch 25, batch 1800, loss[loss=0.2014, simple_loss=0.2777, pruned_loss=0.06255, over 4818.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2457, pruned_loss=0.05037, over 953535.95 frames. ], batch size: 41, lr: 3.01e-03, grad_scale: 32.0 +2023-03-27 06:12:39,471 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139265.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 06:12:39,961 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.974e+01 1.553e+02 1.833e+02 2.131e+02 4.022e+02, threshold=3.667e+02, percent-clipped=3.0 +2023-03-27 06:12:46,662 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139276.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:12:49,091 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.4935, 3.8352, 4.0844, 4.2944, 4.2291, 3.9973, 4.5907, 1.5283], + device='cuda:4'), covar=tensor([0.0804, 0.0891, 0.0934, 0.1046, 0.1334, 0.1651, 0.0669, 0.5829], + device='cuda:4'), in_proj_covar=tensor([0.0351, 0.0249, 0.0282, 0.0297, 0.0339, 0.0287, 0.0307, 0.0303], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:12:55,155 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139289.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:13:02,257 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139299.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:13:13,308 INFO [finetune.py:976] (4/7) Epoch 25, batch 1850, loss[loss=0.1557, simple_loss=0.2335, pruned_loss=0.03892, over 4832.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2467, pruned_loss=0.0504, over 955165.22 frames. ], batch size: 30, lr: 3.01e-03, grad_scale: 32.0 +2023-03-27 06:13:27,854 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139337.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:13:27,904 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139337.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 06:13:30,847 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139342.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:13:30,917 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7956, 2.7128, 2.4287, 2.9482, 2.6566, 2.5900, 2.6388, 3.5380], + device='cuda:4'), covar=tensor([0.3509, 0.4348, 0.3098, 0.3660, 0.3849, 0.2539, 0.3782, 0.1441], + device='cuda:4'), in_proj_covar=tensor([0.0291, 0.0264, 0.0236, 0.0276, 0.0259, 0.0230, 0.0257, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:13:36,167 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139350.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:13:38,903 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139353.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:13:46,545 INFO [finetune.py:976] (4/7) Epoch 25, batch 1900, loss[loss=0.1689, simple_loss=0.2461, pruned_loss=0.04585, over 4885.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2463, pruned_loss=0.05036, over 953648.89 frames. ], batch size: 43, lr: 3.01e-03, grad_scale: 32.0 +2023-03-27 06:13:47,140 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.007e+02 1.553e+02 1.835e+02 2.150e+02 3.557e+02, threshold=3.671e+02, percent-clipped=0.0 +2023-03-27 06:13:48,378 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9231, 1.7492, 2.2506, 1.4788, 2.0803, 2.1671, 1.6526, 2.3823], + device='cuda:4'), covar=tensor([0.1269, 0.1978, 0.1367, 0.1942, 0.0860, 0.1386, 0.2531, 0.0821], + device='cuda:4'), in_proj_covar=tensor([0.0194, 0.0209, 0.0194, 0.0193, 0.0177, 0.0216, 0.0219, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:13:54,180 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7759, 1.3730, 0.8092, 1.6329, 2.1519, 1.4206, 1.7091, 1.7413], + device='cuda:4'), covar=tensor([0.1485, 0.2013, 0.2013, 0.1217, 0.1976, 0.1941, 0.1326, 0.1848], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0110, 0.0092, 0.0119, 0.0093, 0.0097, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 06:13:59,666 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139385.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:14:03,137 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139390.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:14:10,505 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139401.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:14:19,885 INFO [finetune.py:976] (4/7) Epoch 25, batch 1950, loss[loss=0.1327, simple_loss=0.2099, pruned_loss=0.02771, over 4021.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2445, pruned_loss=0.05006, over 953023.39 frames. ], batch size: 17, lr: 3.01e-03, grad_scale: 32.0 +2023-03-27 06:15:09,935 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4664, 2.4218, 2.1556, 1.1815, 2.3433, 1.9068, 1.7387, 2.2647], + device='cuda:4'), covar=tensor([0.0972, 0.0727, 0.1314, 0.1962, 0.1170, 0.1942, 0.2075, 0.0945], + device='cuda:4'), in_proj_covar=tensor([0.0171, 0.0192, 0.0201, 0.0182, 0.0209, 0.0211, 0.0224, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:15:11,618 INFO [finetune.py:976] (4/7) Epoch 25, batch 2000, loss[loss=0.1719, simple_loss=0.2405, pruned_loss=0.05162, over 4817.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2414, pruned_loss=0.04915, over 953074.77 frames. ], batch size: 33, lr: 3.01e-03, grad_scale: 32.0 +2023-03-27 06:15:12,707 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.253e+01 1.362e+02 1.721e+02 2.187e+02 3.038e+02, threshold=3.442e+02, percent-clipped=0.0 +2023-03-27 06:15:28,886 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7658, 1.9783, 1.5774, 1.8529, 2.3004, 2.3055, 2.0017, 1.8972], + device='cuda:4'), covar=tensor([0.0408, 0.0358, 0.0592, 0.0337, 0.0306, 0.0581, 0.0353, 0.0376], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0107, 0.0146, 0.0112, 0.0101, 0.0114, 0.0102, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.8073e-05, 8.1923e-05, 1.1396e-04, 8.5782e-05, 7.8283e-05, 8.4262e-05, + 7.6081e-05, 8.5726e-05], device='cuda:4') +2023-03-27 06:15:45,239 INFO [finetune.py:976] (4/7) Epoch 25, batch 2050, loss[loss=0.1451, simple_loss=0.2158, pruned_loss=0.03718, over 4860.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2388, pruned_loss=0.04853, over 952693.86 frames. ], batch size: 31, lr: 3.01e-03, grad_scale: 32.0 +2023-03-27 06:16:08,066 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-27 06:16:18,436 INFO [finetune.py:976] (4/7) Epoch 25, batch 2100, loss[loss=0.1706, simple_loss=0.2435, pruned_loss=0.04888, over 4930.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2387, pruned_loss=0.04853, over 953809.53 frames. ], batch size: 33, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:16:18,549 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139565.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 06:16:20,121 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.352e+01 1.449e+02 1.714e+02 2.109e+02 3.824e+02, threshold=3.428e+02, percent-clipped=2.0 +2023-03-27 06:16:23,221 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139571.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:16:30,374 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139582.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:16:38,030 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139594.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:16:41,862 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-27 06:16:44,786 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5188, 1.3692, 1.2337, 1.5645, 1.6240, 1.5404, 0.9869, 1.2537], + device='cuda:4'), covar=tensor([0.2322, 0.2223, 0.2100, 0.1719, 0.1673, 0.1265, 0.2627, 0.1965], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0212, 0.0216, 0.0199, 0.0245, 0.0191, 0.0217, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:16:50,722 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139613.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:16:52,382 INFO [finetune.py:976] (4/7) Epoch 25, batch 2150, loss[loss=0.2525, simple_loss=0.3181, pruned_loss=0.09347, over 4791.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2425, pruned_loss=0.04979, over 953498.16 frames. ], batch size: 51, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:17:09,455 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139632.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 06:17:09,490 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139632.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:17:21,671 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139643.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:17:27,568 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139645.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:17:43,826 INFO [finetune.py:976] (4/7) Epoch 25, batch 2200, loss[loss=0.2109, simple_loss=0.2751, pruned_loss=0.07336, over 4796.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2457, pruned_loss=0.05088, over 953856.68 frames. ], batch size: 51, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:17:45,450 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.512e+02 1.789e+02 2.111e+02 3.462e+02, threshold=3.578e+02, percent-clipped=1.0 +2023-03-27 06:18:08,186 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8246, 1.5635, 0.8164, 1.6800, 2.1644, 1.4567, 1.7163, 1.7421], + device='cuda:4'), covar=tensor([0.1545, 0.1960, 0.1961, 0.1223, 0.1913, 0.1858, 0.1354, 0.1960], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0094, 0.0109, 0.0092, 0.0119, 0.0093, 0.0097, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 06:18:17,067 INFO [finetune.py:976] (4/7) Epoch 25, batch 2250, loss[loss=0.1896, simple_loss=0.2613, pruned_loss=0.05893, over 4895.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2474, pruned_loss=0.05176, over 954969.38 frames. ], batch size: 36, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:18:50,822 INFO [finetune.py:976] (4/7) Epoch 25, batch 2300, loss[loss=0.1367, simple_loss=0.2246, pruned_loss=0.02442, over 4778.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2459, pruned_loss=0.05029, over 955285.08 frames. ], batch size: 29, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:18:52,009 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.533e+02 1.822e+02 2.118e+02 3.916e+02, threshold=3.645e+02, percent-clipped=1.0 +2023-03-27 06:18:53,808 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139769.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:19:05,191 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.99 vs. limit=5.0 +2023-03-27 06:19:06,795 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139789.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:19:16,871 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139804.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:19:18,137 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139806.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:19:23,931 INFO [finetune.py:976] (4/7) Epoch 25, batch 2350, loss[loss=0.1651, simple_loss=0.225, pruned_loss=0.0526, over 4042.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2437, pruned_loss=0.04999, over 953139.22 frames. ], batch size: 17, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:19:35,062 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139830.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:19:54,879 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=139850.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:19:54,888 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139850.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:20:07,502 INFO [finetune.py:976] (4/7) Epoch 25, batch 2400, loss[loss=0.1414, simple_loss=0.2103, pruned_loss=0.03629, over 4760.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2414, pruned_loss=0.04932, over 955184.91 frames. ], batch size: 28, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:20:07,632 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139865.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:20:10,613 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.173e+01 1.419e+02 1.768e+02 2.081e+02 3.267e+02, threshold=3.536e+02, percent-clipped=0.0 +2023-03-27 06:20:10,748 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139867.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:20:14,386 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3829, 1.5179, 1.9888, 1.7180, 1.6440, 3.6300, 1.4938, 1.7091], + device='cuda:4'), covar=tensor([0.1092, 0.1767, 0.1083, 0.0969, 0.1588, 0.0213, 0.1440, 0.1738], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 06:20:35,970 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139894.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:20:47,285 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=139911.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:20:50,040 INFO [finetune.py:976] (4/7) Epoch 25, batch 2450, loss[loss=0.1498, simple_loss=0.2164, pruned_loss=0.0416, over 4768.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2388, pruned_loss=0.0483, over 955447.26 frames. ], batch size: 26, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:20:57,354 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139927.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:21:01,386 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139932.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 06:21:05,444 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=139938.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:21:07,845 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.4783, 1.5496, 1.6092, 0.9219, 1.7265, 1.9162, 1.9399, 1.4481], + device='cuda:4'), covar=tensor([0.1186, 0.0801, 0.0703, 0.0676, 0.0576, 0.0776, 0.0380, 0.0832], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0149, 0.0128, 0.0123, 0.0131, 0.0129, 0.0142, 0.0148], + device='cuda:4'), out_proj_covar=tensor([8.9435e-05, 1.0679e-04, 9.1369e-05, 8.6230e-05, 9.1683e-05, 9.2065e-05, + 1.0099e-04, 1.0589e-04], device='cuda:4') +2023-03-27 06:21:08,389 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139942.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:21:10,187 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=139945.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:21:22,676 INFO [finetune.py:976] (4/7) Epoch 25, batch 2500, loss[loss=0.1856, simple_loss=0.2698, pruned_loss=0.05074, over 4861.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2415, pruned_loss=0.0503, over 953052.64 frames. ], batch size: 44, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:21:24,359 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.523e+02 1.884e+02 2.422e+02 3.755e+02, threshold=3.768e+02, percent-clipped=3.0 +2023-03-27 06:21:32,355 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139980.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:21:42,133 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=139993.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:21:57,549 INFO [finetune.py:976] (4/7) Epoch 25, batch 2550, loss[loss=0.1766, simple_loss=0.2573, pruned_loss=0.04792, over 4823.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2437, pruned_loss=0.05048, over 954347.38 frames. ], batch size: 39, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:21:59,092 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.83 vs. limit=5.0 +2023-03-27 06:22:15,808 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5278, 1.4449, 1.9423, 2.9237, 1.9513, 2.2245, 0.9141, 2.4790], + device='cuda:4'), covar=tensor([0.1744, 0.1398, 0.1209, 0.0646, 0.0892, 0.1330, 0.1767, 0.0553], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0133, 0.0163, 0.0101, 0.0135, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 06:22:36,080 INFO [finetune.py:976] (4/7) Epoch 25, batch 2600, loss[loss=0.1895, simple_loss=0.2783, pruned_loss=0.0504, over 4821.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2456, pruned_loss=0.05061, over 954990.10 frames. ], batch size: 38, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:22:42,055 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.019e+02 1.547e+02 1.969e+02 2.320e+02 4.703e+02, threshold=3.938e+02, percent-clipped=1.0 +2023-03-27 06:23:14,149 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140103.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:23:19,337 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140110.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:23:22,262 INFO [finetune.py:976] (4/7) Epoch 25, batch 2650, loss[loss=0.174, simple_loss=0.2488, pruned_loss=0.04961, over 4885.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2469, pruned_loss=0.05085, over 955001.42 frames. ], batch size: 43, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:23:28,792 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140125.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:23:40,094 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.60 vs. limit=5.0 +2023-03-27 06:23:41,812 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140145.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:23:51,791 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140160.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:23:53,482 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140162.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:23:53,562 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1959, 1.8761, 2.2074, 2.2596, 1.9131, 1.9421, 2.1666, 2.1266], + device='cuda:4'), covar=tensor([0.4212, 0.4229, 0.3268, 0.3755, 0.5166, 0.4060, 0.4796, 0.2973], + device='cuda:4'), in_proj_covar=tensor([0.0262, 0.0245, 0.0265, 0.0290, 0.0290, 0.0267, 0.0296, 0.0247], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:23:55,123 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140164.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:23:55,632 INFO [finetune.py:976] (4/7) Epoch 25, batch 2700, loss[loss=0.1735, simple_loss=0.2463, pruned_loss=0.05037, over 4842.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2453, pruned_loss=0.05007, over 955166.06 frames. ], batch size: 47, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:23:56,851 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.476e+02 1.708e+02 2.136e+02 4.297e+02, threshold=3.417e+02, percent-clipped=1.0 +2023-03-27 06:23:59,444 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140171.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:24:01,342 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 06:24:22,849 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140206.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:24:28,675 INFO [finetune.py:976] (4/7) Epoch 25, batch 2750, loss[loss=0.1348, simple_loss=0.2029, pruned_loss=0.03336, over 4760.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2422, pruned_loss=0.04927, over 955801.30 frames. ], batch size: 23, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:24:36,545 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140227.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:24:43,711 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140238.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:24:50,133 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1397, 2.1512, 1.6821, 1.9934, 1.9064, 1.8836, 1.9590, 2.7042], + device='cuda:4'), covar=tensor([0.3574, 0.3497, 0.3306, 0.3565, 0.3878, 0.2443, 0.3575, 0.1502], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0263, 0.0235, 0.0276, 0.0259, 0.0229, 0.0257, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:24:58,264 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140259.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:25:01,789 INFO [finetune.py:976] (4/7) Epoch 25, batch 2800, loss[loss=0.1795, simple_loss=0.2521, pruned_loss=0.05347, over 4692.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2398, pruned_loss=0.04862, over 954489.84 frames. ], batch size: 23, lr: 3.01e-03, grad_scale: 16.0 +2023-03-27 06:25:02,942 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.021e+02 1.479e+02 1.751e+02 2.221e+02 3.486e+02, threshold=3.502e+02, percent-clipped=1.0 +2023-03-27 06:25:10,749 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140275.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:25:22,268 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140286.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:25:53,836 INFO [finetune.py:976] (4/7) Epoch 25, batch 2850, loss[loss=0.1523, simple_loss=0.2324, pruned_loss=0.03613, over 4817.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2384, pruned_loss=0.0485, over 953626.93 frames. ], batch size: 38, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:25:55,178 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6875, 1.6119, 1.4207, 1.8125, 2.0220, 1.8002, 1.3282, 1.4268], + device='cuda:4'), covar=tensor([0.2235, 0.2029, 0.1956, 0.1581, 0.1549, 0.1162, 0.2528, 0.1843], + device='cuda:4'), in_proj_covar=tensor([0.0245, 0.0211, 0.0214, 0.0198, 0.0245, 0.0190, 0.0217, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:25:56,968 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140320.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:26:03,410 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1643, 1.4015, 1.5939, 1.4551, 1.5766, 2.9983, 1.3326, 1.5264], + device='cuda:4'), covar=tensor([0.1081, 0.1755, 0.1086, 0.1010, 0.1619, 0.0278, 0.1545, 0.1908], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0081, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 06:26:11,102 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3957, 1.4385, 1.1964, 1.4617, 1.7019, 1.5731, 1.4083, 1.2783], + device='cuda:4'), covar=tensor([0.0356, 0.0283, 0.0593, 0.0261, 0.0203, 0.0493, 0.0350, 0.0413], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0107, 0.0145, 0.0111, 0.0100, 0.0114, 0.0103, 0.0112], + device='cuda:4'), out_proj_covar=tensor([7.7949e-05, 8.1557e-05, 1.1344e-04, 8.4996e-05, 7.7683e-05, 8.4320e-05, + 7.6142e-05, 8.5273e-05], device='cuda:4') +2023-03-27 06:26:27,557 INFO [finetune.py:976] (4/7) Epoch 25, batch 2900, loss[loss=0.1837, simple_loss=0.2661, pruned_loss=0.05059, over 4836.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.242, pruned_loss=0.04954, over 955182.42 frames. ], batch size: 47, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:26:28,759 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.105e+02 1.583e+02 1.866e+02 2.190e+02 4.311e+02, threshold=3.732e+02, percent-clipped=1.0 +2023-03-27 06:26:41,970 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140387.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:27:01,453 INFO [finetune.py:976] (4/7) Epoch 25, batch 2950, loss[loss=0.1908, simple_loss=0.271, pruned_loss=0.05525, over 4733.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2462, pruned_loss=0.05147, over 953564.47 frames. ], batch size: 54, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:27:06,966 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8638, 1.7954, 1.7550, 1.8511, 1.4359, 3.9079, 1.5883, 2.0514], + device='cuda:4'), covar=tensor([0.3241, 0.2376, 0.1980, 0.2262, 0.1592, 0.0171, 0.2378, 0.1132], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0116, 0.0120, 0.0123, 0.0113, 0.0096, 0.0094, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 06:27:07,564 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140425.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:27:21,223 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140445.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:27:23,048 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140448.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:27:30,140 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140459.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:27:30,780 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140460.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:27:32,483 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140462.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:27:34,686 INFO [finetune.py:976] (4/7) Epoch 25, batch 3000, loss[loss=0.1946, simple_loss=0.2615, pruned_loss=0.06389, over 4815.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2472, pruned_loss=0.05162, over 953163.88 frames. ], batch size: 38, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:27:34,686 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 06:27:48,789 INFO [finetune.py:1010] (4/7) Epoch 25, validation: loss=0.1571, simple_loss=0.2254, pruned_loss=0.04443, over 2265189.00 frames. +2023-03-27 06:27:48,790 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6557MB +2023-03-27 06:27:49,984 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140466.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:27:50,499 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.568e+02 1.888e+02 2.214e+02 4.503e+02, threshold=3.776e+02, percent-clipped=3.0 +2023-03-27 06:27:59,535 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140473.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:28:17,963 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140489.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:28:20,346 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140493.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:28:29,176 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140506.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:28:29,202 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6576, 1.4527, 1.0121, 0.3256, 1.1865, 1.4911, 1.3794, 1.3626], + device='cuda:4'), covar=tensor([0.0866, 0.0803, 0.1401, 0.1796, 0.1335, 0.1992, 0.2239, 0.0878], + device='cuda:4'), in_proj_covar=tensor([0.0171, 0.0192, 0.0201, 0.0181, 0.0210, 0.0211, 0.0224, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:28:30,347 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140508.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:28:31,008 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140509.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:28:31,588 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140510.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:28:31,720 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-27 06:28:34,548 INFO [finetune.py:976] (4/7) Epoch 25, batch 3050, loss[loss=0.1619, simple_loss=0.2466, pruned_loss=0.03864, over 4906.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2482, pruned_loss=0.0515, over 953934.75 frames. ], batch size: 37, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:28:58,618 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140550.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:29:00,943 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140554.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:29:08,053 INFO [finetune.py:976] (4/7) Epoch 25, batch 3100, loss[loss=0.1386, simple_loss=0.2101, pruned_loss=0.03351, over 4846.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2459, pruned_loss=0.05062, over 955704.61 frames. ], batch size: 47, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:29:08,212 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8779, 1.4033, 2.0053, 1.9588, 1.7425, 1.7016, 1.9025, 1.8894], + device='cuda:4'), covar=tensor([0.4058, 0.3929, 0.3143, 0.3611, 0.4581, 0.3854, 0.4310, 0.2881], + device='cuda:4'), in_proj_covar=tensor([0.0264, 0.0247, 0.0267, 0.0292, 0.0292, 0.0269, 0.0298, 0.0249], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:29:09,242 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.313e+01 1.495e+02 1.767e+02 2.180e+02 4.499e+02, threshold=3.535e+02, percent-clipped=1.0 +2023-03-27 06:29:12,210 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140570.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:29:13,547 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-27 06:29:24,922 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.79 vs. limit=5.0 +2023-03-27 06:29:42,051 INFO [finetune.py:976] (4/7) Epoch 25, batch 3150, loss[loss=0.1383, simple_loss=0.2184, pruned_loss=0.02912, over 4774.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2434, pruned_loss=0.04993, over 956511.21 frames. ], batch size: 28, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:29:42,119 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140615.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:29:46,967 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4159, 2.3148, 1.8933, 2.4641, 2.3556, 2.0899, 2.6714, 2.4818], + device='cuda:4'), covar=tensor([0.1302, 0.2126, 0.2894, 0.2343, 0.2454, 0.1704, 0.2831, 0.1600], + device='cuda:4'), in_proj_covar=tensor([0.0188, 0.0190, 0.0235, 0.0253, 0.0248, 0.0205, 0.0214, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:30:15,049 INFO [finetune.py:976] (4/7) Epoch 25, batch 3200, loss[loss=0.1903, simple_loss=0.2555, pruned_loss=0.06255, over 4865.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2398, pruned_loss=0.0487, over 957057.92 frames. ], batch size: 31, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:30:16,220 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.929e+01 1.486e+02 1.750e+02 2.144e+02 4.466e+02, threshold=3.500e+02, percent-clipped=2.0 +2023-03-27 06:30:49,835 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-27 06:31:06,566 INFO [finetune.py:976] (4/7) Epoch 25, batch 3250, loss[loss=0.1842, simple_loss=0.2559, pruned_loss=0.05625, over 4917.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2419, pruned_loss=0.04971, over 957997.85 frames. ], batch size: 37, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:31:11,473 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140723.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:31:24,436 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-03-27 06:31:25,174 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140743.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:31:35,353 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140759.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:31:39,361 INFO [finetune.py:976] (4/7) Epoch 25, batch 3300, loss[loss=0.1502, simple_loss=0.2379, pruned_loss=0.03125, over 4796.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2432, pruned_loss=0.04963, over 958445.24 frames. ], batch size: 29, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:31:40,552 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140766.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:31:41,066 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.629e+02 1.945e+02 2.397e+02 4.021e+02, threshold=3.889e+02, percent-clipped=5.0 +2023-03-27 06:31:43,080 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.08 vs. limit=5.0 +2023-03-27 06:31:53,076 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140784.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:31:56,526 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3219, 1.3986, 1.7344, 1.6601, 1.5691, 3.2780, 1.3642, 1.5717], + device='cuda:4'), covar=tensor([0.1005, 0.1770, 0.1059, 0.0910, 0.1564, 0.0235, 0.1492, 0.1718], + device='cuda:4'), in_proj_covar=tensor([0.0074, 0.0082, 0.0073, 0.0076, 0.0091, 0.0081, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 06:32:08,087 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140807.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:32:09,343 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140809.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:32:12,392 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140814.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:32:12,930 INFO [finetune.py:976] (4/7) Epoch 25, batch 3350, loss[loss=0.1663, simple_loss=0.2527, pruned_loss=0.0399, over 4818.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2456, pruned_loss=0.05044, over 957608.54 frames. ], batch size: 45, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:32:20,668 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9235, 4.6878, 4.4766, 2.5927, 4.6739, 3.5415, 0.8984, 3.3594], + device='cuda:4'), covar=tensor([0.2344, 0.1341, 0.1291, 0.2696, 0.0702, 0.0804, 0.4345, 0.1115], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0178, 0.0161, 0.0130, 0.0160, 0.0123, 0.0147, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 06:32:33,955 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140845.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:32:44,308 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7282, 1.6567, 1.3166, 1.3643, 1.7962, 1.5124, 2.0401, 1.7984], + device='cuda:4'), covar=tensor([0.1688, 0.2042, 0.3540, 0.2656, 0.2965, 0.1853, 0.2370, 0.1905], + device='cuda:4'), in_proj_covar=tensor([0.0187, 0.0188, 0.0233, 0.0251, 0.0246, 0.0204, 0.0212, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:32:46,606 INFO [finetune.py:976] (4/7) Epoch 25, batch 3400, loss[loss=0.206, simple_loss=0.2775, pruned_loss=0.06723, over 4819.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2478, pruned_loss=0.05131, over 958329.34 frames. ], batch size: 39, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:32:46,675 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=140865.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:32:47,793 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.564e+02 1.878e+02 2.236e+02 3.278e+02, threshold=3.756e+02, percent-clipped=0.0 +2023-03-27 06:32:49,700 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140870.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:33:19,370 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.5993, 3.9763, 4.1355, 4.4283, 4.3836, 4.0090, 4.6920, 1.5783], + device='cuda:4'), covar=tensor([0.0847, 0.0823, 0.0942, 0.1002, 0.1256, 0.1849, 0.0679, 0.5900], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0247, 0.0282, 0.0296, 0.0337, 0.0286, 0.0307, 0.0301], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:33:38,074 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9885, 1.8195, 2.1095, 1.2148, 1.9287, 1.9913, 1.9569, 1.6214], + device='cuda:4'), covar=tensor([0.0537, 0.0686, 0.0592, 0.0896, 0.0769, 0.0657, 0.0624, 0.1179], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0136, 0.0140, 0.0119, 0.0126, 0.0137, 0.0138, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:33:39,170 INFO [finetune.py:976] (4/7) Epoch 25, batch 3450, loss[loss=0.1316, simple_loss=0.2128, pruned_loss=0.02521, over 4761.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2472, pruned_loss=0.05049, over 957379.32 frames. ], batch size: 27, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:33:39,276 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=140915.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:33:40,471 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140917.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:34:01,689 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=140947.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:34:11,813 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=140963.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:34:12,944 INFO [finetune.py:976] (4/7) Epoch 25, batch 3500, loss[loss=0.1872, simple_loss=0.2531, pruned_loss=0.06064, over 4930.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2436, pruned_loss=0.04958, over 957892.71 frames. ], batch size: 43, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:34:14,174 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.063e+02 1.468e+02 1.748e+02 2.204e+02 3.629e+02, threshold=3.496e+02, percent-clipped=0.0 +2023-03-27 06:34:20,897 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=140978.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:34:41,918 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141008.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:34:46,050 INFO [finetune.py:976] (4/7) Epoch 25, batch 3550, loss[loss=0.1206, simple_loss=0.2023, pruned_loss=0.01941, over 4869.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2414, pruned_loss=0.04914, over 955843.91 frames. ], batch size: 31, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:34:50,019 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-27 06:35:04,137 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141043.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:35:19,348 INFO [finetune.py:976] (4/7) Epoch 25, batch 3600, loss[loss=0.1767, simple_loss=0.2368, pruned_loss=0.05828, over 4902.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.239, pruned_loss=0.04859, over 952128.04 frames. ], batch size: 36, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:35:20,525 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.456e+02 1.796e+02 2.356e+02 3.995e+02, threshold=3.592e+02, percent-clipped=1.0 +2023-03-27 06:35:28,399 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141079.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:35:36,624 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141091.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:35:37,333 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-03-27 06:35:46,615 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0877, 2.2212, 1.7752, 2.3411, 2.1475, 2.0065, 2.1275, 2.8736], + device='cuda:4'), covar=tensor([0.3952, 0.4548, 0.3373, 0.3871, 0.4166, 0.2430, 0.4062, 0.1635], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0263, 0.0235, 0.0276, 0.0259, 0.0229, 0.0255, 0.0237], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:36:00,473 INFO [finetune.py:976] (4/7) Epoch 25, batch 3650, loss[loss=0.178, simple_loss=0.2544, pruned_loss=0.05082, over 4810.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2421, pruned_loss=0.04984, over 951815.37 frames. ], batch size: 45, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:36:30,936 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141143.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:36:32,096 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141145.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:36:42,944 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-03-27 06:36:44,632 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-27 06:36:46,082 INFO [finetune.py:976] (4/7) Epoch 25, batch 3700, loss[loss=0.2339, simple_loss=0.3024, pruned_loss=0.08272, over 4852.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2455, pruned_loss=0.05106, over 952581.82 frames. ], batch size: 44, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:36:46,156 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141165.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:36:46,188 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141165.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:36:47,287 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.708e+02 2.029e+02 2.382e+02 3.628e+02, threshold=4.058e+02, percent-clipped=1.0 +2023-03-27 06:37:03,882 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141193.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:37:11,228 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141204.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:37:15,095 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.0882, 3.5672, 3.7732, 3.8692, 3.9028, 3.6762, 4.1549, 1.8408], + device='cuda:4'), covar=tensor([0.0800, 0.0873, 0.0823, 0.0997, 0.1038, 0.1270, 0.0700, 0.4901], + device='cuda:4'), in_proj_covar=tensor([0.0351, 0.0248, 0.0284, 0.0297, 0.0339, 0.0288, 0.0308, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:37:18,589 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141213.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:37:19,743 INFO [finetune.py:976] (4/7) Epoch 25, batch 3750, loss[loss=0.195, simple_loss=0.2608, pruned_loss=0.06456, over 4741.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2462, pruned_loss=0.05079, over 952477.55 frames. ], batch size: 54, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:37:45,896 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2065, 3.6378, 3.8129, 4.0273, 3.9813, 3.7152, 4.2757, 1.3652], + device='cuda:4'), covar=tensor([0.0787, 0.0873, 0.0858, 0.1002, 0.1219, 0.1531, 0.0723, 0.5664], + device='cuda:4'), in_proj_covar=tensor([0.0351, 0.0248, 0.0283, 0.0296, 0.0339, 0.0288, 0.0308, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:37:52,659 INFO [finetune.py:976] (4/7) Epoch 25, batch 3800, loss[loss=0.1438, simple_loss=0.2244, pruned_loss=0.0316, over 4761.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2476, pruned_loss=0.05131, over 951265.49 frames. ], batch size: 28, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:37:54,344 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.018e+02 1.508e+02 1.827e+02 2.217e+02 6.513e+02, threshold=3.654e+02, percent-clipped=2.0 +2023-03-27 06:37:58,030 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141273.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:38:19,447 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141303.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:38:32,076 INFO [finetune.py:976] (4/7) Epoch 25, batch 3850, loss[loss=0.1969, simple_loss=0.2659, pruned_loss=0.06394, over 4871.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2458, pruned_loss=0.05089, over 950173.80 frames. ], batch size: 34, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:38:49,944 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141329.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:38:50,633 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.25 vs. limit=5.0 +2023-03-27 06:39:16,920 INFO [finetune.py:976] (4/7) Epoch 25, batch 3900, loss[loss=0.1912, simple_loss=0.247, pruned_loss=0.06768, over 4904.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2433, pruned_loss=0.05052, over 950746.75 frames. ], batch size: 43, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:39:18,108 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.088e+02 1.504e+02 1.773e+02 2.110e+02 6.012e+02, threshold=3.546e+02, percent-clipped=1.0 +2023-03-27 06:39:25,836 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6406, 1.6703, 1.3367, 1.6232, 1.9890, 1.9344, 1.6335, 1.4858], + device='cuda:4'), covar=tensor([0.0345, 0.0364, 0.0664, 0.0353, 0.0214, 0.0414, 0.0350, 0.0410], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0107, 0.0146, 0.0111, 0.0101, 0.0115, 0.0103, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.8092e-05, 8.1913e-05, 1.1369e-04, 8.5441e-05, 7.7918e-05, 8.4623e-05, + 7.6618e-05, 8.5683e-05], device='cuda:4') +2023-03-27 06:39:26,400 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141379.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:39:33,598 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141390.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:39:49,670 INFO [finetune.py:976] (4/7) Epoch 25, batch 3950, loss[loss=0.1507, simple_loss=0.2237, pruned_loss=0.03882, over 4829.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2396, pruned_loss=0.04895, over 953847.35 frames. ], batch size: 38, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:39:51,521 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141417.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 06:39:58,914 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141427.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:40:07,859 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141441.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:40:23,341 INFO [finetune.py:976] (4/7) Epoch 25, batch 4000, loss[loss=0.1234, simple_loss=0.2002, pruned_loss=0.02323, over 4780.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2373, pruned_loss=0.04777, over 954127.96 frames. ], batch size: 28, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:40:23,423 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141465.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:40:24,526 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.071e+02 1.514e+02 1.750e+02 2.113e+02 3.817e+02, threshold=3.500e+02, percent-clipped=1.0 +2023-03-27 06:40:33,183 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141478.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 06:40:46,340 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141499.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:40:48,210 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141502.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:40:55,310 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141513.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:40:56,984 INFO [finetune.py:976] (4/7) Epoch 25, batch 4050, loss[loss=0.1792, simple_loss=0.2519, pruned_loss=0.05321, over 4826.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2427, pruned_loss=0.04961, over 954902.49 frames. ], batch size: 30, lr: 3.00e-03, grad_scale: 16.0 +2023-03-27 06:41:28,244 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0633, 1.4064, 2.0765, 2.0403, 1.8617, 1.8279, 1.9493, 1.9540], + device='cuda:4'), covar=tensor([0.3743, 0.3632, 0.3071, 0.3394, 0.4405, 0.3407, 0.3899, 0.2887], + device='cuda:4'), in_proj_covar=tensor([0.0264, 0.0247, 0.0267, 0.0293, 0.0293, 0.0269, 0.0299, 0.0251], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:41:49,013 INFO [finetune.py:976] (4/7) Epoch 25, batch 4100, loss[loss=0.1801, simple_loss=0.2619, pruned_loss=0.04918, over 4896.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2449, pruned_loss=0.05023, over 954036.77 frames. ], batch size: 37, lr: 3.00e-03, grad_scale: 32.0 +2023-03-27 06:41:50,188 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.144e+02 1.605e+02 1.887e+02 2.173e+02 5.231e+02, threshold=3.774e+02, percent-clipped=3.0 +2023-03-27 06:41:54,376 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141573.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:42:14,755 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141603.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:42:22,402 INFO [finetune.py:976] (4/7) Epoch 25, batch 4150, loss[loss=0.1736, simple_loss=0.2406, pruned_loss=0.05332, over 4865.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2463, pruned_loss=0.05101, over 954182.58 frames. ], batch size: 34, lr: 3.00e-03, grad_scale: 32.0 +2023-03-27 06:42:26,083 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141621.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:42:46,643 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141651.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:42:52,805 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-27 06:42:55,969 INFO [finetune.py:976] (4/7) Epoch 25, batch 4200, loss[loss=0.1443, simple_loss=0.2232, pruned_loss=0.03273, over 4875.00 frames. ], tot_loss[loss=0.1755, simple_loss=0.2481, pruned_loss=0.05143, over 954860.45 frames. ], batch size: 32, lr: 3.00e-03, grad_scale: 32.0 +2023-03-27 06:42:57,196 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.937e+01 1.553e+02 1.832e+02 2.223e+02 5.119e+02, threshold=3.664e+02, percent-clipped=2.0 +2023-03-27 06:43:09,616 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141685.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:43:29,324 INFO [finetune.py:976] (4/7) Epoch 25, batch 4250, loss[loss=0.1734, simple_loss=0.2344, pruned_loss=0.05621, over 4817.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2455, pruned_loss=0.05013, over 954425.22 frames. ], batch size: 30, lr: 3.00e-03, grad_scale: 32.0 +2023-03-27 06:44:21,246 INFO [finetune.py:976] (4/7) Epoch 25, batch 4300, loss[loss=0.1302, simple_loss=0.2067, pruned_loss=0.02689, over 4774.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2426, pruned_loss=0.04975, over 954960.77 frames. ], batch size: 28, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:44:22,429 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.814e+01 1.357e+02 1.630e+02 2.025e+02 3.929e+02, threshold=3.260e+02, percent-clipped=1.0 +2023-03-27 06:44:26,620 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141773.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 06:44:43,605 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=141797.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:44:44,867 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141799.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:44:55,201 INFO [finetune.py:976] (4/7) Epoch 25, batch 4350, loss[loss=0.1429, simple_loss=0.2256, pruned_loss=0.03016, over 4773.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.24, pruned_loss=0.04894, over 955029.13 frames. ], batch size: 28, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:45:08,360 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2690, 3.6962, 3.9257, 4.1456, 4.0683, 3.8322, 4.3538, 1.3851], + device='cuda:4'), covar=tensor([0.0846, 0.0950, 0.0924, 0.1000, 0.1255, 0.1489, 0.0798, 0.5686], + device='cuda:4'), in_proj_covar=tensor([0.0351, 0.0249, 0.0283, 0.0297, 0.0339, 0.0288, 0.0308, 0.0301], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:45:15,258 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5680, 1.5052, 1.3599, 1.6819, 1.6385, 1.6338, 1.1519, 1.3674], + device='cuda:4'), covar=tensor([0.2185, 0.1927, 0.1826, 0.1523, 0.1413, 0.1137, 0.2323, 0.1871], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0211, 0.0215, 0.0198, 0.0245, 0.0191, 0.0217, 0.0206], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:45:15,538 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.31 vs. limit=5.0 +2023-03-27 06:45:17,474 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=141847.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:45:28,846 INFO [finetune.py:976] (4/7) Epoch 25, batch 4400, loss[loss=0.1456, simple_loss=0.2341, pruned_loss=0.02857, over 4930.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2415, pruned_loss=0.05005, over 956035.37 frames. ], batch size: 33, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:45:30,033 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.038e+02 1.595e+02 1.814e+02 2.202e+02 4.275e+02, threshold=3.628e+02, percent-clipped=6.0 +2023-03-27 06:45:36,681 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0921, 1.3614, 1.2137, 1.2558, 1.5758, 2.4970, 1.3769, 1.4974], + device='cuda:4'), covar=tensor([0.1000, 0.1861, 0.1062, 0.0936, 0.1540, 0.0382, 0.1493, 0.1729], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0092, 0.0081, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 06:45:43,854 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6709, 1.6283, 1.5938, 1.6692, 1.2762, 3.7680, 1.4412, 1.7398], + device='cuda:4'), covar=tensor([0.3212, 0.2523, 0.2106, 0.2377, 0.1668, 0.0180, 0.2563, 0.1339], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0116, 0.0120, 0.0123, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 06:46:01,881 INFO [finetune.py:976] (4/7) Epoch 25, batch 4450, loss[loss=0.1712, simple_loss=0.2278, pruned_loss=0.05736, over 4756.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2453, pruned_loss=0.05127, over 957609.57 frames. ], batch size: 23, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:46:02,593 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141916.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:46:10,448 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=141928.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:46:23,193 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-03-27 06:46:43,702 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-27 06:46:46,892 INFO [finetune.py:976] (4/7) Epoch 25, batch 4500, loss[loss=0.1817, simple_loss=0.2539, pruned_loss=0.05474, over 4867.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2464, pruned_loss=0.05117, over 956900.22 frames. ], batch size: 34, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:46:52,641 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.869e+01 1.594e+02 1.826e+02 2.236e+02 4.959e+02, threshold=3.653e+02, percent-clipped=2.0 +2023-03-27 06:47:02,941 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141977.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:47:08,188 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=141985.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:47:10,670 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=141989.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:47:27,105 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4826, 2.7703, 2.5757, 1.7920, 2.4117, 2.6377, 2.7789, 2.2399], + device='cuda:4'), covar=tensor([0.0584, 0.0512, 0.0666, 0.0886, 0.0880, 0.0671, 0.0557, 0.1054], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0136, 0.0140, 0.0119, 0.0126, 0.0137, 0.0138, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:47:30,043 INFO [finetune.py:976] (4/7) Epoch 25, batch 4550, loss[loss=0.1625, simple_loss=0.2473, pruned_loss=0.03886, over 4866.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2469, pruned_loss=0.05113, over 956784.47 frames. ], batch size: 34, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:47:38,437 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1311, 1.9008, 2.1123, 1.3112, 1.9249, 2.0122, 2.1345, 1.5937], + device='cuda:4'), covar=tensor([0.0551, 0.0650, 0.0673, 0.0911, 0.0836, 0.0723, 0.0584, 0.1174], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0136, 0.0140, 0.0119, 0.0126, 0.0137, 0.0139, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:47:41,389 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142033.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:48:03,344 INFO [finetune.py:976] (4/7) Epoch 25, batch 4600, loss[loss=0.1541, simple_loss=0.2311, pruned_loss=0.03852, over 4917.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2458, pruned_loss=0.05079, over 954063.61 frames. ], batch size: 38, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:48:04,587 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 7.716e+01 1.569e+02 1.799e+02 2.271e+02 3.318e+02, threshold=3.598e+02, percent-clipped=0.0 +2023-03-27 06:48:08,723 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=142073.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 06:48:15,639 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-27 06:48:23,761 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=142097.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:48:36,588 INFO [finetune.py:976] (4/7) Epoch 25, batch 4650, loss[loss=0.1376, simple_loss=0.2149, pruned_loss=0.03018, over 4901.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2431, pruned_loss=0.05068, over 953941.80 frames. ], batch size: 43, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:48:40,333 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142121.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 06:48:57,212 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142145.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:49:14,800 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-03-27 06:49:20,007 INFO [finetune.py:976] (4/7) Epoch 25, batch 4700, loss[loss=0.1564, simple_loss=0.2277, pruned_loss=0.04257, over 4932.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2404, pruned_loss=0.04984, over 955753.12 frames. ], batch size: 33, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:49:21,185 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.163e+01 1.384e+02 1.765e+02 2.088e+02 3.764e+02, threshold=3.531e+02, percent-clipped=1.0 +2023-03-27 06:49:47,854 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5764, 1.4860, 1.3356, 1.6612, 1.6184, 1.6140, 0.9392, 1.3520], + device='cuda:4'), covar=tensor([0.2249, 0.2015, 0.1939, 0.1657, 0.1647, 0.1306, 0.2671, 0.1911], + device='cuda:4'), in_proj_covar=tensor([0.0247, 0.0213, 0.0216, 0.0199, 0.0247, 0.0192, 0.0219, 0.0207], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:50:00,898 INFO [finetune.py:976] (4/7) Epoch 25, batch 4750, loss[loss=0.1535, simple_loss=0.2281, pruned_loss=0.0394, over 4843.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2396, pruned_loss=0.04993, over 954795.67 frames. ], batch size: 47, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:50:11,681 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.4430, 3.8971, 4.0610, 4.2744, 4.2000, 3.9415, 4.5550, 1.4345], + device='cuda:4'), covar=tensor([0.0726, 0.0894, 0.0880, 0.1015, 0.1199, 0.1593, 0.0723, 0.5908], + device='cuda:4'), in_proj_covar=tensor([0.0353, 0.0250, 0.0284, 0.0298, 0.0339, 0.0290, 0.0307, 0.0304], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:50:18,427 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.1815, 3.2938, 3.1943, 2.2424, 2.9316, 3.4277, 3.4442, 2.6313], + device='cuda:4'), covar=tensor([0.0500, 0.0425, 0.0579, 0.0776, 0.0869, 0.0528, 0.0517, 0.0891], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0136, 0.0141, 0.0119, 0.0127, 0.0138, 0.0139, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:50:24,427 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6571, 1.5937, 1.5448, 1.6675, 1.1761, 3.5464, 1.3213, 1.7282], + device='cuda:4'), covar=tensor([0.3319, 0.2550, 0.2098, 0.2300, 0.1735, 0.0229, 0.2699, 0.1313], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0117, 0.0121, 0.0124, 0.0113, 0.0096, 0.0095, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 06:50:29,895 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3071, 2.3084, 1.8025, 2.5198, 2.2254, 1.8943, 2.7299, 2.3706], + device='cuda:4'), covar=tensor([0.1336, 0.2340, 0.3039, 0.2674, 0.2763, 0.1793, 0.3292, 0.1764], + device='cuda:4'), in_proj_covar=tensor([0.0189, 0.0191, 0.0237, 0.0256, 0.0251, 0.0207, 0.0216, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:50:34,338 INFO [finetune.py:976] (4/7) Epoch 25, batch 4800, loss[loss=0.1341, simple_loss=0.2116, pruned_loss=0.02825, over 4680.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2415, pruned_loss=0.05005, over 955303.84 frames. ], batch size: 23, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:50:35,544 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.947e+01 1.535e+02 1.762e+02 2.238e+02 3.446e+02, threshold=3.524e+02, percent-clipped=1.0 +2023-03-27 06:50:39,632 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=142272.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:50:47,480 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=142284.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:51:07,516 INFO [finetune.py:976] (4/7) Epoch 25, batch 4850, loss[loss=0.1756, simple_loss=0.2422, pruned_loss=0.05456, over 4807.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2459, pruned_loss=0.05116, over 956130.88 frames. ], batch size: 25, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:51:39,151 INFO [finetune.py:976] (4/7) Epoch 25, batch 4900, loss[loss=0.2041, simple_loss=0.2753, pruned_loss=0.06645, over 4211.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2474, pruned_loss=0.05194, over 951038.27 frames. ], batch size: 65, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:51:40,866 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.096e+02 1.551e+02 1.812e+02 2.135e+02 6.918e+02, threshold=3.624e+02, percent-clipped=2.0 +2023-03-27 06:52:31,143 INFO [finetune.py:976] (4/7) Epoch 25, batch 4950, loss[loss=0.1949, simple_loss=0.2479, pruned_loss=0.07092, over 4285.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2478, pruned_loss=0.05199, over 950217.00 frames. ], batch size: 66, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:53:03,310 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2535, 2.1622, 1.7748, 2.4002, 2.1975, 1.9515, 2.5886, 2.2315], + device='cuda:4'), covar=tensor([0.1285, 0.2157, 0.2973, 0.2301, 0.2521, 0.1708, 0.2777, 0.1919], + device='cuda:4'), in_proj_covar=tensor([0.0189, 0.0191, 0.0237, 0.0256, 0.0251, 0.0207, 0.0216, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:53:03,772 INFO [finetune.py:976] (4/7) Epoch 25, batch 5000, loss[loss=0.1695, simple_loss=0.2387, pruned_loss=0.0501, over 4904.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2458, pruned_loss=0.05103, over 950822.22 frames. ], batch size: 36, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:53:04,978 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.190e+01 1.432e+02 1.813e+02 2.155e+02 3.992e+02, threshold=3.625e+02, percent-clipped=1.0 +2023-03-27 06:53:12,263 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9388, 1.8318, 2.0341, 1.1496, 1.9862, 2.0041, 1.9676, 1.6516], + device='cuda:4'), covar=tensor([0.0577, 0.0672, 0.0638, 0.0910, 0.0818, 0.0670, 0.0649, 0.1149], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0137, 0.0141, 0.0120, 0.0128, 0.0139, 0.0141, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:53:29,348 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1994, 2.1150, 1.8457, 2.0192, 2.0412, 1.9757, 2.0716, 2.7531], + device='cuda:4'), covar=tensor([0.3722, 0.4056, 0.3245, 0.3866, 0.4015, 0.2526, 0.3856, 0.1628], + device='cuda:4'), in_proj_covar=tensor([0.0290, 0.0264, 0.0236, 0.0276, 0.0259, 0.0229, 0.0256, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:53:36,415 INFO [finetune.py:976] (4/7) Epoch 25, batch 5050, loss[loss=0.1517, simple_loss=0.2198, pruned_loss=0.04176, over 4719.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2438, pruned_loss=0.05093, over 951553.30 frames. ], batch size: 23, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:53:37,687 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5263, 3.2847, 3.1610, 1.5373, 3.4044, 2.6123, 0.9429, 2.2927], + device='cuda:4'), covar=tensor([0.2493, 0.2183, 0.1677, 0.3506, 0.1321, 0.1108, 0.4080, 0.1757], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0178, 0.0161, 0.0130, 0.0161, 0.0123, 0.0148, 0.0125], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 06:53:51,079 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.68 vs. limit=5.0 +2023-03-27 06:54:09,846 INFO [finetune.py:976] (4/7) Epoch 25, batch 5100, loss[loss=0.184, simple_loss=0.2572, pruned_loss=0.05545, over 4824.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2397, pruned_loss=0.04937, over 952995.92 frames. ], batch size: 41, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:54:11,047 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.212e+01 1.519e+02 1.807e+02 2.247e+02 4.075e+02, threshold=3.613e+02, percent-clipped=2.0 +2023-03-27 06:54:14,221 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=142572.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:54:14,840 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=142573.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:54:27,719 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=142584.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:54:59,670 INFO [finetune.py:976] (4/7) Epoch 25, batch 5150, loss[loss=0.2094, simple_loss=0.2717, pruned_loss=0.07352, over 4833.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.241, pruned_loss=0.05032, over 952692.94 frames. ], batch size: 33, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:55:01,637 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-27 06:55:03,299 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142620.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:55:10,594 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=142632.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:55:12,818 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=142634.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:55:20,572 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7151, 1.6347, 1.4117, 1.5219, 1.9415, 1.9011, 1.6601, 1.4526], + device='cuda:4'), covar=tensor([0.0354, 0.0300, 0.0591, 0.0337, 0.0209, 0.0513, 0.0375, 0.0408], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0106, 0.0145, 0.0110, 0.0100, 0.0114, 0.0102, 0.0112], + device='cuda:4'), out_proj_covar=tensor([7.7679e-05, 8.0980e-05, 1.1288e-04, 8.4273e-05, 7.7511e-05, 8.4414e-05, + 7.5923e-05, 8.4892e-05], device='cuda:4') +2023-03-27 06:55:33,009 INFO [finetune.py:976] (4/7) Epoch 25, batch 5200, loss[loss=0.227, simple_loss=0.3045, pruned_loss=0.07473, over 4807.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2438, pruned_loss=0.05076, over 952919.30 frames. ], batch size: 51, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:55:34,191 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.084e+02 1.563e+02 1.762e+02 2.093e+02 3.679e+02, threshold=3.523e+02, percent-clipped=1.0 +2023-03-27 06:56:06,168 INFO [finetune.py:976] (4/7) Epoch 25, batch 5250, loss[loss=0.1854, simple_loss=0.2494, pruned_loss=0.06071, over 4863.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2446, pruned_loss=0.05071, over 952918.87 frames. ], batch size: 31, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:56:12,146 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=142724.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:56:39,098 INFO [finetune.py:976] (4/7) Epoch 25, batch 5300, loss[loss=0.2205, simple_loss=0.2706, pruned_loss=0.0852, over 4722.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2458, pruned_loss=0.05087, over 955013.91 frames. ], batch size: 23, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:56:40,275 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.031e+02 1.558e+02 1.826e+02 2.127e+02 3.045e+02, threshold=3.651e+02, percent-clipped=0.0 +2023-03-27 06:56:51,747 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=142785.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:57:19,998 INFO [finetune.py:976] (4/7) Epoch 25, batch 5350, loss[loss=0.1425, simple_loss=0.2294, pruned_loss=0.02774, over 4885.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2447, pruned_loss=0.04951, over 954246.20 frames. ], batch size: 35, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:57:41,739 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6535, 2.5143, 1.9916, 1.0151, 2.1725, 2.1182, 1.8898, 2.2035], + device='cuda:4'), covar=tensor([0.0860, 0.0699, 0.1540, 0.2009, 0.1235, 0.2178, 0.2255, 0.0951], + device='cuda:4'), in_proj_covar=tensor([0.0172, 0.0193, 0.0202, 0.0183, 0.0211, 0.0213, 0.0226, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 06:58:06,041 INFO [finetune.py:976] (4/7) Epoch 25, batch 5400, loss[loss=0.1227, simple_loss=0.1987, pruned_loss=0.02331, over 4835.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2425, pruned_loss=0.04906, over 955671.41 frames. ], batch size: 47, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:58:07,259 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.131e+02 1.487e+02 1.682e+02 2.190e+02 4.832e+02, threshold=3.364e+02, percent-clipped=1.0 +2023-03-27 06:58:38,664 INFO [finetune.py:976] (4/7) Epoch 25, batch 5450, loss[loss=0.1279, simple_loss=0.1952, pruned_loss=0.03024, over 4872.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2398, pruned_loss=0.04846, over 956083.29 frames. ], batch size: 44, lr: 2.99e-03, grad_scale: 32.0 +2023-03-27 06:58:47,651 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=142929.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 06:58:52,346 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3119, 1.4109, 1.4323, 0.7712, 1.5340, 1.6658, 1.7168, 1.3665], + device='cuda:4'), covar=tensor([0.0906, 0.0694, 0.0482, 0.0551, 0.0441, 0.0668, 0.0326, 0.0631], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0147, 0.0127, 0.0122, 0.0130, 0.0129, 0.0142, 0.0148], + device='cuda:4'), out_proj_covar=tensor([8.9069e-05, 1.0600e-04, 9.0818e-05, 8.5537e-05, 9.0888e-05, 9.1768e-05, + 1.0096e-04, 1.0574e-04], device='cuda:4') +2023-03-27 06:58:56,134 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-27 06:59:07,382 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-27 06:59:11,891 INFO [finetune.py:976] (4/7) Epoch 25, batch 5500, loss[loss=0.1567, simple_loss=0.2105, pruned_loss=0.05142, over 4037.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2377, pruned_loss=0.04809, over 955537.70 frames. ], batch size: 65, lr: 2.99e-03, grad_scale: 16.0 +2023-03-27 06:59:13,718 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.198e+01 1.372e+02 1.708e+02 2.223e+02 4.314e+02, threshold=3.415e+02, percent-clipped=3.0 +2023-03-27 06:59:46,284 INFO [finetune.py:976] (4/7) Epoch 25, batch 5550, loss[loss=0.1865, simple_loss=0.2632, pruned_loss=0.05489, over 4861.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2399, pruned_loss=0.04929, over 953256.46 frames. ], batch size: 31, lr: 2.99e-03, grad_scale: 16.0 +2023-03-27 07:00:06,557 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.0512, 4.3582, 4.6319, 4.9559, 4.7741, 4.4378, 5.1664, 1.7408], + device='cuda:4'), covar=tensor([0.0739, 0.0882, 0.0823, 0.0848, 0.1335, 0.1814, 0.0603, 0.5948], + device='cuda:4'), in_proj_covar=tensor([0.0351, 0.0248, 0.0281, 0.0296, 0.0337, 0.0288, 0.0307, 0.0301], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:00:06,841 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-27 07:00:29,926 INFO [finetune.py:976] (4/7) Epoch 25, batch 5600, loss[loss=0.1931, simple_loss=0.2613, pruned_loss=0.06248, over 4935.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2431, pruned_loss=0.04939, over 954262.83 frames. ], batch size: 33, lr: 2.99e-03, grad_scale: 16.0 +2023-03-27 07:00:30,047 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1133, 2.0565, 1.8221, 2.1333, 2.6627, 2.1444, 2.1900, 1.6239], + device='cuda:4'), covar=tensor([0.2083, 0.1733, 0.1825, 0.1542, 0.1714, 0.1216, 0.1844, 0.1805], + device='cuda:4'), in_proj_covar=tensor([0.0245, 0.0211, 0.0216, 0.0199, 0.0245, 0.0193, 0.0217, 0.0206], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:00:30,589 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3984, 0.9683, 0.7830, 1.2832, 1.8408, 0.8276, 1.1608, 1.2997], + device='cuda:4'), covar=tensor([0.1586, 0.2244, 0.1709, 0.1300, 0.2142, 0.1894, 0.1545, 0.2025], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0109, 0.0092, 0.0119, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 07:00:31,670 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.665e+01 1.700e+02 1.937e+02 2.306e+02 4.675e+02, threshold=3.875e+02, percent-clipped=1.0 +2023-03-27 07:00:38,730 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=143080.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:00:51,424 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5404, 1.4340, 1.9355, 1.7712, 1.7664, 3.6117, 1.5666, 1.6689], + device='cuda:4'), covar=tensor([0.1059, 0.1868, 0.1151, 0.0989, 0.1581, 0.0231, 0.1450, 0.1852], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0081, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 07:00:59,197 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3398, 2.8517, 2.6956, 1.3180, 2.9481, 2.2649, 0.8081, 1.9344], + device='cuda:4'), covar=tensor([0.2096, 0.2110, 0.1975, 0.3164, 0.1351, 0.1029, 0.3808, 0.1417], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0179, 0.0161, 0.0130, 0.0161, 0.0123, 0.0148, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 07:01:00,342 INFO [finetune.py:976] (4/7) Epoch 25, batch 5650, loss[loss=0.1604, simple_loss=0.2438, pruned_loss=0.03847, over 4892.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2461, pruned_loss=0.05009, over 955920.63 frames. ], batch size: 43, lr: 2.99e-03, grad_scale: 16.0 +2023-03-27 07:01:23,402 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7810, 2.5770, 2.0292, 1.1100, 2.3102, 2.2701, 2.0163, 2.3315], + device='cuda:4'), covar=tensor([0.0834, 0.0802, 0.1715, 0.2040, 0.1177, 0.2055, 0.2124, 0.0946], + device='cuda:4'), in_proj_covar=tensor([0.0172, 0.0193, 0.0201, 0.0183, 0.0211, 0.0213, 0.0225, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:01:29,842 INFO [finetune.py:976] (4/7) Epoch 25, batch 5700, loss[loss=0.1846, simple_loss=0.2419, pruned_loss=0.0636, over 4049.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2419, pruned_loss=0.04966, over 935012.00 frames. ], batch size: 18, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:01:31,571 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.213e+01 1.339e+02 1.671e+02 2.043e+02 4.216e+02, threshold=3.342e+02, percent-clipped=1.0 +2023-03-27 07:01:58,300 INFO [finetune.py:976] (4/7) Epoch 26, batch 0, loss[loss=0.1479, simple_loss=0.2277, pruned_loss=0.03402, over 4765.00 frames. ], tot_loss[loss=0.1479, simple_loss=0.2277, pruned_loss=0.03402, over 4765.00 frames. ], batch size: 26, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:01:58,300 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 07:02:01,170 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7779, 1.6498, 1.9604, 1.3156, 1.6456, 1.9498, 1.5483, 2.0883], + device='cuda:4'), covar=tensor([0.1308, 0.2322, 0.1460, 0.1867, 0.1092, 0.1523, 0.3077, 0.0957], + device='cuda:4'), in_proj_covar=tensor([0.0190, 0.0205, 0.0190, 0.0190, 0.0173, 0.0212, 0.0215, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:02:03,217 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4970, 1.3184, 1.3294, 1.4069, 1.6699, 1.6565, 1.4052, 1.2974], + device='cuda:4'), covar=tensor([0.0384, 0.0349, 0.0690, 0.0363, 0.0278, 0.0425, 0.0420, 0.0441], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0107, 0.0146, 0.0112, 0.0102, 0.0116, 0.0104, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.8500e-05, 8.1959e-05, 1.1422e-04, 8.5531e-05, 7.8734e-05, 8.5958e-05, + 7.7183e-05, 8.6065e-05], device='cuda:4') +2023-03-27 07:02:06,209 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1197, 1.9259, 1.7270, 1.6873, 1.8648, 1.8807, 1.8778, 2.5598], + device='cuda:4'), covar=tensor([0.3990, 0.4428, 0.3389, 0.3859, 0.4222, 0.2509, 0.3983, 0.1796], + device='cuda:4'), in_proj_covar=tensor([0.0291, 0.0265, 0.0237, 0.0278, 0.0260, 0.0230, 0.0257, 0.0239], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:02:14,281 INFO [finetune.py:1010] (4/7) Epoch 26, validation: loss=0.1591, simple_loss=0.2269, pruned_loss=0.04565, over 2265189.00 frames. +2023-03-27 07:02:14,281 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6557MB +2023-03-27 07:02:43,898 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=143229.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:03:00,608 INFO [finetune.py:976] (4/7) Epoch 26, batch 50, loss[loss=0.1813, simple_loss=0.2454, pruned_loss=0.0586, over 4842.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2434, pruned_loss=0.04759, over 217739.36 frames. ], batch size: 44, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:03:13,061 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1691, 1.3345, 1.4299, 0.7266, 1.3958, 1.6402, 1.6550, 1.3665], + device='cuda:4'), covar=tensor([0.0915, 0.0631, 0.0615, 0.0517, 0.0575, 0.0576, 0.0383, 0.0627], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0147, 0.0128, 0.0122, 0.0130, 0.0129, 0.0141, 0.0148], + device='cuda:4'), out_proj_covar=tensor([8.8943e-05, 1.0601e-04, 9.1068e-05, 8.5568e-05, 9.0961e-05, 9.1380e-05, + 1.0051e-04, 1.0563e-04], device='cuda:4') +2023-03-27 07:03:18,454 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.036e+02 1.460e+02 1.766e+02 2.058e+02 4.416e+02, threshold=3.532e+02, percent-clipped=3.0 +2023-03-27 07:03:19,189 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1383, 1.9533, 1.6818, 1.7636, 2.0404, 1.7832, 2.1301, 2.0911], + device='cuda:4'), covar=tensor([0.1274, 0.1918, 0.2934, 0.2382, 0.2438, 0.1706, 0.2611, 0.1695], + device='cuda:4'), in_proj_covar=tensor([0.0189, 0.0191, 0.0237, 0.0255, 0.0251, 0.0207, 0.0215, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:03:24,477 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=143277.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:03:30,576 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2625, 2.1261, 1.7945, 2.1412, 2.1527, 1.9023, 2.3949, 2.2824], + device='cuda:4'), covar=tensor([0.1257, 0.1961, 0.2682, 0.2369, 0.2373, 0.1709, 0.2678, 0.1595], + device='cuda:4'), in_proj_covar=tensor([0.0189, 0.0191, 0.0237, 0.0255, 0.0252, 0.0208, 0.0216, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:03:34,113 INFO [finetune.py:976] (4/7) Epoch 26, batch 100, loss[loss=0.1463, simple_loss=0.2241, pruned_loss=0.03424, over 4769.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2369, pruned_loss=0.04677, over 379676.28 frames. ], batch size: 28, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:03:34,195 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8094, 3.9865, 3.7331, 1.9505, 4.0231, 3.1531, 1.2777, 2.9506], + device='cuda:4'), covar=tensor([0.2089, 0.1779, 0.1624, 0.3166, 0.1035, 0.0936, 0.3933, 0.1269], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0178, 0.0161, 0.0130, 0.0160, 0.0123, 0.0148, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 07:04:07,506 INFO [finetune.py:976] (4/7) Epoch 26, batch 150, loss[loss=0.146, simple_loss=0.2101, pruned_loss=0.04094, over 4719.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2328, pruned_loss=0.04611, over 508399.80 frames. ], batch size: 23, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:04:15,749 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.92 vs. limit=5.0 +2023-03-27 07:04:25,693 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.744e+01 1.335e+02 1.679e+02 2.114e+02 2.886e+02, threshold=3.358e+02, percent-clipped=0.0 +2023-03-27 07:04:33,624 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=143380.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:04:41,262 INFO [finetune.py:976] (4/7) Epoch 26, batch 200, loss[loss=0.2053, simple_loss=0.2671, pruned_loss=0.07171, over 4827.00 frames. ], tot_loss[loss=0.1623, simple_loss=0.2325, pruned_loss=0.04602, over 606896.26 frames. ], batch size: 33, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:04:46,839 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=143400.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 07:05:05,300 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=143428.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:05:05,379 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4489, 2.4262, 1.9128, 2.5571, 2.4352, 2.0646, 2.9692, 2.5164], + device='cuda:4'), covar=tensor([0.1425, 0.2269, 0.3176, 0.2836, 0.2695, 0.1837, 0.2593, 0.1865], + device='cuda:4'), in_proj_covar=tensor([0.0190, 0.0191, 0.0237, 0.0256, 0.0252, 0.0208, 0.0216, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:05:22,139 INFO [finetune.py:976] (4/7) Epoch 26, batch 250, loss[loss=0.1811, simple_loss=0.2597, pruned_loss=0.05128, over 4804.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2385, pruned_loss=0.04853, over 685257.82 frames. ], batch size: 38, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:05:48,881 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=143461.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 07:05:51,351 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.9466, 3.5897, 3.1314, 2.0876, 3.4974, 3.0681, 2.8579, 3.2548], + device='cuda:4'), covar=tensor([0.0627, 0.0637, 0.1327, 0.1546, 0.0990, 0.1510, 0.1474, 0.0849], + device='cuda:4'), in_proj_covar=tensor([0.0172, 0.0193, 0.0202, 0.0183, 0.0212, 0.0213, 0.0226, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:05:53,069 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.741e+01 1.618e+02 1.961e+02 2.394e+02 5.476e+02, threshold=3.922e+02, percent-clipped=2.0 +2023-03-27 07:05:59,891 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=143473.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 07:05:59,922 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1074, 2.0234, 1.4151, 1.8574, 2.0130, 1.6415, 2.6495, 2.0683], + device='cuda:4'), covar=tensor([0.1356, 0.1819, 0.3291, 0.3108, 0.2650, 0.1797, 0.2192, 0.1858], + device='cuda:4'), in_proj_covar=tensor([0.0189, 0.0191, 0.0236, 0.0255, 0.0251, 0.0207, 0.0215, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:06:12,190 INFO [finetune.py:976] (4/7) Epoch 26, batch 300, loss[loss=0.1547, simple_loss=0.2269, pruned_loss=0.0412, over 4915.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2412, pruned_loss=0.04862, over 744428.32 frames. ], batch size: 36, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:06:40,200 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=143534.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 07:06:44,914 INFO [finetune.py:976] (4/7) Epoch 26, batch 350, loss[loss=0.1552, simple_loss=0.2334, pruned_loss=0.03846, over 4917.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2446, pruned_loss=0.04996, over 791472.18 frames. ], batch size: 38, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:06:51,511 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6851, 2.4789, 2.3797, 1.6200, 2.5879, 2.0412, 1.9050, 2.3308], + device='cuda:4'), covar=tensor([0.1170, 0.0757, 0.1440, 0.1754, 0.1300, 0.2041, 0.1947, 0.0943], + device='cuda:4'), in_proj_covar=tensor([0.0173, 0.0194, 0.0202, 0.0184, 0.0212, 0.0213, 0.0226, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:07:03,059 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.441e+02 1.724e+02 2.077e+02 3.544e+02, threshold=3.448e+02, percent-clipped=0.0 +2023-03-27 07:07:18,099 INFO [finetune.py:976] (4/7) Epoch 26, batch 400, loss[loss=0.1667, simple_loss=0.2326, pruned_loss=0.05036, over 4850.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2454, pruned_loss=0.04997, over 826207.92 frames. ], batch size: 44, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:07:40,603 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.78 vs. limit=5.0 +2023-03-27 07:07:54,070 INFO [finetune.py:976] (4/7) Epoch 26, batch 450, loss[loss=0.1691, simple_loss=0.2384, pruned_loss=0.04991, over 4712.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2455, pruned_loss=0.05046, over 855850.44 frames. ], batch size: 59, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:07:58,550 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4299, 2.1996, 2.2715, 1.9470, 2.4902, 2.7603, 2.7683, 1.5608], + device='cuda:4'), covar=tensor([0.0759, 0.1007, 0.0973, 0.0959, 0.0824, 0.0807, 0.0757, 0.2002], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0137, 0.0142, 0.0119, 0.0128, 0.0138, 0.0140, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:08:22,189 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.245e+01 1.544e+02 1.809e+02 2.165e+02 3.752e+02, threshold=3.619e+02, percent-clipped=3.0 +2023-03-27 07:08:37,482 INFO [finetune.py:976] (4/7) Epoch 26, batch 500, loss[loss=0.1797, simple_loss=0.2458, pruned_loss=0.05677, over 4832.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2422, pruned_loss=0.04946, over 877839.87 frames. ], batch size: 40, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:09:00,518 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7120, 1.8939, 1.4820, 1.4997, 2.1879, 2.2688, 1.7813, 1.8039], + device='cuda:4'), covar=tensor([0.0516, 0.0361, 0.0605, 0.0380, 0.0292, 0.0446, 0.0471, 0.0450], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0107, 0.0146, 0.0111, 0.0101, 0.0115, 0.0102, 0.0112], + device='cuda:4'), out_proj_covar=tensor([7.8055e-05, 8.1436e-05, 1.1360e-04, 8.4894e-05, 7.8390e-05, 8.5180e-05, + 7.5851e-05, 8.5470e-05], device='cuda:4') +2023-03-27 07:09:11,115 INFO [finetune.py:976] (4/7) Epoch 26, batch 550, loss[loss=0.1586, simple_loss=0.2262, pruned_loss=0.04547, over 4912.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2376, pruned_loss=0.04782, over 894435.62 frames. ], batch size: 36, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:09:14,301 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.5081, 4.7018, 5.0105, 5.2863, 5.2299, 4.9421, 5.5928, 1.8407], + device='cuda:4'), covar=tensor([0.0614, 0.0960, 0.0798, 0.0965, 0.1033, 0.1631, 0.0544, 0.5957], + device='cuda:4'), in_proj_covar=tensor([0.0351, 0.0249, 0.0282, 0.0297, 0.0338, 0.0289, 0.0307, 0.0303], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:09:20,246 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=143756.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 07:09:28,910 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.175e+01 1.443e+02 1.723e+02 1.984e+02 5.074e+02, threshold=3.446e+02, percent-clipped=2.0 +2023-03-27 07:09:44,563 INFO [finetune.py:976] (4/7) Epoch 26, batch 600, loss[loss=0.2037, simple_loss=0.279, pruned_loss=0.06423, over 4817.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.239, pruned_loss=0.04864, over 907753.66 frames. ], batch size: 39, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:10:09,699 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=143829.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 07:10:17,584 INFO [finetune.py:976] (4/7) Epoch 26, batch 650, loss[loss=0.1465, simple_loss=0.2321, pruned_loss=0.03043, over 4789.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2426, pruned_loss=0.04936, over 919307.61 frames. ], batch size: 28, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:10:40,474 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.52 vs. limit=5.0 +2023-03-27 07:10:41,254 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.600e+02 1.821e+02 2.293e+02 5.159e+02, threshold=3.642e+02, percent-clipped=4.0 +2023-03-27 07:11:12,408 INFO [finetune.py:976] (4/7) Epoch 26, batch 700, loss[loss=0.1586, simple_loss=0.2238, pruned_loss=0.0467, over 4783.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2446, pruned_loss=0.04969, over 926224.30 frames. ], batch size: 26, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:11:13,855 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-27 07:11:52,958 INFO [finetune.py:976] (4/7) Epoch 26, batch 750, loss[loss=0.2049, simple_loss=0.2816, pruned_loss=0.06414, over 4850.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2467, pruned_loss=0.05027, over 929165.73 frames. ], batch size: 44, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:12:00,354 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9000, 2.0412, 1.6570, 1.6796, 2.3704, 2.5014, 1.9348, 1.9171], + device='cuda:4'), covar=tensor([0.0386, 0.0344, 0.0576, 0.0379, 0.0265, 0.0450, 0.0398, 0.0379], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0107, 0.0146, 0.0111, 0.0102, 0.0116, 0.0103, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.8785e-05, 8.1819e-05, 1.1423e-04, 8.5213e-05, 7.9080e-05, 8.5572e-05, + 7.6285e-05, 8.5937e-05], device='cuda:4') +2023-03-27 07:12:02,812 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5125, 1.4174, 1.2673, 1.5525, 1.6517, 1.5432, 1.0326, 1.2741], + device='cuda:4'), covar=tensor([0.2130, 0.1967, 0.1940, 0.1647, 0.1592, 0.1275, 0.2434, 0.1906], + device='cuda:4'), in_proj_covar=tensor([0.0244, 0.0209, 0.0215, 0.0197, 0.0244, 0.0191, 0.0215, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:12:09,862 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.567e+02 1.788e+02 2.169e+02 3.888e+02, threshold=3.576e+02, percent-clipped=1.0 +2023-03-27 07:12:26,437 INFO [finetune.py:976] (4/7) Epoch 26, batch 800, loss[loss=0.2286, simple_loss=0.2876, pruned_loss=0.08482, over 4799.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2463, pruned_loss=0.04971, over 935930.74 frames. ], batch size: 45, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:13:00,658 INFO [finetune.py:976] (4/7) Epoch 26, batch 850, loss[loss=0.1906, simple_loss=0.2638, pruned_loss=0.05877, over 4818.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2453, pruned_loss=0.04986, over 940259.49 frames. ], batch size: 33, lr: 2.98e-03, grad_scale: 16.0 +2023-03-27 07:13:09,696 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=144056.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 07:13:16,916 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.460e+02 1.746e+02 2.115e+02 7.519e+02, threshold=3.492e+02, percent-clipped=2.0 +2023-03-27 07:13:43,959 INFO [finetune.py:976] (4/7) Epoch 26, batch 900, loss[loss=0.1342, simple_loss=0.2089, pruned_loss=0.02973, over 4744.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2427, pruned_loss=0.04913, over 945130.59 frames. ], batch size: 27, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:13:51,331 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=144104.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 07:14:07,459 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=144129.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 07:14:10,365 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7734, 1.5802, 1.5007, 1.8291, 1.9033, 1.8331, 1.3873, 1.4959], + device='cuda:4'), covar=tensor([0.2024, 0.1889, 0.1833, 0.1609, 0.1706, 0.1087, 0.2288, 0.1762], + device='cuda:4'), in_proj_covar=tensor([0.0244, 0.0209, 0.0214, 0.0197, 0.0244, 0.0190, 0.0215, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:14:16,857 INFO [finetune.py:976] (4/7) Epoch 26, batch 950, loss[loss=0.1659, simple_loss=0.2302, pruned_loss=0.05082, over 4002.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2411, pruned_loss=0.04896, over 946997.37 frames. ], batch size: 65, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:14:21,263 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-27 07:14:33,134 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.128e+02 1.468e+02 1.742e+02 2.065e+02 3.876e+02, threshold=3.485e+02, percent-clipped=2.0 +2023-03-27 07:14:39,229 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=144177.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 07:14:50,350 INFO [finetune.py:976] (4/7) Epoch 26, batch 1000, loss[loss=0.1882, simple_loss=0.2755, pruned_loss=0.05044, over 4833.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2442, pruned_loss=0.05033, over 949076.18 frames. ], batch size: 40, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:15:22,314 INFO [finetune.py:976] (4/7) Epoch 26, batch 1050, loss[loss=0.1815, simple_loss=0.25, pruned_loss=0.05652, over 4864.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2463, pruned_loss=0.05054, over 952183.55 frames. ], batch size: 44, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:15:40,002 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.483e+02 1.785e+02 2.219e+02 5.161e+02, threshold=3.570e+02, percent-clipped=2.0 +2023-03-27 07:16:01,563 INFO [finetune.py:976] (4/7) Epoch 26, batch 1100, loss[loss=0.2241, simple_loss=0.2842, pruned_loss=0.08205, over 4179.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.247, pruned_loss=0.05021, over 953711.45 frames. ], batch size: 65, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:16:38,599 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-27 07:16:56,472 INFO [finetune.py:976] (4/7) Epoch 26, batch 1150, loss[loss=0.169, simple_loss=0.2443, pruned_loss=0.04686, over 4817.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2469, pruned_loss=0.05046, over 952485.11 frames. ], batch size: 33, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:17:13,862 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.072e+02 1.499e+02 1.760e+02 2.197e+02 4.327e+02, threshold=3.521e+02, percent-clipped=2.0 +2023-03-27 07:17:20,673 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6627, 1.5014, 1.1143, 0.3341, 1.3115, 1.5144, 1.4555, 1.4175], + device='cuda:4'), covar=tensor([0.0924, 0.0757, 0.1326, 0.1869, 0.1300, 0.2142, 0.2101, 0.0896], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0191, 0.0200, 0.0181, 0.0210, 0.0210, 0.0223, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:17:30,187 INFO [finetune.py:976] (4/7) Epoch 26, batch 1200, loss[loss=0.1426, simple_loss=0.2154, pruned_loss=0.03493, over 4909.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2452, pruned_loss=0.0495, over 955029.65 frames. ], batch size: 37, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:17:43,710 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=144411.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:18:03,374 INFO [finetune.py:976] (4/7) Epoch 26, batch 1250, loss[loss=0.1735, simple_loss=0.2424, pruned_loss=0.05231, over 4816.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.243, pruned_loss=0.04926, over 952693.73 frames. ], batch size: 41, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:18:21,714 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.894e+01 1.606e+02 1.805e+02 2.274e+02 3.881e+02, threshold=3.611e+02, percent-clipped=1.0 +2023-03-27 07:18:24,280 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=144472.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:18:37,271 INFO [finetune.py:976] (4/7) Epoch 26, batch 1300, loss[loss=0.1786, simple_loss=0.2481, pruned_loss=0.05453, over 4903.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.239, pruned_loss=0.04779, over 954479.26 frames. ], batch size: 43, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:19:12,365 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4954, 2.3951, 2.0616, 2.6062, 2.4960, 2.2279, 2.7212, 2.4961], + device='cuda:4'), covar=tensor([0.1296, 0.1969, 0.2914, 0.2197, 0.2363, 0.1658, 0.2485, 0.1685], + device='cuda:4'), in_proj_covar=tensor([0.0188, 0.0190, 0.0235, 0.0253, 0.0249, 0.0206, 0.0214, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:19:21,353 INFO [finetune.py:976] (4/7) Epoch 26, batch 1350, loss[loss=0.1716, simple_loss=0.2487, pruned_loss=0.0472, over 4178.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2404, pruned_loss=0.04868, over 950614.90 frames. ], batch size: 65, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:19:39,470 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.034e+02 1.499e+02 1.803e+02 2.073e+02 4.281e+02, threshold=3.607e+02, percent-clipped=1.0 +2023-03-27 07:19:41,998 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1174, 3.5908, 3.7411, 4.0019, 3.9201, 3.6385, 4.2076, 1.2931], + device='cuda:4'), covar=tensor([0.0919, 0.1031, 0.0889, 0.1014, 0.1410, 0.1811, 0.0892, 0.6301], + device='cuda:4'), in_proj_covar=tensor([0.0352, 0.0249, 0.0281, 0.0297, 0.0338, 0.0288, 0.0307, 0.0303], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:19:54,498 INFO [finetune.py:976] (4/7) Epoch 26, batch 1400, loss[loss=0.1959, simple_loss=0.2655, pruned_loss=0.06313, over 4931.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2438, pruned_loss=0.04948, over 953512.32 frames. ], batch size: 33, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:20:27,739 INFO [finetune.py:976] (4/7) Epoch 26, batch 1450, loss[loss=0.1354, simple_loss=0.1956, pruned_loss=0.03761, over 4195.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2452, pruned_loss=0.04966, over 953989.47 frames. ], batch size: 18, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:20:42,310 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-27 07:20:45,824 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.092e+01 1.571e+02 1.855e+02 2.334e+02 4.645e+02, threshold=3.710e+02, percent-clipped=2.0 +2023-03-27 07:20:51,892 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2652, 1.9947, 2.2665, 1.5415, 2.0061, 2.2873, 2.2130, 1.6400], + device='cuda:4'), covar=tensor([0.0481, 0.0636, 0.0561, 0.0791, 0.0755, 0.0563, 0.0532, 0.1163], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0137, 0.0141, 0.0119, 0.0128, 0.0138, 0.0139, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:21:00,868 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6898, 1.6355, 1.6025, 1.7062, 1.4334, 4.2719, 1.5658, 2.0456], + device='cuda:4'), covar=tensor([0.3197, 0.2407, 0.2106, 0.2347, 0.1559, 0.0131, 0.2447, 0.1145], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0124, 0.0113, 0.0095, 0.0094, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 07:21:01,363 INFO [finetune.py:976] (4/7) Epoch 26, batch 1500, loss[loss=0.1432, simple_loss=0.2261, pruned_loss=0.03019, over 4926.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2463, pruned_loss=0.04951, over 955539.15 frames. ], batch size: 42, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:21:50,337 INFO [finetune.py:976] (4/7) Epoch 26, batch 1550, loss[loss=0.1725, simple_loss=0.2477, pruned_loss=0.0486, over 4891.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2456, pruned_loss=0.04911, over 955713.58 frames. ], batch size: 35, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:22:18,423 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=144767.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:22:18,964 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.202e+02 1.547e+02 1.850e+02 2.044e+02 4.068e+02, threshold=3.700e+02, percent-clipped=1.0 +2023-03-27 07:22:35,467 INFO [finetune.py:976] (4/7) Epoch 26, batch 1600, loss[loss=0.1216, simple_loss=0.1984, pruned_loss=0.02243, over 4804.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2442, pruned_loss=0.0488, over 955627.46 frames. ], batch size: 25, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:22:40,551 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4638, 1.3778, 1.3896, 1.3799, 0.8913, 2.2085, 0.7243, 1.2243], + device='cuda:4'), covar=tensor([0.3041, 0.2373, 0.2050, 0.2368, 0.1806, 0.0382, 0.2749, 0.1288], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0116, 0.0121, 0.0124, 0.0113, 0.0095, 0.0094, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 07:23:09,211 INFO [finetune.py:976] (4/7) Epoch 26, batch 1650, loss[loss=0.1575, simple_loss=0.2255, pruned_loss=0.04474, over 4807.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2416, pruned_loss=0.04808, over 957680.91 frames. ], batch size: 51, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:23:22,971 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5750, 1.6330, 1.4052, 1.5210, 2.0129, 1.8970, 1.6701, 1.4583], + device='cuda:4'), covar=tensor([0.0392, 0.0348, 0.0615, 0.0356, 0.0213, 0.0456, 0.0348, 0.0412], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0108, 0.0148, 0.0113, 0.0103, 0.0117, 0.0104, 0.0114], + device='cuda:4'), out_proj_covar=tensor([7.9678e-05, 8.2638e-05, 1.1579e-04, 8.6349e-05, 7.9614e-05, 8.6397e-05, + 7.7215e-05, 8.6961e-05], device='cuda:4') +2023-03-27 07:23:26,322 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.503e+01 1.460e+02 1.738e+02 2.010e+02 3.428e+02, threshold=3.475e+02, percent-clipped=0.0 +2023-03-27 07:23:42,421 INFO [finetune.py:976] (4/7) Epoch 26, batch 1700, loss[loss=0.1308, simple_loss=0.2052, pruned_loss=0.02821, over 4381.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2386, pruned_loss=0.04684, over 956883.73 frames. ], batch size: 19, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:23:54,453 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7318, 2.8492, 2.6297, 1.8354, 2.7272, 3.0020, 2.8712, 2.3722], + device='cuda:4'), covar=tensor([0.0543, 0.0511, 0.0666, 0.0814, 0.0662, 0.0562, 0.0533, 0.0930], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0137, 0.0141, 0.0119, 0.0128, 0.0138, 0.0140, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:24:19,026 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6249, 1.1755, 0.8622, 1.4792, 1.9519, 1.3127, 1.3824, 1.5351], + device='cuda:4'), covar=tensor([0.1489, 0.2013, 0.1876, 0.1231, 0.2017, 0.1950, 0.1439, 0.1788], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0110, 0.0092, 0.0119, 0.0093, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 07:24:25,822 INFO [finetune.py:976] (4/7) Epoch 26, batch 1750, loss[loss=0.1879, simple_loss=0.2651, pruned_loss=0.05532, over 4833.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2395, pruned_loss=0.04751, over 955186.43 frames. ], batch size: 33, lr: 2.97e-03, grad_scale: 16.0 +2023-03-27 07:24:30,646 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9024, 1.8087, 1.5568, 1.6320, 1.7250, 1.7016, 1.7582, 2.4060], + device='cuda:4'), covar=tensor([0.3579, 0.3882, 0.3124, 0.3694, 0.3698, 0.2360, 0.3447, 0.1604], + device='cuda:4'), in_proj_covar=tensor([0.0290, 0.0265, 0.0236, 0.0277, 0.0259, 0.0229, 0.0257, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:24:42,914 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.967e+01 1.592e+02 1.823e+02 2.389e+02 4.337e+02, threshold=3.645e+02, percent-clipped=3.0 +2023-03-27 07:24:59,584 INFO [finetune.py:976] (4/7) Epoch 26, batch 1800, loss[loss=0.1765, simple_loss=0.2556, pruned_loss=0.04874, over 4730.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2423, pruned_loss=0.04856, over 956663.96 frames. ], batch size: 59, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:25:02,601 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=144996.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:25:21,846 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2710, 1.2768, 1.5727, 1.0479, 1.3561, 1.4321, 1.2844, 1.6275], + device='cuda:4'), covar=tensor([0.1162, 0.2315, 0.1270, 0.1561, 0.0921, 0.1241, 0.3394, 0.0867], + device='cuda:4'), in_proj_covar=tensor([0.0190, 0.0206, 0.0191, 0.0189, 0.0174, 0.0212, 0.0216, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:25:23,065 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145027.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:25:33,497 INFO [finetune.py:976] (4/7) Epoch 26, batch 1850, loss[loss=0.2067, simple_loss=0.2788, pruned_loss=0.06729, over 4841.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2447, pruned_loss=0.04948, over 957198.22 frames. ], batch size: 47, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:25:40,774 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2980, 2.1113, 1.7347, 2.0867, 2.1266, 1.8726, 2.4050, 2.2694], + device='cuda:4'), covar=tensor([0.1183, 0.1781, 0.2685, 0.2378, 0.2445, 0.1560, 0.2719, 0.1425], + device='cuda:4'), in_proj_covar=tensor([0.0187, 0.0189, 0.0235, 0.0253, 0.0250, 0.0207, 0.0214, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:25:43,164 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145057.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 07:25:48,490 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145066.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:25:49,086 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=145067.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:25:49,582 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.533e+02 1.829e+02 2.183e+02 4.392e+02, threshold=3.659e+02, percent-clipped=3.0 +2023-03-27 07:26:03,652 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145088.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:26:06,519 INFO [finetune.py:976] (4/7) Epoch 26, batch 1900, loss[loss=0.1572, simple_loss=0.2331, pruned_loss=0.04068, over 4862.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2464, pruned_loss=0.05008, over 955982.60 frames. ], batch size: 31, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:26:08,810 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5555, 1.4573, 2.1028, 3.3173, 2.1556, 2.3014, 0.8969, 2.8046], + device='cuda:4'), covar=tensor([0.1682, 0.1402, 0.1245, 0.0505, 0.0859, 0.1862, 0.1895, 0.0453], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0163, 0.0101, 0.0134, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 07:26:20,772 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.9560, 0.9015, 0.8615, 1.0069, 1.1142, 1.0647, 0.9235, 0.8516], + device='cuda:4'), covar=tensor([0.0456, 0.0374, 0.0705, 0.0360, 0.0307, 0.0521, 0.0385, 0.0485], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0108, 0.0149, 0.0113, 0.0102, 0.0117, 0.0104, 0.0114], + device='cuda:4'), out_proj_covar=tensor([7.9699e-05, 8.2735e-05, 1.1600e-04, 8.6256e-05, 7.9516e-05, 8.6305e-05, + 7.7119e-05, 8.6935e-05], device='cuda:4') +2023-03-27 07:26:21,298 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=145115.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:26:29,581 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145127.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:26:36,357 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.49 vs. limit=5.0 +2023-03-27 07:26:46,688 INFO [finetune.py:976] (4/7) Epoch 26, batch 1950, loss[loss=0.154, simple_loss=0.2291, pruned_loss=0.03942, over 4755.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2462, pruned_loss=0.05062, over 956481.34 frames. ], batch size: 26, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:27:15,918 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.125e+02 1.577e+02 1.831e+02 2.188e+02 4.363e+02, threshold=3.662e+02, percent-clipped=3.0 +2023-03-27 07:27:36,122 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.2845, 2.8909, 3.0500, 3.1952, 3.0845, 2.9183, 3.3182, 0.9765], + device='cuda:4'), covar=tensor([0.1058, 0.1057, 0.1031, 0.1145, 0.1497, 0.1776, 0.1122, 0.5618], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0249, 0.0279, 0.0295, 0.0336, 0.0286, 0.0304, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:27:40,099 INFO [finetune.py:976] (4/7) Epoch 26, batch 2000, loss[loss=0.1642, simple_loss=0.2326, pruned_loss=0.04791, over 4808.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2435, pruned_loss=0.05019, over 956959.29 frames. ], batch size: 25, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:27:52,177 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.16 vs. limit=5.0 +2023-03-27 07:28:13,278 INFO [finetune.py:976] (4/7) Epoch 26, batch 2050, loss[loss=0.1535, simple_loss=0.2323, pruned_loss=0.0374, over 4934.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2412, pruned_loss=0.04933, over 956707.63 frames. ], batch size: 33, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:28:30,375 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.187e+01 1.436e+02 1.792e+02 2.264e+02 4.038e+02, threshold=3.583e+02, percent-clipped=1.0 +2023-03-27 07:28:45,865 INFO [finetune.py:976] (4/7) Epoch 26, batch 2100, loss[loss=0.1663, simple_loss=0.2404, pruned_loss=0.04617, over 4858.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2416, pruned_loss=0.04986, over 954956.33 frames. ], batch size: 44, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:29:19,663 INFO [finetune.py:976] (4/7) Epoch 26, batch 2150, loss[loss=0.1502, simple_loss=0.2263, pruned_loss=0.03704, over 4763.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.243, pruned_loss=0.04969, over 954400.19 frames. ], batch size: 28, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:29:28,920 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145352.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 07:29:47,543 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.022e+02 1.494e+02 1.709e+02 2.298e+02 6.165e+02, threshold=3.419e+02, percent-clipped=3.0 +2023-03-27 07:29:49,486 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145371.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:29:56,781 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145383.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:30:02,628 INFO [finetune.py:976] (4/7) Epoch 26, batch 2200, loss[loss=0.1741, simple_loss=0.2484, pruned_loss=0.04986, over 4834.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2443, pruned_loss=0.04992, over 952413.58 frames. ], batch size: 30, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:30:20,825 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6867, 3.3944, 3.3087, 1.9288, 3.5336, 2.8395, 1.4585, 2.6057], + device='cuda:4'), covar=tensor([0.2852, 0.2177, 0.1545, 0.2789, 0.1019, 0.0938, 0.3528, 0.1347], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0179, 0.0160, 0.0130, 0.0160, 0.0124, 0.0149, 0.0123], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 07:30:23,253 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145422.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:30:29,361 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145432.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:30:30,783 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-03-27 07:30:36,242 INFO [finetune.py:976] (4/7) Epoch 26, batch 2250, loss[loss=0.1997, simple_loss=0.262, pruned_loss=0.06873, over 4925.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2456, pruned_loss=0.05019, over 953215.32 frames. ], batch size: 33, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:30:53,946 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.189e+01 1.488e+02 1.760e+02 2.143e+02 3.776e+02, threshold=3.521e+02, percent-clipped=2.0 +2023-03-27 07:31:08,986 INFO [finetune.py:976] (4/7) Epoch 26, batch 2300, loss[loss=0.2005, simple_loss=0.2711, pruned_loss=0.06497, over 4813.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2455, pruned_loss=0.04998, over 952564.18 frames. ], batch size: 38, lr: 2.97e-03, grad_scale: 32.0 +2023-03-27 07:31:42,488 INFO [finetune.py:976] (4/7) Epoch 26, batch 2350, loss[loss=0.1064, simple_loss=0.1826, pruned_loss=0.01513, over 4817.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2432, pruned_loss=0.04864, over 955860.27 frames. ], batch size: 25, lr: 2.96e-03, grad_scale: 32.0 +2023-03-27 07:31:57,292 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2701, 1.8767, 2.2827, 2.2610, 1.9669, 1.9860, 2.2789, 2.1532], + device='cuda:4'), covar=tensor([0.3835, 0.3704, 0.2905, 0.3629, 0.4754, 0.3932, 0.4365, 0.2955], + device='cuda:4'), in_proj_covar=tensor([0.0266, 0.0248, 0.0267, 0.0295, 0.0295, 0.0271, 0.0301, 0.0252], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:32:06,625 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.544e+02 1.840e+02 2.226e+02 4.643e+02, threshold=3.680e+02, percent-clipped=1.0 +2023-03-27 07:32:34,348 INFO [finetune.py:976] (4/7) Epoch 26, batch 2400, loss[loss=0.1873, simple_loss=0.2465, pruned_loss=0.06409, over 4832.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2405, pruned_loss=0.04801, over 957226.47 frames. ], batch size: 39, lr: 2.96e-03, grad_scale: 32.0 +2023-03-27 07:32:35,657 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145594.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:32:37,936 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8709, 1.7696, 1.6334, 2.0348, 2.2683, 1.9458, 1.5626, 1.5460], + device='cuda:4'), covar=tensor([0.2037, 0.1864, 0.1827, 0.1534, 0.1473, 0.1142, 0.2274, 0.1788], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0211, 0.0216, 0.0199, 0.0246, 0.0192, 0.0218, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:33:17,103 INFO [finetune.py:976] (4/7) Epoch 26, batch 2450, loss[loss=0.164, simple_loss=0.2338, pruned_loss=0.04712, over 4815.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2388, pruned_loss=0.04794, over 958012.50 frames. ], batch size: 38, lr: 2.96e-03, grad_scale: 32.0 +2023-03-27 07:33:17,820 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145643.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:33:23,855 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=145652.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 07:33:26,195 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145655.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 07:33:34,916 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.521e+01 1.449e+02 1.824e+02 2.163e+02 4.630e+02, threshold=3.648e+02, percent-clipped=2.0 +2023-03-27 07:33:45,644 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=145683.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:33:51,041 INFO [finetune.py:976] (4/7) Epoch 26, batch 2500, loss[loss=0.1786, simple_loss=0.2389, pruned_loss=0.05919, over 4909.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2411, pruned_loss=0.04921, over 957156.37 frames. ], batch size: 32, lr: 2.96e-03, grad_scale: 32.0 +2023-03-27 07:33:55,944 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=145700.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:33:58,914 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145704.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:34:11,630 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=145722.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:34:15,150 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145727.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:34:16,468 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.13 vs. limit=5.0 +2023-03-27 07:34:17,566 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=145731.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:34:24,605 INFO [finetune.py:976] (4/7) Epoch 26, batch 2550, loss[loss=0.1622, simple_loss=0.2304, pruned_loss=0.04699, over 4779.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2445, pruned_loss=0.05011, over 957389.41 frames. ], batch size: 26, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:34:42,325 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.531e+01 1.551e+02 1.807e+02 2.106e+02 4.459e+02, threshold=3.615e+02, percent-clipped=2.0 +2023-03-27 07:34:43,554 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=145770.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:35:08,873 INFO [finetune.py:976] (4/7) Epoch 26, batch 2600, loss[loss=0.1486, simple_loss=0.2292, pruned_loss=0.034, over 4856.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2462, pruned_loss=0.0508, over 957516.06 frames. ], batch size: 44, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:35:10,431 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-27 07:35:28,220 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=145821.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 07:35:42,708 INFO [finetune.py:976] (4/7) Epoch 26, batch 2650, loss[loss=0.1565, simple_loss=0.2272, pruned_loss=0.04286, over 4797.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2475, pruned_loss=0.0511, over 955704.93 frames. ], batch size: 29, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:36:00,020 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.986e+01 1.512e+02 1.783e+02 2.110e+02 4.476e+02, threshold=3.566e+02, percent-clipped=1.0 +2023-03-27 07:36:09,584 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=145882.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 07:36:16,426 INFO [finetune.py:976] (4/7) Epoch 26, batch 2700, loss[loss=0.1367, simple_loss=0.2058, pruned_loss=0.03379, over 4748.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2462, pruned_loss=0.0509, over 954156.96 frames. ], batch size: 27, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:36:49,651 INFO [finetune.py:976] (4/7) Epoch 26, batch 2750, loss[loss=0.171, simple_loss=0.2282, pruned_loss=0.05685, over 4824.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2434, pruned_loss=0.05048, over 953511.68 frames. ], batch size: 30, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:36:55,117 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145950.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 07:37:07,588 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.569e+02 1.804e+02 2.200e+02 3.850e+02, threshold=3.609e+02, percent-clipped=2.0 +2023-03-27 07:37:29,452 INFO [finetune.py:976] (4/7) Epoch 26, batch 2800, loss[loss=0.1722, simple_loss=0.245, pruned_loss=0.04973, over 4906.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2398, pruned_loss=0.04921, over 954103.82 frames. ], batch size: 46, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:37:39,467 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=145999.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:38:14,436 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146027.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:38:24,986 INFO [finetune.py:976] (4/7) Epoch 26, batch 2850, loss[loss=0.2198, simple_loss=0.2802, pruned_loss=0.07973, over 4937.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2388, pruned_loss=0.04922, over 951835.61 frames. ], batch size: 33, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:38:42,241 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.185e+01 1.523e+02 1.819e+02 2.110e+02 4.930e+02, threshold=3.638e+02, percent-clipped=2.0 +2023-03-27 07:38:46,430 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=146075.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:38:58,304 INFO [finetune.py:976] (4/7) Epoch 26, batch 2900, loss[loss=0.1887, simple_loss=0.2627, pruned_loss=0.05733, over 4925.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.242, pruned_loss=0.05036, over 952746.28 frames. ], batch size: 38, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:39:03,970 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-27 07:39:25,177 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1349, 1.7774, 2.5861, 4.1455, 2.8005, 2.8895, 0.9754, 3.6610], + device='cuda:4'), covar=tensor([0.1711, 0.1501, 0.1440, 0.0575, 0.0778, 0.1555, 0.2100, 0.0350], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0163, 0.0101, 0.0134, 0.0125, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 07:39:31,502 INFO [finetune.py:976] (4/7) Epoch 26, batch 2950, loss[loss=0.2014, simple_loss=0.2701, pruned_loss=0.06634, over 4826.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2458, pruned_loss=0.05174, over 952567.52 frames. ], batch size: 30, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:39:36,362 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=146149.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:39:49,294 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.099e+02 1.550e+02 1.859e+02 2.106e+02 3.478e+02, threshold=3.719e+02, percent-clipped=0.0 +2023-03-27 07:39:54,696 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=146177.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 07:40:04,824 INFO [finetune.py:976] (4/7) Epoch 26, batch 3000, loss[loss=0.1794, simple_loss=0.2649, pruned_loss=0.0469, over 4912.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2466, pruned_loss=0.05123, over 953739.44 frames. ], batch size: 36, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:40:04,824 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 07:40:19,936 INFO [finetune.py:1010] (4/7) Epoch 26, validation: loss=0.1577, simple_loss=0.2252, pruned_loss=0.04507, over 2265189.00 frames. +2023-03-27 07:40:19,936 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6557MB +2023-03-27 07:40:35,418 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=146210.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:40:43,355 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-27 07:40:56,715 INFO [finetune.py:976] (4/7) Epoch 26, batch 3050, loss[loss=0.1781, simple_loss=0.2513, pruned_loss=0.05248, over 4861.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2469, pruned_loss=0.05039, over 955612.47 frames. ], batch size: 34, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:41:02,141 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146250.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:41:02,242 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-27 07:41:14,892 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.646e+01 1.408e+02 1.795e+02 2.163e+02 4.679e+02, threshold=3.589e+02, percent-clipped=3.0 +2023-03-27 07:41:29,843 INFO [finetune.py:976] (4/7) Epoch 26, batch 3100, loss[loss=0.1847, simple_loss=0.2494, pruned_loss=0.05997, over 4838.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2449, pruned_loss=0.04993, over 954987.51 frames. ], batch size: 49, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:41:30,580 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0763, 1.0162, 1.0019, 0.4075, 0.9537, 1.1460, 1.1938, 1.0172], + device='cuda:4'), covar=tensor([0.0894, 0.0627, 0.0602, 0.0585, 0.0596, 0.0665, 0.0430, 0.0678], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0148, 0.0128, 0.0123, 0.0130, 0.0129, 0.0142, 0.0149], + device='cuda:4'), out_proj_covar=tensor([8.9023e-05, 1.0654e-04, 9.1137e-05, 8.6342e-05, 9.1146e-05, 9.1843e-05, + 1.0121e-04, 1.0643e-04], device='cuda:4') +2023-03-27 07:41:34,031 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=146298.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:41:34,687 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146299.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:41:42,618 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-27 07:42:02,608 INFO [finetune.py:976] (4/7) Epoch 26, batch 3150, loss[loss=0.1863, simple_loss=0.2411, pruned_loss=0.0658, over 4844.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2424, pruned_loss=0.04933, over 955568.57 frames. ], batch size: 25, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:42:06,564 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=146347.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:42:15,238 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3130, 1.2863, 1.4563, 0.7441, 1.4608, 1.7244, 1.6161, 1.3477], + device='cuda:4'), covar=tensor([0.1197, 0.1046, 0.0658, 0.0633, 0.0608, 0.0569, 0.0518, 0.0793], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0149, 0.0128, 0.0123, 0.0130, 0.0130, 0.0143, 0.0149], + device='cuda:4'), out_proj_covar=tensor([8.9315e-05, 1.0687e-04, 9.1304e-05, 8.6614e-05, 9.1327e-05, 9.2011e-05, + 1.0161e-04, 1.0683e-04], device='cuda:4') +2023-03-27 07:42:21,140 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.044e+02 1.464e+02 1.851e+02 2.163e+02 3.423e+02, threshold=3.701e+02, percent-clipped=0.0 +2023-03-27 07:42:38,055 INFO [finetune.py:976] (4/7) Epoch 26, batch 3200, loss[loss=0.2022, simple_loss=0.2652, pruned_loss=0.06961, over 4838.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2393, pruned_loss=0.04802, over 955257.89 frames. ], batch size: 39, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:43:35,344 INFO [finetune.py:976] (4/7) Epoch 26, batch 3250, loss[loss=0.2031, simple_loss=0.2818, pruned_loss=0.06216, over 4211.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2412, pruned_loss=0.04907, over 953323.22 frames. ], batch size: 65, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:43:48,550 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 07:43:53,729 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.041e+02 1.451e+02 1.808e+02 2.175e+02 4.535e+02, threshold=3.616e+02, percent-clipped=3.0 +2023-03-27 07:43:58,659 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146477.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 07:44:08,655 INFO [finetune.py:976] (4/7) Epoch 26, batch 3300, loss[loss=0.1839, simple_loss=0.2538, pruned_loss=0.05702, over 4199.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2458, pruned_loss=0.05045, over 952443.81 frames. ], batch size: 65, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:44:17,128 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=146505.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:44:30,710 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=146525.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 07:44:41,544 INFO [finetune.py:976] (4/7) Epoch 26, batch 3350, loss[loss=0.1584, simple_loss=0.2348, pruned_loss=0.04102, over 4903.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2453, pruned_loss=0.04936, over 952543.49 frames. ], batch size: 37, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:44:53,457 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6559, 2.5563, 2.2136, 1.1281, 2.3225, 1.9856, 1.9189, 2.3716], + device='cuda:4'), covar=tensor([0.1052, 0.0735, 0.1588, 0.2031, 0.1478, 0.2267, 0.2087, 0.0911], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0191, 0.0199, 0.0181, 0.0208, 0.0209, 0.0223, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:45:00,357 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.852e+01 1.562e+02 1.841e+02 2.282e+02 4.006e+02, threshold=3.682e+02, percent-clipped=1.0 +2023-03-27 07:45:14,935 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=146591.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:45:15,431 INFO [finetune.py:976] (4/7) Epoch 26, batch 3400, loss[loss=0.1169, simple_loss=0.1896, pruned_loss=0.0221, over 4685.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2464, pruned_loss=0.0498, over 953921.95 frames. ], batch size: 23, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:45:46,304 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0636, 1.7740, 2.3142, 1.4620, 2.0334, 2.3853, 1.7336, 2.4133], + device='cuda:4'), covar=tensor([0.1241, 0.2206, 0.1390, 0.2088, 0.1001, 0.1228, 0.2835, 0.0859], + device='cuda:4'), in_proj_covar=tensor([0.0191, 0.0206, 0.0193, 0.0190, 0.0175, 0.0213, 0.0217, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:45:48,152 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=146626.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:45:58,658 INFO [finetune.py:976] (4/7) Epoch 26, batch 3450, loss[loss=0.1878, simple_loss=0.2608, pruned_loss=0.05744, over 4858.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2451, pruned_loss=0.04926, over 954242.23 frames. ], batch size: 34, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:46:03,013 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8138, 4.0873, 3.8069, 1.8040, 4.0957, 3.0475, 0.9813, 2.8957], + device='cuda:4'), covar=tensor([0.2225, 0.2117, 0.1483, 0.3483, 0.1021, 0.0978, 0.4427, 0.1345], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0181, 0.0162, 0.0131, 0.0163, 0.0125, 0.0151, 0.0125], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 07:46:04,887 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=146652.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:46:17,057 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.040e+02 1.450e+02 1.708e+02 2.017e+02 4.995e+02, threshold=3.417e+02, percent-clipped=1.0 +2023-03-27 07:46:27,902 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7830, 3.7846, 3.5019, 1.9943, 3.8713, 2.8237, 0.8822, 2.6469], + device='cuda:4'), covar=tensor([0.2198, 0.1893, 0.1603, 0.3147, 0.1067, 0.1076, 0.4555, 0.1551], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0180, 0.0161, 0.0131, 0.0162, 0.0124, 0.0150, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 07:46:28,565 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=146687.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 07:46:32,392 INFO [finetune.py:976] (4/7) Epoch 26, batch 3500, loss[loss=0.1752, simple_loss=0.2522, pruned_loss=0.04912, over 4803.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2414, pruned_loss=0.0481, over 954402.59 frames. ], batch size: 51, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:47:05,281 INFO [finetune.py:976] (4/7) Epoch 26, batch 3550, loss[loss=0.1322, simple_loss=0.2054, pruned_loss=0.0295, over 4305.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2392, pruned_loss=0.0477, over 955397.12 frames. ], batch size: 19, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:47:15,011 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0706, 2.1036, 1.6871, 1.7712, 2.5085, 2.5626, 2.0854, 2.0181], + device='cuda:4'), covar=tensor([0.0443, 0.0358, 0.0629, 0.0391, 0.0244, 0.0488, 0.0363, 0.0404], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0107, 0.0147, 0.0112, 0.0102, 0.0117, 0.0104, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.9298e-05, 8.2150e-05, 1.1478e-04, 8.5592e-05, 7.8677e-05, 8.6100e-05, + 7.7227e-05, 8.6056e-05], device='cuda:4') +2023-03-27 07:47:22,721 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.051e+02 1.430e+02 1.748e+02 2.315e+02 5.079e+02, threshold=3.497e+02, percent-clipped=6.0 +2023-03-27 07:47:28,057 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9564, 1.9840, 1.6837, 2.0902, 2.5587, 2.0965, 1.9668, 1.5687], + device='cuda:4'), covar=tensor([0.2106, 0.1834, 0.1847, 0.1565, 0.1536, 0.1135, 0.2034, 0.1891], + device='cuda:4'), in_proj_covar=tensor([0.0247, 0.0212, 0.0216, 0.0200, 0.0247, 0.0192, 0.0219, 0.0206], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:47:38,113 INFO [finetune.py:976] (4/7) Epoch 26, batch 3600, loss[loss=0.204, simple_loss=0.2808, pruned_loss=0.06363, over 4804.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2377, pruned_loss=0.04714, over 956427.84 frames. ], batch size: 45, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:47:47,319 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=146805.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:48:24,649 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.80 vs. limit=5.0 +2023-03-27 07:48:26,742 INFO [finetune.py:976] (4/7) Epoch 26, batch 3650, loss[loss=0.1413, simple_loss=0.1932, pruned_loss=0.04476, over 3910.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.242, pruned_loss=0.04941, over 953655.25 frames. ], batch size: 17, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:48:28,102 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=146844.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:48:39,244 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=146853.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:48:53,566 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.488e+01 1.511e+02 1.813e+02 2.229e+02 3.524e+02, threshold=3.627e+02, percent-clipped=1.0 +2023-03-27 07:49:03,902 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-27 07:49:12,953 INFO [finetune.py:976] (4/7) Epoch 26, batch 3700, loss[loss=0.1866, simple_loss=0.2686, pruned_loss=0.05227, over 4849.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2462, pruned_loss=0.05071, over 954627.50 frames. ], batch size: 49, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:49:21,437 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=146905.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:49:46,520 INFO [finetune.py:976] (4/7) Epoch 26, batch 3750, loss[loss=0.1599, simple_loss=0.2341, pruned_loss=0.0428, over 4834.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.247, pruned_loss=0.05099, over 954426.70 frames. ], batch size: 49, lr: 2.96e-03, grad_scale: 16.0 +2023-03-27 07:49:49,617 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=146947.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:50:03,846 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.820e+01 1.503e+02 1.791e+02 2.461e+02 5.017e+02, threshold=3.581e+02, percent-clipped=5.0 +2023-03-27 07:50:08,697 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7889, 1.3588, 1.8681, 1.8798, 1.6491, 1.6165, 1.8342, 1.7946], + device='cuda:4'), covar=tensor([0.3842, 0.3603, 0.3001, 0.3542, 0.4450, 0.3739, 0.4325, 0.2896], + device='cuda:4'), in_proj_covar=tensor([0.0265, 0.0246, 0.0266, 0.0294, 0.0295, 0.0270, 0.0301, 0.0251], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:50:12,646 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=146982.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 07:50:19,584 INFO [finetune.py:976] (4/7) Epoch 26, batch 3800, loss[loss=0.1448, simple_loss=0.2232, pruned_loss=0.03314, over 4772.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2477, pruned_loss=0.05129, over 955123.07 frames. ], batch size: 27, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:50:55,311 INFO [finetune.py:976] (4/7) Epoch 26, batch 3850, loss[loss=0.1757, simple_loss=0.2435, pruned_loss=0.05401, over 4904.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2465, pruned_loss=0.05092, over 954377.98 frames. ], batch size: 43, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:51:03,546 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 07:51:21,146 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.517e+02 1.854e+02 2.266e+02 5.483e+02, threshold=3.707e+02, percent-clipped=2.0 +2023-03-27 07:51:37,020 INFO [finetune.py:976] (4/7) Epoch 26, batch 3900, loss[loss=0.1941, simple_loss=0.2588, pruned_loss=0.0647, over 4899.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2433, pruned_loss=0.04997, over 954999.53 frames. ], batch size: 32, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:52:06,854 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-27 07:52:09,602 INFO [finetune.py:976] (4/7) Epoch 26, batch 3950, loss[loss=0.1881, simple_loss=0.2578, pruned_loss=0.05923, over 4778.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2398, pruned_loss=0.04853, over 956347.81 frames. ], batch size: 26, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:52:14,973 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8825, 1.7100, 1.5285, 1.4501, 1.6624, 1.6526, 1.6597, 2.2302], + device='cuda:4'), covar=tensor([0.3507, 0.3589, 0.2854, 0.3158, 0.3661, 0.2206, 0.3230, 0.1735], + device='cuda:4'), in_proj_covar=tensor([0.0292, 0.0266, 0.0238, 0.0278, 0.0261, 0.0231, 0.0259, 0.0240], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:52:27,907 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.492e+02 1.682e+02 1.976e+02 2.814e+02, threshold=3.365e+02, percent-clipped=0.0 +2023-03-27 07:52:34,858 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-27 07:52:42,787 INFO [finetune.py:976] (4/7) Epoch 26, batch 4000, loss[loss=0.2199, simple_loss=0.2763, pruned_loss=0.08175, over 4811.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2406, pruned_loss=0.04963, over 955935.95 frames. ], batch size: 40, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:52:48,872 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147200.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:53:00,217 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.64 vs. limit=2.0 +2023-03-27 07:53:15,828 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147241.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:53:16,325 INFO [finetune.py:976] (4/7) Epoch 26, batch 4050, loss[loss=0.1186, simple_loss=0.1861, pruned_loss=0.02555, over 4766.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.243, pruned_loss=0.04979, over 954801.38 frames. ], batch size: 23, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:53:21,636 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147247.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:53:41,925 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1724, 1.7828, 2.4115, 4.1412, 2.7216, 2.8000, 0.7752, 3.5129], + device='cuda:4'), covar=tensor([0.1739, 0.1470, 0.1472, 0.0573, 0.0879, 0.1361, 0.2203, 0.0409], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0162, 0.0100, 0.0133, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 07:53:48,587 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.158e+02 1.660e+02 1.920e+02 2.375e+02 4.575e+02, threshold=3.840e+02, percent-clipped=6.0 +2023-03-27 07:53:58,884 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147282.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:54:08,318 INFO [finetune.py:976] (4/7) Epoch 26, batch 4100, loss[loss=0.1999, simple_loss=0.2636, pruned_loss=0.06804, over 4830.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2462, pruned_loss=0.05072, over 955765.90 frames. ], batch size: 33, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:54:13,960 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=147295.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:54:18,780 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147302.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:54:24,628 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147310.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:54:37,224 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=147330.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:54:44,911 INFO [finetune.py:976] (4/7) Epoch 26, batch 4150, loss[loss=0.1633, simple_loss=0.2348, pruned_loss=0.04591, over 4760.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2467, pruned_loss=0.05084, over 955275.70 frames. ], batch size: 28, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:54:53,405 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-03-27 07:55:03,460 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.520e+02 1.873e+02 2.208e+02 5.004e+02, threshold=3.746e+02, percent-clipped=1.0 +2023-03-27 07:55:05,301 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147371.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:55:09,574 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1464, 1.3429, 1.4003, 0.7892, 1.4040, 1.6771, 1.6791, 1.3307], + device='cuda:4'), covar=tensor([0.0859, 0.0578, 0.0558, 0.0517, 0.0450, 0.0598, 0.0359, 0.0734], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0148, 0.0128, 0.0122, 0.0130, 0.0129, 0.0141, 0.0149], + device='cuda:4'), out_proj_covar=tensor([8.8712e-05, 1.0656e-04, 9.1380e-05, 8.5734e-05, 9.0928e-05, 9.1456e-05, + 1.0061e-04, 1.0637e-04], device='cuda:4') +2023-03-27 07:55:18,278 INFO [finetune.py:976] (4/7) Epoch 26, batch 4200, loss[loss=0.1665, simple_loss=0.238, pruned_loss=0.04747, over 4842.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2471, pruned_loss=0.05064, over 956332.37 frames. ], batch size: 44, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:55:28,554 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147407.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:55:46,438 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147434.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:55:51,222 INFO [finetune.py:976] (4/7) Epoch 26, batch 4250, loss[loss=0.2016, simple_loss=0.2697, pruned_loss=0.06678, over 4812.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2445, pruned_loss=0.0497, over 956809.53 frames. ], batch size: 41, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:56:01,250 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3130, 2.3413, 2.2130, 2.6068, 2.6706, 2.4954, 2.2692, 1.9391], + device='cuda:4'), covar=tensor([0.1877, 0.1615, 0.1499, 0.1339, 0.1585, 0.0960, 0.1803, 0.1608], + device='cuda:4'), in_proj_covar=tensor([0.0245, 0.0210, 0.0215, 0.0198, 0.0245, 0.0190, 0.0217, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:56:16,332 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147468.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:56:16,792 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.473e+02 1.815e+02 2.255e+02 8.587e+02, threshold=3.630e+02, percent-clipped=2.0 +2023-03-27 07:56:34,775 INFO [finetune.py:976] (4/7) Epoch 26, batch 4300, loss[loss=0.1478, simple_loss=0.219, pruned_loss=0.03828, over 4760.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2415, pruned_loss=0.0485, over 958469.53 frames. ], batch size: 28, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:56:36,682 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0979, 1.7295, 2.3574, 3.8937, 2.6529, 2.7434, 0.8296, 3.1929], + device='cuda:4'), covar=tensor([0.1526, 0.1372, 0.1359, 0.0578, 0.0729, 0.1876, 0.1920, 0.0389], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0163, 0.0101, 0.0134, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 07:56:36,728 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147495.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:56:37,815 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7920, 1.6123, 2.1909, 3.7077, 2.4348, 2.5719, 1.0581, 3.0412], + device='cuda:4'), covar=tensor([0.1799, 0.1406, 0.1436, 0.0619, 0.0838, 0.1409, 0.2050, 0.0481], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0163, 0.0101, 0.0134, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 07:56:40,204 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147500.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:57:07,239 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-27 07:57:08,538 INFO [finetune.py:976] (4/7) Epoch 26, batch 4350, loss[loss=0.158, simple_loss=0.229, pruned_loss=0.04345, over 4777.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2383, pruned_loss=0.04737, over 958874.60 frames. ], batch size: 26, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:57:11,719 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9741, 1.8373, 2.2842, 1.4692, 2.0603, 2.2241, 1.6528, 2.3871], + device='cuda:4'), covar=tensor([0.1163, 0.1758, 0.1217, 0.1782, 0.0886, 0.1366, 0.2775, 0.0799], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0208, 0.0194, 0.0191, 0.0177, 0.0215, 0.0218, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:57:12,266 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=147548.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:57:13,049 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-27 07:57:17,684 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5902, 1.4843, 1.9671, 1.7823, 1.5884, 3.4682, 1.4172, 1.6112], + device='cuda:4'), covar=tensor([0.0920, 0.1690, 0.1044, 0.0893, 0.1558, 0.0207, 0.1456, 0.1769], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0073, 0.0076, 0.0091, 0.0080, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 07:57:26,970 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.870e+01 1.462e+02 1.667e+02 1.917e+02 5.708e+02, threshold=3.333e+02, percent-clipped=2.0 +2023-03-27 07:57:38,205 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6386, 3.3731, 3.1389, 1.4868, 3.5003, 2.6034, 0.6685, 2.3299], + device='cuda:4'), covar=tensor([0.2495, 0.2642, 0.1842, 0.3748, 0.1394, 0.1183, 0.4773, 0.1771], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0179, 0.0160, 0.0129, 0.0161, 0.0124, 0.0149, 0.0125], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 07:57:42,396 INFO [finetune.py:976] (4/7) Epoch 26, batch 4400, loss[loss=0.1958, simple_loss=0.2769, pruned_loss=0.05733, over 4871.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2398, pruned_loss=0.04848, over 957851.41 frames. ], batch size: 34, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:57:45,507 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147597.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:57:58,101 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6663, 2.7967, 2.6326, 1.9158, 2.8073, 2.9900, 2.8008, 2.4159], + device='cuda:4'), covar=tensor([0.0603, 0.0610, 0.0707, 0.0889, 0.0604, 0.0674, 0.0693, 0.0963], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0135, 0.0139, 0.0118, 0.0126, 0.0137, 0.0138, 0.0160], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:58:16,305 INFO [finetune.py:976] (4/7) Epoch 26, batch 4450, loss[loss=0.1753, simple_loss=0.2454, pruned_loss=0.05261, over 4833.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2428, pruned_loss=0.04874, over 957546.00 frames. ], batch size: 33, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:58:21,318 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8658, 1.6754, 1.5088, 1.9545, 2.4629, 1.9787, 1.7936, 1.5039], + device='cuda:4'), covar=tensor([0.1951, 0.1913, 0.1859, 0.1534, 0.1449, 0.1148, 0.2113, 0.1785], + device='cuda:4'), in_proj_covar=tensor([0.0245, 0.0210, 0.0214, 0.0198, 0.0245, 0.0190, 0.0217, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:58:27,315 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147659.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:58:29,637 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7883, 1.8876, 1.5722, 1.5664, 2.1818, 2.2376, 1.9781, 1.8870], + device='cuda:4'), covar=tensor([0.0442, 0.0377, 0.0592, 0.0358, 0.0333, 0.0609, 0.0400, 0.0384], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0106, 0.0146, 0.0111, 0.0100, 0.0115, 0.0103, 0.0112], + device='cuda:4'), out_proj_covar=tensor([7.8126e-05, 8.1033e-05, 1.1382e-04, 8.4722e-05, 7.7814e-05, 8.4885e-05, + 7.6446e-05, 8.5333e-05], device='cuda:4') +2023-03-27 07:58:34,817 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147666.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:58:36,566 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.542e+02 1.864e+02 2.192e+02 3.736e+02, threshold=3.727e+02, percent-clipped=4.0 +2023-03-27 07:59:04,420 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4549, 3.3503, 3.1147, 1.4231, 3.4970, 2.5965, 0.7526, 2.2667], + device='cuda:4'), covar=tensor([0.2484, 0.2123, 0.1667, 0.3406, 0.1281, 0.1099, 0.4304, 0.1513], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0179, 0.0160, 0.0129, 0.0161, 0.0124, 0.0148, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 07:59:06,169 INFO [finetune.py:976] (4/7) Epoch 26, batch 4500, loss[loss=0.1824, simple_loss=0.2584, pruned_loss=0.05317, over 4887.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2446, pruned_loss=0.04961, over 956405.53 frames. ], batch size: 32, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 07:59:37,027 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147720.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 07:59:41,148 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0867, 1.9097, 1.6653, 1.8479, 1.8020, 1.8039, 1.8674, 2.5280], + device='cuda:4'), covar=tensor([0.3931, 0.3971, 0.3282, 0.3862, 0.4048, 0.2494, 0.3505, 0.1871], + device='cuda:4'), in_proj_covar=tensor([0.0290, 0.0265, 0.0237, 0.0277, 0.0260, 0.0230, 0.0258, 0.0239], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:59:52,140 INFO [finetune.py:976] (4/7) Epoch 26, batch 4550, loss[loss=0.1725, simple_loss=0.244, pruned_loss=0.05053, over 4795.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2463, pruned_loss=0.05033, over 955390.80 frames. ], batch size: 25, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 07:59:54,705 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7471, 1.0253, 1.8378, 1.7443, 1.5981, 1.5083, 1.6863, 1.7074], + device='cuda:4'), covar=tensor([0.3291, 0.3641, 0.2882, 0.3287, 0.4318, 0.3394, 0.3597, 0.2704], + device='cuda:4'), in_proj_covar=tensor([0.0266, 0.0247, 0.0267, 0.0295, 0.0295, 0.0270, 0.0301, 0.0251], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 07:59:59,896 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=2.01 vs. limit=2.0 +2023-03-27 08:00:04,827 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147763.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:00:09,335 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.722e+01 1.488e+02 1.756e+02 2.293e+02 4.562e+02, threshold=3.512e+02, percent-clipped=2.0 +2023-03-27 08:00:24,541 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=147790.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:00:25,690 INFO [finetune.py:976] (4/7) Epoch 26, batch 4600, loss[loss=0.137, simple_loss=0.2118, pruned_loss=0.0311, over 4786.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2463, pruned_loss=0.05046, over 957240.61 frames. ], batch size: 29, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:00:39,913 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147815.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:00:59,127 INFO [finetune.py:976] (4/7) Epoch 26, batch 4650, loss[loss=0.1502, simple_loss=0.2211, pruned_loss=0.03963, over 4750.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.243, pruned_loss=0.04922, over 956434.07 frames. ], batch size: 28, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:01:08,856 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147857.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:01:15,990 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 7.955e+01 1.458e+02 1.712e+02 2.175e+02 4.467e+02, threshold=3.424e+02, percent-clipped=3.0 +2023-03-27 08:01:21,844 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147876.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:01:39,630 INFO [finetune.py:976] (4/7) Epoch 26, batch 4700, loss[loss=0.1549, simple_loss=0.224, pruned_loss=0.04291, over 4744.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2393, pruned_loss=0.04803, over 956593.96 frames. ], batch size: 59, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:01:46,512 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147897.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:01:55,416 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=147911.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:01:59,713 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147918.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:02:16,231 INFO [finetune.py:976] (4/7) Epoch 26, batch 4750, loss[loss=0.2217, simple_loss=0.2817, pruned_loss=0.08087, over 4912.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2376, pruned_loss=0.04764, over 956811.91 frames. ], batch size: 36, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:02:18,627 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=147945.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:02:32,361 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=147966.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:02:34,089 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.372e+01 1.470e+02 1.654e+02 2.049e+02 2.990e+02, threshold=3.309e+02, percent-clipped=0.0 +2023-03-27 08:02:36,001 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=147972.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:02:41,816 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8780, 1.8651, 2.4714, 2.0501, 2.0304, 4.6494, 2.1167, 2.0641], + device='cuda:4'), covar=tensor([0.0875, 0.1598, 0.0973, 0.0885, 0.1453, 0.0110, 0.1210, 0.1616], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0081, 0.0072, 0.0076, 0.0090, 0.0080, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 08:02:50,081 INFO [finetune.py:976] (4/7) Epoch 26, batch 4800, loss[loss=0.1536, simple_loss=0.2388, pruned_loss=0.03414, over 4768.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2409, pruned_loss=0.04866, over 954724.31 frames. ], batch size: 28, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:03:04,513 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1575, 2.0994, 1.6542, 1.9358, 1.9801, 1.9772, 1.9796, 2.6698], + device='cuda:4'), covar=tensor([0.3923, 0.3642, 0.3323, 0.3988, 0.4115, 0.2468, 0.3843, 0.1842], + device='cuda:4'), in_proj_covar=tensor([0.0290, 0.0266, 0.0238, 0.0278, 0.0261, 0.0231, 0.0259, 0.0240], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:03:06,226 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148014.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:03:06,847 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148015.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:03:09,934 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2228, 2.2215, 1.8944, 2.2952, 2.0830, 2.1345, 2.0957, 2.8873], + device='cuda:4'), covar=tensor([0.3645, 0.4903, 0.3471, 0.4447, 0.4353, 0.2444, 0.4406, 0.1715], + device='cuda:4'), in_proj_covar=tensor([0.0290, 0.0265, 0.0238, 0.0278, 0.0261, 0.0231, 0.0259, 0.0240], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:03:17,888 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-27 08:03:24,535 INFO [finetune.py:976] (4/7) Epoch 26, batch 4850, loss[loss=0.2285, simple_loss=0.2896, pruned_loss=0.08374, over 4839.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2455, pruned_loss=0.04953, over 955687.40 frames. ], batch size: 49, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:03:39,259 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148063.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:03:42,782 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.530e+02 1.908e+02 2.333e+02 3.886e+02, threshold=3.817e+02, percent-clipped=4.0 +2023-03-27 08:04:03,598 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-03-27 08:04:04,142 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148090.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:04:05,254 INFO [finetune.py:976] (4/7) Epoch 26, batch 4900, loss[loss=0.1451, simple_loss=0.2291, pruned_loss=0.03049, over 4796.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.246, pruned_loss=0.04949, over 954127.52 frames. ], batch size: 29, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:04:16,932 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6749, 2.3527, 1.8947, 1.0017, 2.1144, 2.1703, 1.9956, 2.1917], + device='cuda:4'), covar=tensor([0.0756, 0.0713, 0.1269, 0.1772, 0.1109, 0.1764, 0.1821, 0.0705], + device='cuda:4'), in_proj_covar=tensor([0.0173, 0.0193, 0.0201, 0.0183, 0.0210, 0.0212, 0.0225, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:04:30,820 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148111.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:04:57,743 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148138.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:05:00,561 INFO [finetune.py:976] (4/7) Epoch 26, batch 4950, loss[loss=0.1627, simple_loss=0.2422, pruned_loss=0.04161, over 4881.00 frames. ], tot_loss[loss=0.172, simple_loss=0.246, pruned_loss=0.04899, over 954449.31 frames. ], batch size: 32, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:05:03,126 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0418, 1.8280, 1.5954, 1.6219, 1.6782, 1.6964, 1.7858, 2.4501], + device='cuda:4'), covar=tensor([0.3338, 0.3447, 0.2923, 0.3388, 0.3615, 0.2080, 0.3203, 0.1557], + device='cuda:4'), in_proj_covar=tensor([0.0288, 0.0264, 0.0236, 0.0276, 0.0259, 0.0229, 0.0257, 0.0239], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:05:18,896 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.030e+01 1.572e+02 1.871e+02 2.257e+02 5.603e+02, threshold=3.742e+02, percent-clipped=1.0 +2023-03-27 08:05:20,235 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148171.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:05:20,940 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5932, 0.7358, 1.7288, 1.6341, 1.5478, 1.4447, 1.5884, 1.6611], + device='cuda:4'), covar=tensor([0.3591, 0.3599, 0.2864, 0.3251, 0.4251, 0.3541, 0.3600, 0.2665], + device='cuda:4'), in_proj_covar=tensor([0.0265, 0.0246, 0.0267, 0.0294, 0.0294, 0.0270, 0.0300, 0.0251], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:05:23,364 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6854, 1.5273, 1.0756, 0.3092, 1.2957, 1.5589, 1.5264, 1.5292], + device='cuda:4'), covar=tensor([0.0867, 0.0814, 0.1345, 0.1921, 0.1383, 0.2308, 0.2214, 0.0751], + device='cuda:4'), in_proj_covar=tensor([0.0171, 0.0192, 0.0200, 0.0182, 0.0209, 0.0211, 0.0224, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:05:33,993 INFO [finetune.py:976] (4/7) Epoch 26, batch 5000, loss[loss=0.1605, simple_loss=0.2212, pruned_loss=0.04991, over 4905.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2442, pruned_loss=0.04881, over 954674.03 frames. ], batch size: 32, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:05:48,838 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148213.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:06:07,383 INFO [finetune.py:976] (4/7) Epoch 26, batch 5050, loss[loss=0.1342, simple_loss=0.2021, pruned_loss=0.03315, over 4098.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2422, pruned_loss=0.04847, over 955139.24 frames. ], batch size: 17, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:06:25,054 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148267.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:06:26,159 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.004e+02 1.490e+02 1.796e+02 2.082e+02 4.496e+02, threshold=3.592e+02, percent-clipped=3.0 +2023-03-27 08:06:37,579 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148287.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:06:40,483 INFO [finetune.py:976] (4/7) Epoch 26, batch 5100, loss[loss=0.1147, simple_loss=0.1846, pruned_loss=0.02237, over 4719.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2399, pruned_loss=0.04791, over 952872.51 frames. ], batch size: 23, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:06:54,199 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1654, 3.6187, 3.8221, 3.9539, 3.9364, 3.6495, 4.2239, 1.3603], + device='cuda:4'), covar=tensor([0.0732, 0.0896, 0.0858, 0.0982, 0.1153, 0.1518, 0.0738, 0.5478], + device='cuda:4'), in_proj_covar=tensor([0.0356, 0.0251, 0.0285, 0.0299, 0.0341, 0.0289, 0.0308, 0.0304], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:07:02,533 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148315.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:07:22,782 INFO [finetune.py:976] (4/7) Epoch 26, batch 5150, loss[loss=0.1612, simple_loss=0.2273, pruned_loss=0.04751, over 4709.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2395, pruned_loss=0.04833, over 952137.03 frames. ], batch size: 23, lr: 2.95e-03, grad_scale: 32.0 +2023-03-27 08:07:27,011 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148348.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:07:36,970 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148363.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:07:41,443 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.549e+02 1.828e+02 2.233e+02 5.689e+02, threshold=3.657e+02, percent-clipped=4.0 +2023-03-27 08:07:55,902 INFO [finetune.py:976] (4/7) Epoch 26, batch 5200, loss[loss=0.2151, simple_loss=0.2871, pruned_loss=0.07152, over 4741.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2442, pruned_loss=0.05018, over 951838.05 frames. ], batch size: 54, lr: 2.95e-03, grad_scale: 16.0 +2023-03-27 08:08:13,159 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-27 08:08:15,689 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-03-27 08:08:21,945 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148429.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:08:29,695 INFO [finetune.py:976] (4/7) Epoch 26, batch 5250, loss[loss=0.2337, simple_loss=0.2997, pruned_loss=0.0838, over 4190.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.245, pruned_loss=0.05028, over 951756.75 frames. ], batch size: 66, lr: 2.94e-03, grad_scale: 16.0 +2023-03-27 08:08:48,645 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.102e+02 1.514e+02 1.742e+02 2.193e+02 4.299e+02, threshold=3.484e+02, percent-clipped=1.0 +2023-03-27 08:08:49,333 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148471.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:09:02,321 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148490.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:09:03,426 INFO [finetune.py:976] (4/7) Epoch 26, batch 5300, loss[loss=0.188, simple_loss=0.2596, pruned_loss=0.05823, over 4699.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2457, pruned_loss=0.04989, over 952957.90 frames. ], batch size: 59, lr: 2.94e-03, grad_scale: 16.0 +2023-03-27 08:09:07,101 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148498.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:09:20,063 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148513.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:09:28,391 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148519.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:09:54,000 INFO [finetune.py:976] (4/7) Epoch 26, batch 5350, loss[loss=0.1649, simple_loss=0.228, pruned_loss=0.05092, over 4845.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2461, pruned_loss=0.04972, over 954303.29 frames. ], batch size: 49, lr: 2.94e-03, grad_scale: 16.0 +2023-03-27 08:10:12,427 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148559.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:10:13,592 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148561.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:10:17,756 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148567.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:10:19,467 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.029e+02 1.412e+02 1.653e+02 1.941e+02 3.220e+02, threshold=3.306e+02, percent-clipped=0.0 +2023-03-27 08:10:34,702 INFO [finetune.py:976] (4/7) Epoch 26, batch 5400, loss[loss=0.1411, simple_loss=0.2045, pruned_loss=0.03878, over 4811.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2434, pruned_loss=0.04932, over 953784.48 frames. ], batch size: 25, lr: 2.94e-03, grad_scale: 16.0 +2023-03-27 08:10:49,745 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148615.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:11:07,907 INFO [finetune.py:976] (4/7) Epoch 26, batch 5450, loss[loss=0.1606, simple_loss=0.2404, pruned_loss=0.04035, over 4855.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2413, pruned_loss=0.04907, over 955094.99 frames. ], batch size: 44, lr: 2.94e-03, grad_scale: 16.0 +2023-03-27 08:11:08,560 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148643.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:11:13,463 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148651.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:11:18,995 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1164, 2.0328, 1.7265, 1.8966, 2.1504, 1.8608, 2.2326, 2.1252], + device='cuda:4'), covar=tensor([0.1154, 0.1797, 0.2538, 0.2135, 0.2008, 0.1418, 0.2506, 0.1449], + device='cuda:4'), in_proj_covar=tensor([0.0188, 0.0189, 0.0237, 0.0253, 0.0248, 0.0206, 0.0214, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:11:25,813 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.571e+02 1.855e+02 2.174e+02 4.125e+02, threshold=3.711e+02, percent-clipped=2.0 +2023-03-27 08:11:30,032 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3862, 2.2554, 1.8038, 2.3738, 2.2917, 2.0469, 2.6199, 2.3656], + device='cuda:4'), covar=tensor([0.1195, 0.1986, 0.2815, 0.2277, 0.2414, 0.1602, 0.2618, 0.1457], + device='cuda:4'), in_proj_covar=tensor([0.0189, 0.0190, 0.0237, 0.0253, 0.0249, 0.0206, 0.0214, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:11:41,099 INFO [finetune.py:976] (4/7) Epoch 26, batch 5500, loss[loss=0.1456, simple_loss=0.2271, pruned_loss=0.03206, over 4738.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2395, pruned_loss=0.04899, over 955624.48 frames. ], batch size: 54, lr: 2.94e-03, grad_scale: 16.0 +2023-03-27 08:11:51,718 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-27 08:11:53,976 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148712.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:11:56,385 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=148716.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:12:24,628 INFO [finetune.py:976] (4/7) Epoch 26, batch 5550, loss[loss=0.1644, simple_loss=0.2424, pruned_loss=0.04316, over 4867.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2405, pruned_loss=0.0492, over 954690.79 frames. ], batch size: 31, lr: 2.94e-03, grad_scale: 16.0 +2023-03-27 08:12:42,332 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.664e+01 1.578e+02 1.917e+02 2.288e+02 4.413e+02, threshold=3.834e+02, percent-clipped=2.0 +2023-03-27 08:12:47,520 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=148777.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:12:52,459 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148785.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:12:56,502 INFO [finetune.py:976] (4/7) Epoch 26, batch 5600, loss[loss=0.2034, simple_loss=0.2744, pruned_loss=0.06616, over 4917.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2439, pruned_loss=0.04979, over 956198.15 frames. ], batch size: 36, lr: 2.94e-03, grad_scale: 16.0 +2023-03-27 08:12:57,925 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.94 vs. limit=5.0 +2023-03-27 08:13:25,694 INFO [finetune.py:976] (4/7) Epoch 26, batch 5650, loss[loss=0.1436, simple_loss=0.2096, pruned_loss=0.03882, over 4670.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2455, pruned_loss=0.05008, over 955404.95 frames. ], batch size: 23, lr: 2.94e-03, grad_scale: 16.0 +2023-03-27 08:13:32,825 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=148854.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:13:42,328 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.500e+02 1.770e+02 2.150e+02 4.859e+02, threshold=3.539e+02, percent-clipped=1.0 +2023-03-27 08:13:55,306 INFO [finetune.py:976] (4/7) Epoch 26, batch 5700, loss[loss=0.1684, simple_loss=0.2258, pruned_loss=0.05551, over 4268.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2411, pruned_loss=0.04896, over 935447.88 frames. ], batch size: 18, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:14:24,190 INFO [finetune.py:976] (4/7) Epoch 27, batch 0, loss[loss=0.184, simple_loss=0.2667, pruned_loss=0.05063, over 4888.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2667, pruned_loss=0.05063, over 4888.00 frames. ], batch size: 46, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:14:24,190 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 08:14:36,091 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6683, 1.2362, 0.8958, 1.6624, 2.1487, 1.1466, 1.5699, 1.5801], + device='cuda:4'), covar=tensor([0.1484, 0.1953, 0.1822, 0.1170, 0.1878, 0.2129, 0.1309, 0.2030], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0110, 0.0092, 0.0120, 0.0094, 0.0099, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 08:14:36,566 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5786, 1.4351, 1.3823, 1.4871, 1.7448, 1.7082, 1.4919, 1.3564], + device='cuda:4'), covar=tensor([0.0427, 0.0332, 0.0651, 0.0332, 0.0294, 0.0362, 0.0350, 0.0424], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0105, 0.0146, 0.0110, 0.0100, 0.0114, 0.0103, 0.0112], + device='cuda:4'), out_proj_covar=tensor([7.7685e-05, 8.0337e-05, 1.1349e-04, 8.4398e-05, 7.7821e-05, 8.4339e-05, + 7.6199e-05, 8.4954e-05], device='cuda:4') +2023-03-27 08:14:40,697 INFO [finetune.py:1010] (4/7) Epoch 27, validation: loss=0.1593, simple_loss=0.2269, pruned_loss=0.04586, over 2265189.00 frames. +2023-03-27 08:14:40,698 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6557MB +2023-03-27 08:14:57,045 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=148943.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:15:27,446 INFO [finetune.py:976] (4/7) Epoch 27, batch 50, loss[loss=0.1997, simple_loss=0.2801, pruned_loss=0.05961, over 4811.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2445, pruned_loss=0.04874, over 216128.59 frames. ], batch size: 38, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:15:28,075 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.253e+01 1.427e+02 1.731e+02 2.058e+02 3.661e+02, threshold=3.462e+02, percent-clipped=4.0 +2023-03-27 08:15:34,818 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-27 08:15:44,091 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=148991.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:15:53,997 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=149007.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:16:03,653 INFO [finetune.py:976] (4/7) Epoch 27, batch 100, loss[loss=0.1417, simple_loss=0.2095, pruned_loss=0.03693, over 4700.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2379, pruned_loss=0.04726, over 379742.40 frames. ], batch size: 23, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:16:36,428 INFO [finetune.py:976] (4/7) Epoch 27, batch 150, loss[loss=0.1317, simple_loss=0.2072, pruned_loss=0.02809, over 4862.00 frames. ], tot_loss[loss=0.1614, simple_loss=0.2326, pruned_loss=0.04514, over 508136.08 frames. ], batch size: 49, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:16:37,489 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.570e+01 1.451e+02 1.770e+02 2.054e+02 3.397e+02, threshold=3.539e+02, percent-clipped=0.0 +2023-03-27 08:16:39,141 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=149072.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:16:47,480 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=149085.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:16:54,785 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1345, 1.3172, 1.4663, 0.7345, 1.4289, 1.6535, 1.6664, 1.3997], + device='cuda:4'), covar=tensor([0.0872, 0.0657, 0.0513, 0.0479, 0.0452, 0.0521, 0.0304, 0.0691], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0149, 0.0130, 0.0124, 0.0131, 0.0131, 0.0143, 0.0150], + device='cuda:4'), out_proj_covar=tensor([8.9437e-05, 1.0712e-04, 9.2253e-05, 8.6978e-05, 9.1978e-05, 9.2468e-05, + 1.0158e-04, 1.0762e-04], device='cuda:4') +2023-03-27 08:16:57,495 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-27 08:16:58,390 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7746, 1.7961, 1.6888, 1.8530, 1.4780, 4.2262, 1.7116, 2.0819], + device='cuda:4'), covar=tensor([0.3362, 0.2325, 0.2092, 0.2342, 0.1627, 0.0170, 0.2634, 0.1130], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0117, 0.0121, 0.0124, 0.0113, 0.0095, 0.0094, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 08:17:09,497 INFO [finetune.py:976] (4/7) Epoch 27, batch 200, loss[loss=0.1839, simple_loss=0.2552, pruned_loss=0.05632, over 4872.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2338, pruned_loss=0.0463, over 607252.10 frames. ], batch size: 31, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:17:19,406 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=149133.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:17:39,176 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=149154.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:17:52,876 INFO [finetune.py:976] (4/7) Epoch 27, batch 250, loss[loss=0.1539, simple_loss=0.2275, pruned_loss=0.04015, over 4868.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2382, pruned_loss=0.04776, over 685663.95 frames. ], batch size: 34, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:17:53,480 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.507e+01 1.547e+02 1.763e+02 2.073e+02 3.560e+02, threshold=3.526e+02, percent-clipped=1.0 +2023-03-27 08:18:14,503 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.41 vs. limit=5.0 +2023-03-27 08:18:14,781 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=149202.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:18:25,586 INFO [finetune.py:976] (4/7) Epoch 27, batch 300, loss[loss=0.1544, simple_loss=0.2382, pruned_loss=0.03526, over 4823.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.243, pruned_loss=0.04918, over 746503.36 frames. ], batch size: 40, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:18:45,466 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8626, 2.0702, 1.6610, 1.7080, 2.3893, 2.3267, 1.9709, 1.9096], + device='cuda:4'), covar=tensor([0.0411, 0.0326, 0.0578, 0.0337, 0.0234, 0.0506, 0.0398, 0.0367], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0107, 0.0147, 0.0112, 0.0102, 0.0116, 0.0105, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.8546e-05, 8.1570e-05, 1.1494e-04, 8.5492e-05, 7.8742e-05, 8.5252e-05, + 7.7678e-05, 8.5762e-05], device='cuda:4') +2023-03-27 08:18:46,656 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1201, 1.2333, 1.3732, 1.2224, 1.3057, 2.4551, 1.2011, 1.3446], + device='cuda:4'), covar=tensor([0.1026, 0.1929, 0.1083, 0.0966, 0.1778, 0.0359, 0.1623, 0.1898], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 08:18:55,227 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.6763, 4.9417, 5.2066, 5.3470, 5.3690, 5.1351, 5.7180, 2.1988], + device='cuda:4'), covar=tensor([0.0612, 0.0787, 0.0668, 0.0908, 0.0945, 0.1433, 0.0481, 0.4974], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0248, 0.0284, 0.0296, 0.0336, 0.0288, 0.0305, 0.0299], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:18:58,798 INFO [finetune.py:976] (4/7) Epoch 27, batch 350, loss[loss=0.1397, simple_loss=0.2117, pruned_loss=0.03384, over 4792.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2444, pruned_loss=0.04945, over 791940.01 frames. ], batch size: 29, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:18:59,398 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.129e+02 1.601e+02 1.876e+02 2.140e+02 5.128e+02, threshold=3.753e+02, percent-clipped=1.0 +2023-03-27 08:19:14,647 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0410, 2.0142, 1.6220, 1.7709, 1.8452, 1.8186, 1.9176, 2.5197], + device='cuda:4'), covar=tensor([0.3594, 0.3438, 0.3134, 0.3488, 0.3697, 0.2246, 0.3259, 0.1633], + device='cuda:4'), in_proj_covar=tensor([0.0288, 0.0263, 0.0235, 0.0275, 0.0259, 0.0229, 0.0257, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:19:24,863 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=149307.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:19:32,721 INFO [finetune.py:976] (4/7) Epoch 27, batch 400, loss[loss=0.178, simple_loss=0.2513, pruned_loss=0.05238, over 4888.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2455, pruned_loss=0.0496, over 828333.68 frames. ], batch size: 35, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:19:34,681 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.2158, 1.3192, 1.4258, 0.7920, 1.4015, 1.5463, 1.6604, 1.3822], + device='cuda:4'), covar=tensor([0.0861, 0.0665, 0.0649, 0.0502, 0.0567, 0.0736, 0.0406, 0.0690], + device='cuda:4'), in_proj_covar=tensor([0.0121, 0.0147, 0.0128, 0.0122, 0.0130, 0.0129, 0.0141, 0.0149], + device='cuda:4'), out_proj_covar=tensor([8.7996e-05, 1.0552e-04, 9.1176e-05, 8.5589e-05, 9.0718e-05, 9.1437e-05, + 1.0023e-04, 1.0660e-04], device='cuda:4') +2023-03-27 08:19:35,665 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=5.03 vs. limit=5.0 +2023-03-27 08:20:07,341 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=149355.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:20:10,499 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9304, 1.7689, 1.5258, 1.3345, 1.7223, 1.7276, 1.7107, 2.2672], + device='cuda:4'), covar=tensor([0.3830, 0.4027, 0.3395, 0.3709, 0.3593, 0.2406, 0.3641, 0.1841], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0263, 0.0235, 0.0275, 0.0259, 0.0229, 0.0258, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:20:16,475 INFO [finetune.py:976] (4/7) Epoch 27, batch 450, loss[loss=0.1626, simple_loss=0.2203, pruned_loss=0.0524, over 4902.00 frames. ], tot_loss[loss=0.171, simple_loss=0.244, pruned_loss=0.04902, over 857200.24 frames. ], batch size: 32, lr: 2.94e-03, grad_scale: 8.0 +2023-03-27 08:20:17,065 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.956e+01 1.496e+02 1.736e+02 2.126e+02 4.914e+02, threshold=3.471e+02, percent-clipped=1.0 +2023-03-27 08:20:17,774 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=149372.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:20:57,079 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-27 08:21:04,695 INFO [finetune.py:976] (4/7) Epoch 27, batch 500, loss[loss=0.1609, simple_loss=0.2288, pruned_loss=0.0465, over 4874.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2416, pruned_loss=0.04872, over 877588.40 frames. ], batch size: 34, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:21:04,764 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=149420.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:21:38,464 INFO [finetune.py:976] (4/7) Epoch 27, batch 550, loss[loss=0.1701, simple_loss=0.238, pruned_loss=0.05107, over 4759.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2388, pruned_loss=0.04816, over 895035.19 frames. ], batch size: 59, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:21:39,063 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.918e+01 1.466e+02 1.717e+02 2.125e+02 3.295e+02, threshold=3.435e+02, percent-clipped=0.0 +2023-03-27 08:22:02,498 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.2537, 2.9003, 3.0327, 3.1796, 3.0309, 2.8317, 3.3101, 0.9204], + device='cuda:4'), covar=tensor([0.1163, 0.1116, 0.1162, 0.1258, 0.1677, 0.2053, 0.1212, 0.6252], + device='cuda:4'), in_proj_covar=tensor([0.0352, 0.0249, 0.0285, 0.0297, 0.0337, 0.0289, 0.0306, 0.0301], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:22:03,746 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4150, 1.4332, 1.8779, 1.7073, 1.5418, 3.3600, 1.3980, 1.5641], + device='cuda:4'), covar=tensor([0.1080, 0.1919, 0.1101, 0.0979, 0.1676, 0.0235, 0.1535, 0.1826], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 08:22:08,656 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=149514.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:22:12,133 INFO [finetune.py:976] (4/7) Epoch 27, batch 600, loss[loss=0.1757, simple_loss=0.2381, pruned_loss=0.05667, over 4907.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2395, pruned_loss=0.04869, over 909575.27 frames. ], batch size: 37, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:22:45,137 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=149565.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:22:48,086 INFO [finetune.py:976] (4/7) Epoch 27, batch 650, loss[loss=0.1894, simple_loss=0.2632, pruned_loss=0.05778, over 4845.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2421, pruned_loss=0.04902, over 917231.85 frames. ], batch size: 49, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:22:53,196 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.592e+02 1.975e+02 2.434e+02 4.045e+02, threshold=3.949e+02, percent-clipped=4.0 +2023-03-27 08:22:55,787 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=149575.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:23:29,820 INFO [finetune.py:976] (4/7) Epoch 27, batch 700, loss[loss=0.2219, simple_loss=0.2932, pruned_loss=0.07533, over 4809.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.244, pruned_loss=0.04942, over 927428.17 frames. ], batch size: 40, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:23:33,637 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=149626.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 08:23:41,651 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.57 vs. limit=5.0 +2023-03-27 08:23:59,608 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3676, 2.3509, 2.0260, 2.5178, 2.3477, 2.3040, 2.3033, 3.1351], + device='cuda:4'), covar=tensor([0.3731, 0.4845, 0.3404, 0.4028, 0.4331, 0.2481, 0.4128, 0.1666], + device='cuda:4'), in_proj_covar=tensor([0.0290, 0.0265, 0.0237, 0.0277, 0.0261, 0.0231, 0.0260, 0.0241], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:24:03,080 INFO [finetune.py:976] (4/7) Epoch 27, batch 750, loss[loss=0.2074, simple_loss=0.2672, pruned_loss=0.07385, over 4895.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2445, pruned_loss=0.04916, over 935463.83 frames. ], batch size: 37, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:24:03,699 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.520e+02 1.783e+02 2.094e+02 3.998e+02, threshold=3.567e+02, percent-clipped=1.0 +2023-03-27 08:24:36,878 INFO [finetune.py:976] (4/7) Epoch 27, batch 800, loss[loss=0.1514, simple_loss=0.2223, pruned_loss=0.04018, over 4852.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2448, pruned_loss=0.04959, over 939373.36 frames. ], batch size: 49, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:24:54,917 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=149746.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 08:25:20,610 INFO [finetune.py:976] (4/7) Epoch 27, batch 850, loss[loss=0.1558, simple_loss=0.2271, pruned_loss=0.04227, over 4747.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2429, pruned_loss=0.04903, over 943200.13 frames. ], batch size: 54, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:25:21,212 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.021e+02 1.421e+02 1.714e+02 1.950e+02 4.580e+02, threshold=3.429e+02, percent-clipped=2.0 +2023-03-27 08:25:29,323 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.72 vs. limit=2.0 +2023-03-27 08:25:46,957 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3964, 1.5877, 1.2181, 1.4724, 1.6834, 1.7246, 1.4860, 1.4545], + device='cuda:4'), covar=tensor([0.0446, 0.0265, 0.0549, 0.0301, 0.0280, 0.0375, 0.0343, 0.0367], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0105, 0.0146, 0.0111, 0.0101, 0.0115, 0.0103, 0.0112], + device='cuda:4'), out_proj_covar=tensor([7.7680e-05, 8.0618e-05, 1.1389e-04, 8.4658e-05, 7.8134e-05, 8.4801e-05, + 7.6603e-05, 8.5240e-05], device='cuda:4') +2023-03-27 08:25:56,771 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=149807.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 08:26:09,033 INFO [finetune.py:976] (4/7) Epoch 27, batch 900, loss[loss=0.1232, simple_loss=0.1889, pruned_loss=0.0287, over 4745.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2398, pruned_loss=0.0483, over 945122.78 frames. ], batch size: 54, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:26:22,568 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1286, 1.6920, 2.2168, 2.2121, 1.9291, 1.9036, 2.1012, 2.0400], + device='cuda:4'), covar=tensor([0.3845, 0.3883, 0.3307, 0.3508, 0.5188, 0.4113, 0.4436, 0.3120], + device='cuda:4'), in_proj_covar=tensor([0.0268, 0.0249, 0.0269, 0.0297, 0.0297, 0.0273, 0.0302, 0.0254], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:26:42,230 INFO [finetune.py:976] (4/7) Epoch 27, batch 950, loss[loss=0.2051, simple_loss=0.278, pruned_loss=0.06614, over 4827.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2389, pruned_loss=0.04841, over 946743.66 frames. ], batch size: 41, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:26:42,298 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=149870.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:26:42,815 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.619e+01 1.503e+02 1.866e+02 2.296e+02 3.689e+02, threshold=3.732e+02, percent-clipped=3.0 +2023-03-27 08:27:15,532 INFO [finetune.py:976] (4/7) Epoch 27, batch 1000, loss[loss=0.1558, simple_loss=0.2316, pruned_loss=0.03996, over 4752.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2407, pruned_loss=0.04877, over 950051.32 frames. ], batch size: 27, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:27:16,181 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=149921.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 08:27:27,177 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5447, 1.4078, 2.0136, 3.1137, 2.1446, 2.2395, 0.9004, 2.6910], + device='cuda:4'), covar=tensor([0.1694, 0.1413, 0.1255, 0.0553, 0.0805, 0.1552, 0.1834, 0.0457], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0163, 0.0101, 0.0136, 0.0124, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 08:27:37,066 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-27 08:27:48,832 INFO [finetune.py:976] (4/7) Epoch 27, batch 1050, loss[loss=0.2097, simple_loss=0.2768, pruned_loss=0.07127, over 4868.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2436, pruned_loss=0.04928, over 952191.00 frames. ], batch size: 31, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:27:49,414 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.561e+02 1.767e+02 2.240e+02 3.870e+02, threshold=3.534e+02, percent-clipped=1.0 +2023-03-27 08:27:52,452 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.3807, 4.7341, 4.9925, 5.1820, 5.1463, 4.8496, 5.5179, 1.7697], + device='cuda:4'), covar=tensor([0.0762, 0.0806, 0.0741, 0.1011, 0.1197, 0.1552, 0.0568, 0.6086], + device='cuda:4'), in_proj_covar=tensor([0.0353, 0.0249, 0.0284, 0.0297, 0.0336, 0.0289, 0.0306, 0.0303], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:28:22,091 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.5379, 1.5666, 1.4713, 0.8279, 1.6179, 1.8394, 1.8488, 1.4078], + device='cuda:4'), covar=tensor([0.0828, 0.0581, 0.0557, 0.0547, 0.0406, 0.0560, 0.0284, 0.0699], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0147, 0.0128, 0.0123, 0.0130, 0.0130, 0.0142, 0.0150], + device='cuda:4'), out_proj_covar=tensor([8.8584e-05, 1.0592e-04, 9.1402e-05, 8.6310e-05, 9.1337e-05, 9.1946e-05, + 1.0105e-04, 1.0725e-04], device='cuda:4') +2023-03-27 08:28:33,250 INFO [finetune.py:976] (4/7) Epoch 27, batch 1100, loss[loss=0.1876, simple_loss=0.2596, pruned_loss=0.05783, over 4852.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.245, pruned_loss=0.04892, over 953637.47 frames. ], batch size: 31, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:28:38,601 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6433, 1.2509, 0.7310, 1.5508, 2.1348, 1.3918, 1.4362, 1.5465], + device='cuda:4'), covar=tensor([0.1559, 0.2168, 0.2069, 0.1291, 0.1774, 0.1896, 0.1542, 0.2126], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0093, 0.0109, 0.0091, 0.0119, 0.0092, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 08:28:51,981 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5769, 1.5920, 1.8818, 1.8115, 1.7191, 3.3505, 1.4556, 1.5652], + device='cuda:4'), covar=tensor([0.1040, 0.1757, 0.1193, 0.0944, 0.1446, 0.0230, 0.1519, 0.1801], + device='cuda:4'), in_proj_covar=tensor([0.0074, 0.0081, 0.0073, 0.0076, 0.0090, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 08:28:52,313 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-27 08:29:06,479 INFO [finetune.py:976] (4/7) Epoch 27, batch 1150, loss[loss=0.1938, simple_loss=0.2652, pruned_loss=0.06125, over 4822.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2465, pruned_loss=0.04993, over 953270.80 frames. ], batch size: 33, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:29:07,080 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.913e+01 1.470e+02 1.766e+02 2.217e+02 3.439e+02, threshold=3.531e+02, percent-clipped=0.0 +2023-03-27 08:29:13,045 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150079.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:29:26,985 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=150102.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 08:29:39,274 INFO [finetune.py:976] (4/7) Epoch 27, batch 1200, loss[loss=0.1806, simple_loss=0.2514, pruned_loss=0.05491, over 4901.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2455, pruned_loss=0.0499, over 954388.35 frames. ], batch size: 35, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:29:52,974 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8860, 1.7121, 1.6025, 1.6155, 2.0572, 2.1605, 1.7876, 1.6370], + device='cuda:4'), covar=tensor([0.0448, 0.0396, 0.0553, 0.0370, 0.0284, 0.0466, 0.0400, 0.0451], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0106, 0.0147, 0.0111, 0.0101, 0.0116, 0.0104, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.7952e-05, 8.1007e-05, 1.1431e-04, 8.5052e-05, 7.8425e-05, 8.5461e-05, + 7.7114e-05, 8.5893e-05], device='cuda:4') +2023-03-27 08:29:52,981 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150140.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:30:14,458 INFO [finetune.py:976] (4/7) Epoch 27, batch 1250, loss[loss=0.1532, simple_loss=0.224, pruned_loss=0.04124, over 4799.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2422, pruned_loss=0.04884, over 954168.51 frames. ], batch size: 29, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:30:15,046 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150170.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:30:15,537 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.012e+02 1.554e+02 1.886e+02 2.235e+02 6.588e+02, threshold=3.772e+02, percent-clipped=2.0 +2023-03-27 08:30:55,492 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-27 08:30:56,391 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=150218.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:30:57,559 INFO [finetune.py:976] (4/7) Epoch 27, batch 1300, loss[loss=0.1321, simple_loss=0.2087, pruned_loss=0.02773, over 4795.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2394, pruned_loss=0.04773, over 955624.61 frames. ], batch size: 25, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:31:02,865 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150221.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:31:23,041 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2234, 1.9945, 1.8192, 2.1497, 2.7066, 2.2469, 2.1964, 1.6890], + device='cuda:4'), covar=tensor([0.1998, 0.1877, 0.1821, 0.1599, 0.1605, 0.1105, 0.1934, 0.1792], + device='cuda:4'), in_proj_covar=tensor([0.0243, 0.0210, 0.0213, 0.0198, 0.0244, 0.0190, 0.0215, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:31:30,846 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0221, 1.3478, 1.4365, 1.2684, 1.4249, 2.4308, 1.2321, 1.4340], + device='cuda:4'), covar=tensor([0.1155, 0.1925, 0.1046, 0.1000, 0.1715, 0.0398, 0.1642, 0.1955], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0090, 0.0080, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 08:31:42,216 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=150269.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:31:42,769 INFO [finetune.py:976] (4/7) Epoch 27, batch 1350, loss[loss=0.1837, simple_loss=0.2537, pruned_loss=0.05681, over 4925.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2392, pruned_loss=0.04771, over 956314.87 frames. ], batch size: 36, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:31:43,343 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.453e+02 1.768e+02 2.125e+02 3.830e+02, threshold=3.537e+02, percent-clipped=1.0 +2023-03-27 08:31:45,180 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.2396, 2.8994, 2.7187, 1.4004, 2.8008, 2.3097, 2.3190, 2.7710], + device='cuda:4'), covar=tensor([0.0859, 0.0704, 0.1627, 0.2177, 0.1501, 0.2103, 0.2023, 0.0953], + device='cuda:4'), in_proj_covar=tensor([0.0172, 0.0192, 0.0200, 0.0182, 0.0209, 0.0211, 0.0224, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:31:45,780 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150274.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:31:49,887 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-27 08:32:03,989 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0631, 1.9895, 1.6119, 1.7495, 1.8550, 1.8687, 1.9295, 2.5312], + device='cuda:4'), covar=tensor([0.3258, 0.3166, 0.2975, 0.3362, 0.3813, 0.2094, 0.3201, 0.1461], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0263, 0.0236, 0.0277, 0.0259, 0.0229, 0.0258, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:32:16,594 INFO [finetune.py:976] (4/7) Epoch 27, batch 1400, loss[loss=0.1881, simple_loss=0.2607, pruned_loss=0.05773, over 4819.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2428, pruned_loss=0.04912, over 956707.36 frames. ], batch size: 39, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:32:28,272 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150335.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:32:37,812 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5369, 1.4849, 2.1177, 1.7993, 1.8148, 4.0528, 1.4144, 1.6563], + device='cuda:4'), covar=tensor([0.0997, 0.1884, 0.1358, 0.1009, 0.1634, 0.0172, 0.1643, 0.1929], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 08:32:49,841 INFO [finetune.py:976] (4/7) Epoch 27, batch 1450, loss[loss=0.1934, simple_loss=0.2616, pruned_loss=0.06258, over 4915.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.244, pruned_loss=0.04922, over 953922.18 frames. ], batch size: 36, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:32:50,438 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.045e+02 1.587e+02 1.925e+02 2.309e+02 4.827e+02, threshold=3.851e+02, percent-clipped=3.0 +2023-03-27 08:32:59,292 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3501, 2.0146, 2.7038, 1.7586, 2.4180, 2.5033, 1.8544, 2.6959], + device='cuda:4'), covar=tensor([0.1379, 0.1981, 0.1488, 0.2138, 0.0907, 0.1712, 0.2789, 0.0878], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0208, 0.0194, 0.0191, 0.0176, 0.0214, 0.0219, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:33:11,846 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150402.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 08:33:24,832 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7773, 1.3547, 0.8499, 1.6221, 2.2055, 1.3393, 1.6513, 1.6155], + device='cuda:4'), covar=tensor([0.1410, 0.1923, 0.1757, 0.1098, 0.1774, 0.1809, 0.1267, 0.1848], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0093, 0.0109, 0.0091, 0.0119, 0.0092, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 08:33:26,914 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-27 08:33:29,004 INFO [finetune.py:976] (4/7) Epoch 27, batch 1500, loss[loss=0.1346, simple_loss=0.2121, pruned_loss=0.02854, over 4781.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2446, pruned_loss=0.04938, over 953324.07 frames. ], batch size: 29, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:33:42,988 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=150435.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:33:53,601 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=150450.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 08:33:56,685 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-27 08:34:01,432 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6092, 1.5699, 1.5137, 1.6400, 1.0744, 3.5160, 1.2646, 1.8620], + device='cuda:4'), covar=tensor([0.3250, 0.2553, 0.2130, 0.2306, 0.1801, 0.0194, 0.2569, 0.1135], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0116, 0.0120, 0.0124, 0.0113, 0.0095, 0.0094, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 08:34:05,554 INFO [finetune.py:976] (4/7) Epoch 27, batch 1550, loss[loss=0.1771, simple_loss=0.249, pruned_loss=0.05255, over 4687.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2446, pruned_loss=0.04941, over 952379.17 frames. ], batch size: 23, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:34:06,131 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.290e+01 1.580e+02 1.863e+02 2.206e+02 4.598e+02, threshold=3.727e+02, percent-clipped=2.0 +2023-03-27 08:34:20,902 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2533, 1.1699, 0.9887, 1.1543, 1.5454, 1.4016, 1.2456, 1.0485], + device='cuda:4'), covar=tensor([0.0388, 0.0351, 0.0851, 0.0388, 0.0244, 0.0578, 0.0331, 0.0505], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0106, 0.0147, 0.0111, 0.0101, 0.0116, 0.0104, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.8210e-05, 8.1286e-05, 1.1441e-04, 8.5169e-05, 7.8454e-05, 8.5394e-05, + 7.7433e-05, 8.6143e-05], device='cuda:4') +2023-03-27 08:34:27,930 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4351, 1.3219, 1.3102, 1.4061, 1.6525, 1.6133, 1.4148, 1.2280], + device='cuda:4'), covar=tensor([0.0412, 0.0343, 0.0651, 0.0323, 0.0267, 0.0489, 0.0354, 0.0462], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0106, 0.0147, 0.0111, 0.0101, 0.0116, 0.0104, 0.0114], + device='cuda:4'), out_proj_covar=tensor([7.8252e-05, 8.1353e-05, 1.1445e-04, 8.5237e-05, 7.8499e-05, 8.5498e-05, + 7.7505e-05, 8.6270e-05], device='cuda:4') +2023-03-27 08:34:30,944 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7127, 1.3089, 0.7747, 1.5562, 2.1082, 1.3022, 1.6025, 1.5917], + device='cuda:4'), covar=tensor([0.1543, 0.2150, 0.1942, 0.1242, 0.1980, 0.1881, 0.1388, 0.1984], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0093, 0.0109, 0.0091, 0.0119, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 08:34:38,731 INFO [finetune.py:976] (4/7) Epoch 27, batch 1600, loss[loss=0.1951, simple_loss=0.253, pruned_loss=0.06858, over 4851.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2418, pruned_loss=0.04813, over 952396.17 frames. ], batch size: 47, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:34:50,051 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150537.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:34:53,148 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-03-27 08:35:02,220 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.46 vs. limit=5.0 +2023-03-27 08:35:11,519 INFO [finetune.py:976] (4/7) Epoch 27, batch 1650, loss[loss=0.1354, simple_loss=0.213, pruned_loss=0.0289, over 4773.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2394, pruned_loss=0.04736, over 953259.60 frames. ], batch size: 28, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:35:12,132 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.130e+02 1.535e+02 1.741e+02 2.182e+02 5.670e+02, threshold=3.482e+02, percent-clipped=1.0 +2023-03-27 08:35:37,680 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150598.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:35:54,939 INFO [finetune.py:976] (4/7) Epoch 27, batch 1700, loss[loss=0.132, simple_loss=0.207, pruned_loss=0.02854, over 4768.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2376, pruned_loss=0.04699, over 954959.63 frames. ], batch size: 28, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:36:01,036 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=150630.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:36:02,299 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0403, 1.7454, 2.0953, 2.0953, 1.7879, 1.8283, 2.0067, 1.9429], + device='cuda:4'), covar=tensor([0.4175, 0.4138, 0.3206, 0.3915, 0.5417, 0.4290, 0.4848, 0.2991], + device='cuda:4'), in_proj_covar=tensor([0.0265, 0.0247, 0.0267, 0.0295, 0.0295, 0.0272, 0.0301, 0.0252], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:36:41,945 INFO [finetune.py:976] (4/7) Epoch 27, batch 1750, loss[loss=0.1363, simple_loss=0.1956, pruned_loss=0.03846, over 4768.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2399, pruned_loss=0.04785, over 953793.28 frames. ], batch size: 26, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:36:42,540 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.512e+01 1.530e+02 1.821e+02 2.198e+02 3.521e+02, threshold=3.642e+02, percent-clipped=1.0 +2023-03-27 08:37:15,428 INFO [finetune.py:976] (4/7) Epoch 27, batch 1800, loss[loss=0.2124, simple_loss=0.2808, pruned_loss=0.07197, over 4722.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2435, pruned_loss=0.04873, over 954617.83 frames. ], batch size: 59, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:37:27,765 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150732.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:37:33,303 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150735.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:37:57,191 INFO [finetune.py:976] (4/7) Epoch 27, batch 1850, loss[loss=0.1293, simple_loss=0.2101, pruned_loss=0.02429, over 4736.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2455, pruned_loss=0.04984, over 954665.61 frames. ], batch size: 27, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:37:57,786 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.063e+02 1.537e+02 1.800e+02 2.248e+02 4.542e+02, threshold=3.600e+02, percent-clipped=6.0 +2023-03-27 08:38:05,039 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=150783.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:38:11,138 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150793.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 08:38:30,230 INFO [finetune.py:976] (4/7) Epoch 27, batch 1900, loss[loss=0.1635, simple_loss=0.2332, pruned_loss=0.04691, over 4718.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2466, pruned_loss=0.05015, over 955692.89 frames. ], batch size: 59, lr: 2.93e-03, grad_scale: 8.0 +2023-03-27 08:39:02,353 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.2726, 2.9041, 3.0687, 3.2204, 3.0815, 2.8407, 3.3104, 0.9681], + device='cuda:4'), covar=tensor([0.1132, 0.1054, 0.1097, 0.1087, 0.1486, 0.1943, 0.1220, 0.5545], + device='cuda:4'), in_proj_covar=tensor([0.0355, 0.0251, 0.0285, 0.0299, 0.0338, 0.0290, 0.0309, 0.0305], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:39:14,069 INFO [finetune.py:976] (4/7) Epoch 27, batch 1950, loss[loss=0.1739, simple_loss=0.2378, pruned_loss=0.05503, over 4892.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.247, pruned_loss=0.05037, over 957232.56 frames. ], batch size: 32, lr: 2.92e-03, grad_scale: 8.0 +2023-03-27 08:39:14,651 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.033e+02 1.460e+02 1.651e+02 1.933e+02 3.642e+02, threshold=3.302e+02, percent-clipped=1.0 +2023-03-27 08:39:28,774 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=150893.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:39:31,243 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6923, 3.7758, 3.6193, 1.9893, 3.8693, 2.9246, 0.8571, 2.7278], + device='cuda:4'), covar=tensor([0.2502, 0.1997, 0.1548, 0.3205, 0.0914, 0.1030, 0.4457, 0.1507], + device='cuda:4'), in_proj_covar=tensor([0.0149, 0.0179, 0.0160, 0.0129, 0.0160, 0.0124, 0.0148, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 08:39:33,113 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=150900.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:39:39,479 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-27 08:39:47,875 INFO [finetune.py:976] (4/7) Epoch 27, batch 2000, loss[loss=0.1716, simple_loss=0.2387, pruned_loss=0.05221, over 4930.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2451, pruned_loss=0.05036, over 955598.40 frames. ], batch size: 33, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:39:54,517 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=150930.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:40:15,357 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=150961.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:40:21,603 INFO [finetune.py:976] (4/7) Epoch 27, batch 2050, loss[loss=0.1131, simple_loss=0.1891, pruned_loss=0.01851, over 4818.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2412, pruned_loss=0.04873, over 957367.65 frames. ], batch size: 25, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:40:22,191 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.304e+01 1.432e+02 1.658e+02 2.071e+02 3.830e+02, threshold=3.317e+02, percent-clipped=1.0 +2023-03-27 08:40:27,052 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=150978.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:40:56,344 INFO [finetune.py:976] (4/7) Epoch 27, batch 2100, loss[loss=0.1864, simple_loss=0.2514, pruned_loss=0.0607, over 4828.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2405, pruned_loss=0.04877, over 956331.41 frames. ], batch size: 33, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:41:15,833 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.57 vs. limit=2.0 +2023-03-27 08:41:17,055 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-27 08:41:47,143 INFO [finetune.py:976] (4/7) Epoch 27, batch 2150, loss[loss=0.214, simple_loss=0.2889, pruned_loss=0.06954, over 4898.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2439, pruned_loss=0.0501, over 956193.25 frames. ], batch size: 37, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:41:48,291 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.039e+02 1.525e+02 1.813e+02 2.166e+02 3.448e+02, threshold=3.626e+02, percent-clipped=1.0 +2023-03-27 08:41:52,753 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1825, 1.8127, 1.8243, 0.9196, 2.0353, 2.2865, 2.0595, 1.7228], + device='cuda:4'), covar=tensor([0.1080, 0.0764, 0.0583, 0.0717, 0.0649, 0.0795, 0.0442, 0.0786], + device='cuda:4'), in_proj_covar=tensor([0.0121, 0.0147, 0.0128, 0.0122, 0.0131, 0.0129, 0.0141, 0.0149], + device='cuda:4'), out_proj_covar=tensor([8.8389e-05, 1.0519e-04, 9.1050e-05, 8.5936e-05, 9.1479e-05, 9.1523e-05, + 1.0022e-04, 1.0698e-04], device='cuda:4') +2023-03-27 08:42:02,497 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151088.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 08:42:23,618 INFO [finetune.py:976] (4/7) Epoch 27, batch 2200, loss[loss=0.1671, simple_loss=0.24, pruned_loss=0.0471, over 4833.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2466, pruned_loss=0.05066, over 956663.37 frames. ], batch size: 49, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:43:04,289 INFO [finetune.py:976] (4/7) Epoch 27, batch 2250, loss[loss=0.1894, simple_loss=0.2671, pruned_loss=0.05586, over 4862.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2466, pruned_loss=0.05043, over 957252.86 frames. ], batch size: 31, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:43:04,891 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.120e+01 1.457e+02 1.754e+02 2.221e+02 3.820e+02, threshold=3.509e+02, percent-clipped=1.0 +2023-03-27 08:43:05,512 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3027, 2.0150, 1.5215, 0.6507, 1.7056, 1.9637, 1.8522, 1.9410], + device='cuda:4'), covar=tensor([0.0856, 0.0824, 0.1402, 0.2051, 0.1376, 0.2380, 0.2293, 0.0826], + device='cuda:4'), in_proj_covar=tensor([0.0172, 0.0192, 0.0201, 0.0182, 0.0209, 0.0211, 0.0224, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:43:14,788 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151184.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:43:20,808 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151193.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:43:37,566 INFO [finetune.py:976] (4/7) Epoch 27, batch 2300, loss[loss=0.1425, simple_loss=0.2046, pruned_loss=0.04019, over 4751.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2463, pruned_loss=0.04978, over 954962.49 frames. ], batch size: 28, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:43:51,733 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151241.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:43:54,688 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151245.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:44:06,984 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151256.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:44:18,928 INFO [finetune.py:976] (4/7) Epoch 27, batch 2350, loss[loss=0.1881, simple_loss=0.2578, pruned_loss=0.05922, over 4736.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2452, pruned_loss=0.04971, over 955017.04 frames. ], batch size: 54, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:44:19,967 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 7.208e+01 1.503e+02 1.827e+02 2.189e+02 3.264e+02, threshold=3.653e+02, percent-clipped=0.0 +2023-03-27 08:44:21,311 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9924, 2.4594, 2.4637, 1.2920, 2.7261, 2.1801, 1.9263, 2.3735], + device='cuda:4'), covar=tensor([0.1169, 0.1221, 0.2027, 0.2415, 0.1701, 0.2222, 0.2497, 0.1372], + device='cuda:4'), in_proj_covar=tensor([0.0172, 0.0192, 0.0202, 0.0183, 0.0210, 0.0212, 0.0225, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:44:28,775 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151282.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:44:39,442 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151298.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:44:52,610 INFO [finetune.py:976] (4/7) Epoch 27, batch 2400, loss[loss=0.1412, simple_loss=0.2134, pruned_loss=0.03445, over 4766.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2423, pruned_loss=0.04894, over 954957.95 frames. ], batch size: 23, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:45:09,230 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151343.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:45:19,411 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151359.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:45:26,016 INFO [finetune.py:976] (4/7) Epoch 27, batch 2450, loss[loss=0.1888, simple_loss=0.2625, pruned_loss=0.05758, over 4781.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2399, pruned_loss=0.04835, over 955328.67 frames. ], batch size: 59, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:45:26,602 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.415e+02 1.689e+02 1.968e+02 4.441e+02, threshold=3.378e+02, percent-clipped=1.0 +2023-03-27 08:45:27,930 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6929, 1.6164, 1.5616, 1.6513, 1.0865, 3.5996, 1.3396, 1.8109], + device='cuda:4'), covar=tensor([0.3270, 0.2540, 0.2258, 0.2482, 0.1867, 0.0181, 0.2687, 0.1272], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0124, 0.0113, 0.0096, 0.0094, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 08:45:38,467 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151388.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:45:49,374 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-27 08:45:58,929 INFO [finetune.py:976] (4/7) Epoch 27, batch 2500, loss[loss=0.1729, simple_loss=0.2414, pruned_loss=0.0522, over 4875.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2406, pruned_loss=0.04868, over 956442.08 frames. ], batch size: 34, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:46:02,015 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0217, 1.9308, 1.5984, 0.8939, 1.7459, 1.6740, 1.5839, 1.8256], + device='cuda:4'), covar=tensor([0.0867, 0.0645, 0.1448, 0.1625, 0.1084, 0.1742, 0.1927, 0.0756], + device='cuda:4'), in_proj_covar=tensor([0.0171, 0.0191, 0.0201, 0.0182, 0.0209, 0.0211, 0.0225, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:46:12,822 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151436.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:46:48,696 INFO [finetune.py:976] (4/7) Epoch 27, batch 2550, loss[loss=0.1603, simple_loss=0.2384, pruned_loss=0.04111, over 4817.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2429, pruned_loss=0.04902, over 955950.35 frames. ], batch size: 40, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:46:49,279 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.209e+01 1.472e+02 1.881e+02 2.470e+02 3.912e+02, threshold=3.762e+02, percent-clipped=2.0 +2023-03-27 08:47:13,715 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.69 vs. limit=5.0 +2023-03-27 08:47:24,842 INFO [finetune.py:976] (4/7) Epoch 27, batch 2600, loss[loss=0.1956, simple_loss=0.2678, pruned_loss=0.06174, over 4814.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2456, pruned_loss=0.05013, over 957410.62 frames. ], batch size: 33, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:47:42,188 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151540.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:48:03,458 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151556.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:48:16,181 INFO [finetune.py:976] (4/7) Epoch 27, batch 2650, loss[loss=0.152, simple_loss=0.229, pruned_loss=0.03751, over 4921.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2451, pruned_loss=0.04948, over 953866.84 frames. ], batch size: 33, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:48:16,787 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.199e+02 1.649e+02 1.887e+02 2.270e+02 4.456e+02, threshold=3.774e+02, percent-clipped=3.0 +2023-03-27 08:48:39,921 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151604.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:48:49,946 INFO [finetune.py:976] (4/7) Epoch 27, batch 2700, loss[loss=0.1235, simple_loss=0.1998, pruned_loss=0.02361, over 4769.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2441, pruned_loss=0.04884, over 954694.99 frames. ], batch size: 28, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:49:01,310 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151638.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:49:12,819 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=151654.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:49:29,776 INFO [finetune.py:976] (4/7) Epoch 27, batch 2750, loss[loss=0.1817, simple_loss=0.2537, pruned_loss=0.0549, over 4932.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2427, pruned_loss=0.04908, over 954155.19 frames. ], batch size: 38, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:49:30,372 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.940e+01 1.418e+02 1.693e+02 2.178e+02 3.976e+02, threshold=3.385e+02, percent-clipped=1.0 +2023-03-27 08:49:41,787 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-27 08:49:44,716 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151688.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:50:06,331 INFO [finetune.py:976] (4/7) Epoch 27, batch 2800, loss[loss=0.1424, simple_loss=0.2104, pruned_loss=0.03719, over 4834.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2395, pruned_loss=0.04823, over 953632.68 frames. ], batch size: 38, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:50:09,959 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2137, 2.0573, 1.7977, 1.9103, 1.9415, 1.9649, 1.9873, 2.6746], + device='cuda:4'), covar=tensor([0.3506, 0.3862, 0.3229, 0.3419, 0.3643, 0.2198, 0.3685, 0.1533], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0263, 0.0236, 0.0275, 0.0260, 0.0229, 0.0258, 0.0237], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:50:24,986 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151749.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:50:39,489 INFO [finetune.py:976] (4/7) Epoch 27, batch 2850, loss[loss=0.1602, simple_loss=0.241, pruned_loss=0.03969, over 4820.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2375, pruned_loss=0.04777, over 954818.12 frames. ], batch size: 51, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:50:40,099 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.092e+02 1.485e+02 1.795e+02 2.169e+02 3.375e+02, threshold=3.589e+02, percent-clipped=0.0 +2023-03-27 08:50:42,149 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1151, 1.9339, 1.7890, 1.7335, 1.8755, 1.9386, 1.9028, 2.5273], + device='cuda:4'), covar=tensor([0.3604, 0.3921, 0.3150, 0.3333, 0.3485, 0.2497, 0.3435, 0.1754], + device='cuda:4'), in_proj_covar=tensor([0.0291, 0.0264, 0.0237, 0.0276, 0.0260, 0.0230, 0.0259, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:50:59,490 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=151801.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:51:12,915 INFO [finetune.py:976] (4/7) Epoch 27, batch 2900, loss[loss=0.177, simple_loss=0.2571, pruned_loss=0.04841, over 4860.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.24, pruned_loss=0.0487, over 955900.94 frames. ], batch size: 44, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:51:25,534 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151840.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:51:54,447 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=151862.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:52:04,136 INFO [finetune.py:976] (4/7) Epoch 27, batch 2950, loss[loss=0.166, simple_loss=0.2442, pruned_loss=0.04388, over 4884.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2435, pruned_loss=0.0495, over 956819.98 frames. ], batch size: 35, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:52:04,749 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.132e+02 1.531e+02 1.876e+02 2.281e+02 4.815e+02, threshold=3.752e+02, percent-clipped=2.0 +2023-03-27 08:52:15,644 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151888.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:52:29,419 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6135, 1.5557, 1.4980, 1.6109, 1.4024, 3.2739, 1.3868, 1.7404], + device='cuda:4'), covar=tensor([0.3270, 0.2492, 0.2127, 0.2276, 0.1517, 0.0263, 0.2600, 0.1199], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0124, 0.0113, 0.0095, 0.0094, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 08:52:33,984 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.52 vs. limit=2.0 +2023-03-27 08:52:36,672 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-27 08:52:37,407 INFO [finetune.py:976] (4/7) Epoch 27, batch 3000, loss[loss=0.1799, simple_loss=0.2518, pruned_loss=0.05407, over 4870.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2464, pruned_loss=0.05077, over 954396.02 frames. ], batch size: 34, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:52:37,408 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 08:52:50,759 INFO [finetune.py:1010] (4/7) Epoch 27, validation: loss=0.1572, simple_loss=0.2248, pruned_loss=0.04486, over 2265189.00 frames. +2023-03-27 08:52:50,759 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6557MB +2023-03-27 08:52:51,515 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4679, 1.5611, 1.2352, 1.4917, 1.7987, 1.8112, 1.5748, 1.3706], + device='cuda:4'), covar=tensor([0.0421, 0.0278, 0.0694, 0.0295, 0.0237, 0.0420, 0.0304, 0.0386], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0106, 0.0147, 0.0111, 0.0101, 0.0115, 0.0103, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.7585e-05, 8.0872e-05, 1.1412e-04, 8.4656e-05, 7.8246e-05, 8.4994e-05, + 7.6257e-05, 8.5715e-05], device='cuda:4') +2023-03-27 08:53:01,835 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151938.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:53:05,596 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.05 vs. limit=5.0 +2023-03-27 08:53:14,181 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=151954.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:53:32,164 INFO [finetune.py:976] (4/7) Epoch 27, batch 3050, loss[loss=0.1457, simple_loss=0.2229, pruned_loss=0.03424, over 4800.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2471, pruned_loss=0.05093, over 953954.32 frames. ], batch size: 25, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:53:32,746 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.022e+02 1.546e+02 1.837e+02 2.199e+02 4.500e+02, threshold=3.674e+02, percent-clipped=2.0 +2023-03-27 08:53:44,360 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=151986.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:53:45,634 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0047, 1.9985, 1.6691, 1.7873, 1.8160, 1.7617, 1.9084, 2.4864], + device='cuda:4'), covar=tensor([0.3339, 0.3369, 0.2957, 0.3272, 0.3569, 0.2198, 0.3190, 0.1512], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0263, 0.0236, 0.0274, 0.0259, 0.0229, 0.0258, 0.0237], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:53:55,892 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152002.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:53:59,515 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6081, 1.4031, 2.0603, 3.1441, 2.0700, 2.3032, 1.1422, 2.7096], + device='cuda:4'), covar=tensor([0.1675, 0.1389, 0.1237, 0.0555, 0.0853, 0.1470, 0.1667, 0.0409], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0114, 0.0131, 0.0162, 0.0100, 0.0134, 0.0123, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 08:54:06,689 INFO [finetune.py:976] (4/7) Epoch 27, batch 3100, loss[loss=0.1487, simple_loss=0.2225, pruned_loss=0.03749, over 4908.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2453, pruned_loss=0.04968, over 954538.41 frames. ], batch size: 36, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:54:18,668 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-27 08:54:18,919 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2324, 1.7920, 1.8876, 0.8217, 2.2148, 2.1826, 2.0784, 1.7657], + device='cuda:4'), covar=tensor([0.0864, 0.0782, 0.0477, 0.0694, 0.0442, 0.0657, 0.0446, 0.0803], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0147, 0.0129, 0.0123, 0.0131, 0.0130, 0.0142, 0.0150], + device='cuda:4'), out_proj_covar=tensor([8.8814e-05, 1.0554e-04, 9.1981e-05, 8.6245e-05, 9.1774e-05, 9.1957e-05, + 1.0062e-04, 1.0737e-04], device='cuda:4') +2023-03-27 08:54:23,672 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152044.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:54:32,749 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7786, 1.9491, 1.6608, 1.7539, 2.3741, 2.3078, 2.0812, 1.9140], + device='cuda:4'), covar=tensor([0.0464, 0.0346, 0.0606, 0.0343, 0.0247, 0.0508, 0.0426, 0.0346], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0107, 0.0147, 0.0112, 0.0102, 0.0116, 0.0104, 0.0114], + device='cuda:4'), out_proj_covar=tensor([7.8303e-05, 8.1594e-05, 1.1476e-04, 8.5342e-05, 7.8813e-05, 8.5596e-05, + 7.6928e-05, 8.6496e-05], device='cuda:4') +2023-03-27 08:54:40,564 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4962, 1.6265, 2.1657, 1.8039, 1.7789, 4.2457, 1.5482, 1.7781], + device='cuda:4'), covar=tensor([0.1154, 0.1920, 0.1344, 0.1030, 0.1692, 0.0192, 0.1601, 0.1852], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0083, 0.0074, 0.0077, 0.0092, 0.0080, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 08:54:41,692 INFO [finetune.py:976] (4/7) Epoch 27, batch 3150, loss[loss=0.1536, simple_loss=0.2289, pruned_loss=0.03919, over 4819.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2428, pruned_loss=0.04902, over 954769.89 frames. ], batch size: 38, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:54:42,285 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.945e+01 1.491e+02 1.827e+02 2.202e+02 3.039e+02, threshold=3.654e+02, percent-clipped=0.0 +2023-03-27 08:54:51,572 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.59 vs. limit=5.0 +2023-03-27 08:55:15,178 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8287, 3.7593, 3.5955, 1.7748, 3.8742, 3.0035, 0.9984, 2.6869], + device='cuda:4'), covar=tensor([0.2213, 0.2636, 0.1509, 0.3443, 0.1124, 0.0968, 0.4366, 0.1655], + device='cuda:4'), in_proj_covar=tensor([0.0149, 0.0179, 0.0159, 0.0129, 0.0160, 0.0124, 0.0148, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 08:55:21,810 INFO [finetune.py:976] (4/7) Epoch 27, batch 3200, loss[loss=0.1692, simple_loss=0.2489, pruned_loss=0.04475, over 4933.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2391, pruned_loss=0.04767, over 953345.99 frames. ], batch size: 38, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:55:46,798 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152157.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:55:54,656 INFO [finetune.py:976] (4/7) Epoch 27, batch 3250, loss[loss=0.1915, simple_loss=0.2595, pruned_loss=0.06176, over 4760.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2406, pruned_loss=0.04841, over 954726.59 frames. ], batch size: 54, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:55:55,264 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.795e+01 1.453e+02 1.756e+02 2.073e+02 3.538e+02, threshold=3.512e+02, percent-clipped=0.0 +2023-03-27 08:55:56,043 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.57 vs. limit=2.0 +2023-03-27 08:56:10,652 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3451, 2.3267, 1.9882, 2.6179, 2.3518, 2.1025, 2.7490, 2.4700], + device='cuda:4'), covar=tensor([0.1118, 0.2111, 0.2511, 0.2137, 0.2104, 0.1408, 0.2673, 0.1364], + device='cuda:4'), in_proj_covar=tensor([0.0188, 0.0189, 0.0235, 0.0252, 0.0248, 0.0206, 0.0214, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 08:56:32,279 INFO [finetune.py:976] (4/7) Epoch 27, batch 3300, loss[loss=0.1839, simple_loss=0.2595, pruned_loss=0.0542, over 4805.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2448, pruned_loss=0.05002, over 951289.55 frames. ], batch size: 41, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:56:35,416 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152225.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:56:50,149 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.7102, 1.6334, 1.6785, 0.9377, 1.7813, 2.1334, 2.0210, 1.4983], + device='cuda:4'), covar=tensor([0.0880, 0.0670, 0.0558, 0.0559, 0.0499, 0.0530, 0.0329, 0.0809], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0148, 0.0130, 0.0124, 0.0132, 0.0131, 0.0143, 0.0151], + device='cuda:4'), out_proj_covar=tensor([8.9637e-05, 1.0644e-04, 9.2813e-05, 8.6908e-05, 9.2727e-05, 9.2582e-05, + 1.0144e-04, 1.0807e-04], device='cuda:4') +2023-03-27 08:57:13,840 INFO [finetune.py:976] (4/7) Epoch 27, batch 3350, loss[loss=0.1515, simple_loss=0.2387, pruned_loss=0.03219, over 4887.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2459, pruned_loss=0.05003, over 953906.33 frames. ], batch size: 32, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:57:14,396 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.581e+01 1.606e+02 1.884e+02 2.337e+02 3.345e+02, threshold=3.768e+02, percent-clipped=0.0 +2023-03-27 08:57:20,898 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152273.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:57:33,564 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152286.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:57:52,064 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8946, 1.4344, 0.8299, 1.6979, 2.2095, 1.6149, 1.7848, 1.7842], + device='cuda:4'), covar=tensor([0.1595, 0.2072, 0.1980, 0.1258, 0.2004, 0.1989, 0.1374, 0.2044], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0092, 0.0108, 0.0090, 0.0118, 0.0092, 0.0097, 0.0087], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 08:57:54,509 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6991, 1.6807, 1.5255, 1.6531, 1.5434, 4.4720, 1.6342, 1.8876], + device='cuda:4'), covar=tensor([0.3550, 0.2684, 0.2336, 0.2596, 0.1657, 0.0123, 0.2463, 0.1306], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0124, 0.0113, 0.0095, 0.0094, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 08:58:01,023 INFO [finetune.py:976] (4/7) Epoch 27, batch 3400, loss[loss=0.2102, simple_loss=0.2813, pruned_loss=0.06957, over 4737.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2476, pruned_loss=0.0511, over 952590.86 frames. ], batch size: 54, lr: 2.92e-03, grad_scale: 16.0 +2023-03-27 08:58:09,667 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152334.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:58:16,693 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=152344.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:58:36,170 INFO [finetune.py:976] (4/7) Epoch 27, batch 3450, loss[loss=0.1474, simple_loss=0.2182, pruned_loss=0.03826, over 4741.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2469, pruned_loss=0.05044, over 952587.18 frames. ], batch size: 54, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 08:58:36,742 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.916e+01 1.467e+02 1.787e+02 2.252e+02 4.149e+02, threshold=3.573e+02, percent-clipped=3.0 +2023-03-27 08:58:58,954 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152392.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:59:14,831 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5361, 1.6278, 2.1625, 1.9118, 1.9681, 4.0901, 1.7114, 1.8330], + device='cuda:4'), covar=tensor([0.0974, 0.1854, 0.1143, 0.0896, 0.1399, 0.0247, 0.1345, 0.1623], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0083, 0.0074, 0.0077, 0.0092, 0.0080, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 08:59:18,827 INFO [finetune.py:976] (4/7) Epoch 27, batch 3500, loss[loss=0.1623, simple_loss=0.2424, pruned_loss=0.04113, over 4776.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2439, pruned_loss=0.04944, over 953462.80 frames. ], batch size: 26, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 08:59:34,582 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152445.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:59:43,781 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=152457.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 08:59:52,116 INFO [finetune.py:976] (4/7) Epoch 27, batch 3550, loss[loss=0.1579, simple_loss=0.224, pruned_loss=0.04589, over 4923.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2407, pruned_loss=0.04835, over 953478.93 frames. ], batch size: 38, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 08:59:52,706 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.528e+01 1.381e+02 1.664e+02 2.040e+02 3.997e+02, threshold=3.328e+02, percent-clipped=1.0 +2023-03-27 09:00:01,532 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7040, 1.6947, 2.2044, 3.2736, 2.2807, 2.4529, 1.2167, 2.8067], + device='cuda:4'), covar=tensor([0.1730, 0.1339, 0.1264, 0.0652, 0.0771, 0.1230, 0.1746, 0.0455], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0163, 0.0100, 0.0135, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 09:00:25,119 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152505.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:00:26,814 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152506.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:00:36,269 INFO [finetune.py:976] (4/7) Epoch 27, batch 3600, loss[loss=0.1751, simple_loss=0.2432, pruned_loss=0.05343, over 4819.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2389, pruned_loss=0.04782, over 954588.49 frames. ], batch size: 40, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 09:00:37,664 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1515, 1.7473, 2.1381, 2.1324, 1.8422, 1.8754, 2.0741, 2.0462], + device='cuda:4'), covar=tensor([0.4502, 0.4474, 0.3465, 0.4222, 0.5801, 0.4471, 0.5234, 0.3142], + device='cuda:4'), in_proj_covar=tensor([0.0266, 0.0247, 0.0268, 0.0296, 0.0295, 0.0272, 0.0301, 0.0253], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:00:42,404 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6439, 1.5979, 2.3479, 3.3383, 2.2759, 2.5144, 1.1924, 2.8263], + device='cuda:4'), covar=tensor([0.1703, 0.1290, 0.1161, 0.0519, 0.0759, 0.1575, 0.1648, 0.0446], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0163, 0.0100, 0.0135, 0.0124, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 09:00:47,993 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2275, 2.1511, 1.8817, 2.1296, 2.0099, 1.9977, 2.1100, 2.7222], + device='cuda:4'), covar=tensor([0.3728, 0.4017, 0.3173, 0.3683, 0.3768, 0.2436, 0.3519, 0.1572], + device='cuda:4'), in_proj_covar=tensor([0.0290, 0.0264, 0.0238, 0.0277, 0.0261, 0.0230, 0.0259, 0.0240], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:01:10,220 INFO [finetune.py:976] (4/7) Epoch 27, batch 3650, loss[loss=0.1972, simple_loss=0.2765, pruned_loss=0.0589, over 4809.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2414, pruned_loss=0.04903, over 953516.46 frames. ], batch size: 41, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 09:01:10,828 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.982e+01 1.594e+02 1.907e+02 2.265e+02 4.160e+02, threshold=3.814e+02, percent-clipped=3.0 +2023-03-27 09:01:17,560 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152581.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:01:18,211 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152582.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:01:18,829 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3907, 1.3765, 1.3903, 0.9624, 1.4599, 1.7068, 1.7302, 1.3201], + device='cuda:4'), covar=tensor([0.0927, 0.0685, 0.0575, 0.0495, 0.0484, 0.0501, 0.0268, 0.0723], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0147, 0.0130, 0.0122, 0.0131, 0.0130, 0.0142, 0.0150], + device='cuda:4'), out_proj_covar=tensor([8.8748e-05, 1.0559e-04, 9.2449e-05, 8.5966e-05, 9.1973e-05, 9.2186e-05, + 1.0059e-04, 1.0732e-04], device='cuda:4') +2023-03-27 09:01:22,465 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152589.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:01:46,436 INFO [finetune.py:976] (4/7) Epoch 27, batch 3700, loss[loss=0.206, simple_loss=0.2757, pruned_loss=0.06818, over 4730.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2436, pruned_loss=0.04921, over 952080.12 frames. ], batch size: 54, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 09:01:52,533 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152629.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:02:01,173 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152643.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:02:05,468 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152650.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:02:22,198 INFO [finetune.py:976] (4/7) Epoch 27, batch 3750, loss[loss=0.2632, simple_loss=0.3194, pruned_loss=0.1036, over 4839.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2448, pruned_loss=0.0491, over 952663.25 frames. ], batch size: 44, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 09:02:22,805 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.041e+02 1.491e+02 1.751e+02 2.166e+02 4.226e+02, threshold=3.502e+02, percent-clipped=3.0 +2023-03-27 09:02:50,062 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2728, 2.2295, 2.1200, 1.5623, 2.2674, 2.3513, 2.3711, 1.9063], + device='cuda:4'), covar=tensor([0.0559, 0.0659, 0.0871, 0.0932, 0.0592, 0.0738, 0.0616, 0.1077], + device='cuda:4'), in_proj_covar=tensor([0.0134, 0.0139, 0.0143, 0.0121, 0.0130, 0.0140, 0.0142, 0.0165], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:03:12,604 INFO [finetune.py:976] (4/7) Epoch 27, batch 3800, loss[loss=0.1667, simple_loss=0.2343, pruned_loss=0.04952, over 4910.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2456, pruned_loss=0.04957, over 953174.51 frames. ], batch size: 36, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 09:03:16,894 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=152726.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:03:24,709 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.2853, 4.5317, 4.8262, 5.1653, 4.9820, 4.6014, 5.4058, 1.5528], + device='cuda:4'), covar=tensor([0.0794, 0.0816, 0.0859, 0.0911, 0.1219, 0.1857, 0.0534, 0.6293], + device='cuda:4'), in_proj_covar=tensor([0.0351, 0.0246, 0.0281, 0.0294, 0.0332, 0.0286, 0.0303, 0.0300], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:03:26,576 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8621, 1.3956, 1.9464, 1.9341, 1.7129, 1.6764, 1.8738, 1.8975], + device='cuda:4'), covar=tensor([0.4233, 0.4086, 0.3224, 0.3509, 0.4706, 0.3867, 0.4456, 0.2911], + device='cuda:4'), in_proj_covar=tensor([0.0265, 0.0247, 0.0268, 0.0296, 0.0295, 0.0272, 0.0301, 0.0252], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:03:45,580 INFO [finetune.py:976] (4/7) Epoch 27, batch 3850, loss[loss=0.1749, simple_loss=0.2464, pruned_loss=0.05167, over 4734.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2446, pruned_loss=0.04935, over 951817.77 frames. ], batch size: 54, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 09:03:46,653 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.206e+01 1.335e+02 1.631e+02 2.144e+02 3.589e+02, threshold=3.262e+02, percent-clipped=1.0 +2023-03-27 09:03:52,623 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0796, 1.8931, 1.7338, 1.6316, 1.8002, 1.8615, 1.8677, 2.5280], + device='cuda:4'), covar=tensor([0.3790, 0.4047, 0.3144, 0.3631, 0.3913, 0.2343, 0.3381, 0.1732], + device='cuda:4'), in_proj_covar=tensor([0.0288, 0.0262, 0.0235, 0.0274, 0.0259, 0.0229, 0.0257, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:03:59,988 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=152787.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:04:12,776 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152801.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:04:24,006 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4047, 2.2368, 2.7905, 1.8402, 2.2815, 2.6369, 1.9565, 2.8089], + device='cuda:4'), covar=tensor([0.1523, 0.2112, 0.1621, 0.2222, 0.1064, 0.1709, 0.2830, 0.0872], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0207, 0.0194, 0.0190, 0.0175, 0.0214, 0.0218, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:04:25,546 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.12 vs. limit=5.0 +2023-03-27 09:04:28,201 INFO [finetune.py:976] (4/7) Epoch 27, batch 3900, loss[loss=0.1226, simple_loss=0.1936, pruned_loss=0.02578, over 4312.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2419, pruned_loss=0.04882, over 951778.48 frames. ], batch size: 65, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 09:05:01,436 INFO [finetune.py:976] (4/7) Epoch 27, batch 3950, loss[loss=0.1824, simple_loss=0.2427, pruned_loss=0.06102, over 4761.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2389, pruned_loss=0.04826, over 953945.77 frames. ], batch size: 23, lr: 2.91e-03, grad_scale: 16.0 +2023-03-27 09:05:02,045 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.044e+02 1.440e+02 1.686e+02 2.039e+02 3.105e+02, threshold=3.372e+02, percent-clipped=0.0 +2023-03-27 09:05:09,752 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=152881.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:05:43,211 INFO [finetune.py:976] (4/7) Epoch 27, batch 4000, loss[loss=0.1493, simple_loss=0.2245, pruned_loss=0.03709, over 4785.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2377, pruned_loss=0.0476, over 954930.28 frames. ], batch size: 29, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:05:49,734 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152929.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:05:49,760 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=152929.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:05:56,146 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152938.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:06:00,336 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=152945.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:06:16,490 INFO [finetune.py:976] (4/7) Epoch 27, batch 4050, loss[loss=0.1949, simple_loss=0.2738, pruned_loss=0.05802, over 4821.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2413, pruned_loss=0.04873, over 951894.88 frames. ], batch size: 39, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:06:17,095 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.469e+02 1.767e+02 2.180e+02 3.425e+02, threshold=3.534e+02, percent-clipped=1.0 +2023-03-27 09:06:20,836 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=152977.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:06:22,608 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7010, 1.5831, 2.0588, 2.0578, 1.8066, 3.6707, 1.5269, 1.7624], + device='cuda:4'), covar=tensor([0.0947, 0.1835, 0.1171, 0.0893, 0.1601, 0.0219, 0.1528, 0.1807], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0076, 0.0091, 0.0080, 0.0086, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 09:06:38,268 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 09:06:49,252 INFO [finetune.py:976] (4/7) Epoch 27, batch 4100, loss[loss=0.2297, simple_loss=0.2965, pruned_loss=0.08141, over 4927.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2435, pruned_loss=0.04943, over 953468.42 frames. ], batch size: 38, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:07:17,475 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1240, 3.6062, 3.7733, 3.9689, 3.8626, 3.6228, 4.1997, 1.4037], + device='cuda:4'), covar=tensor([0.0774, 0.0862, 0.0859, 0.0937, 0.1184, 0.1589, 0.0723, 0.5562], + device='cuda:4'), in_proj_covar=tensor([0.0355, 0.0250, 0.0285, 0.0297, 0.0336, 0.0290, 0.0307, 0.0305], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:07:21,744 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3602, 2.1486, 2.3177, 1.1570, 2.7054, 2.9490, 2.5652, 2.0114], + device='cuda:4'), covar=tensor([0.1098, 0.0986, 0.0609, 0.0812, 0.0606, 0.0650, 0.0441, 0.0962], + device='cuda:4'), in_proj_covar=tensor([0.0123, 0.0148, 0.0130, 0.0124, 0.0133, 0.0131, 0.0143, 0.0151], + device='cuda:4'), out_proj_covar=tensor([8.9437e-05, 1.0616e-04, 9.2971e-05, 8.6832e-05, 9.2892e-05, 9.2961e-05, + 1.0172e-04, 1.0806e-04], device='cuda:4') +2023-03-27 09:07:22,843 INFO [finetune.py:976] (4/7) Epoch 27, batch 4150, loss[loss=0.189, simple_loss=0.2695, pruned_loss=0.05423, over 4860.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2426, pruned_loss=0.04819, over 953192.34 frames. ], batch size: 31, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:07:23,442 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.058e+02 1.644e+02 1.926e+02 2.373e+02 3.999e+02, threshold=3.851e+02, percent-clipped=3.0 +2023-03-27 09:07:31,208 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=153082.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:07:40,296 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2326, 2.1238, 1.9853, 2.3488, 2.6738, 2.3422, 2.2869, 1.9342], + device='cuda:4'), covar=tensor([0.2176, 0.1780, 0.1753, 0.1624, 0.1400, 0.1008, 0.1784, 0.1787], + device='cuda:4'), in_proj_covar=tensor([0.0248, 0.0213, 0.0217, 0.0201, 0.0248, 0.0193, 0.0219, 0.0207], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:07:51,106 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153101.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:08:07,054 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=153117.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:08:13,347 INFO [finetune.py:976] (4/7) Epoch 27, batch 4200, loss[loss=0.1612, simple_loss=0.2418, pruned_loss=0.04029, over 4905.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2425, pruned_loss=0.0474, over 952827.72 frames. ], batch size: 37, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:08:36,710 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153149.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:08:49,827 INFO [finetune.py:976] (4/7) Epoch 27, batch 4250, loss[loss=0.1588, simple_loss=0.23, pruned_loss=0.04382, over 4114.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2416, pruned_loss=0.04759, over 953215.71 frames. ], batch size: 18, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:08:50,417 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.682e+01 1.570e+02 1.909e+02 2.227e+02 3.978e+02, threshold=3.818e+02, percent-clipped=1.0 +2023-03-27 09:08:54,807 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=153178.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:09:33,254 INFO [finetune.py:976] (4/7) Epoch 27, batch 4300, loss[loss=0.148, simple_loss=0.2242, pruned_loss=0.0359, over 4790.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2394, pruned_loss=0.04735, over 954489.31 frames. ], batch size: 29, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:09:36,429 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9131, 1.4185, 2.0300, 1.9750, 1.7724, 1.7225, 1.8720, 1.9298], + device='cuda:4'), covar=tensor([0.4247, 0.4111, 0.3220, 0.3669, 0.4977, 0.3909, 0.4667, 0.2967], + device='cuda:4'), in_proj_covar=tensor([0.0266, 0.0247, 0.0268, 0.0297, 0.0295, 0.0272, 0.0302, 0.0253], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:09:44,699 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153238.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:09:48,792 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6525, 2.5734, 2.1118, 2.8377, 2.5790, 2.2509, 3.0847, 2.7480], + device='cuda:4'), covar=tensor([0.1233, 0.2182, 0.2642, 0.2310, 0.2449, 0.1567, 0.2559, 0.1514], + device='cuda:4'), in_proj_covar=tensor([0.0189, 0.0190, 0.0235, 0.0253, 0.0248, 0.0206, 0.0215, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:09:49,805 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153245.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:09:54,027 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-27 09:10:06,730 INFO [finetune.py:976] (4/7) Epoch 27, batch 4350, loss[loss=0.1166, simple_loss=0.1899, pruned_loss=0.02162, over 4786.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2362, pruned_loss=0.04584, over 952547.43 frames. ], batch size: 26, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:10:07,330 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.429e+01 1.453e+02 1.753e+02 2.112e+02 4.699e+02, threshold=3.507e+02, percent-clipped=1.0 +2023-03-27 09:10:10,465 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6317, 1.5402, 1.4671, 1.6299, 1.2660, 3.4766, 1.3950, 1.7230], + device='cuda:4'), covar=tensor([0.3216, 0.2512, 0.2110, 0.2306, 0.1698, 0.0198, 0.2521, 0.1234], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0116, 0.0120, 0.0124, 0.0113, 0.0095, 0.0094, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 09:10:11,664 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=153278.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:10:16,464 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153286.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:10:21,197 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153293.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:10:41,648 INFO [finetune.py:976] (4/7) Epoch 27, batch 4400, loss[loss=0.2081, simple_loss=0.2758, pruned_loss=0.07025, over 4821.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2378, pruned_loss=0.04684, over 952318.55 frames. ], batch size: 39, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:10:51,689 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5718, 1.1208, 0.8309, 1.4592, 1.9848, 1.1215, 1.3669, 1.4878], + device='cuda:4'), covar=tensor([0.1502, 0.2122, 0.1905, 0.1177, 0.1902, 0.1930, 0.1402, 0.1818], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0093, 0.0110, 0.0091, 0.0120, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 09:11:01,179 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=153339.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:11:16,397 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2252, 2.3187, 2.2614, 1.6474, 2.2349, 2.5514, 2.3420, 2.0041], + device='cuda:4'), covar=tensor([0.0613, 0.0603, 0.0745, 0.0931, 0.1200, 0.0680, 0.0657, 0.1091], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0139, 0.0142, 0.0121, 0.0130, 0.0140, 0.0141, 0.0164], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:11:23,048 INFO [finetune.py:976] (4/7) Epoch 27, batch 4450, loss[loss=0.1652, simple_loss=0.239, pruned_loss=0.04566, over 4836.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2406, pruned_loss=0.04756, over 949908.95 frames. ], batch size: 30, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:11:23,634 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.494e+02 1.793e+02 2.132e+02 3.020e+02, threshold=3.586e+02, percent-clipped=0.0 +2023-03-27 09:11:30,906 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153382.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:11:56,789 INFO [finetune.py:976] (4/7) Epoch 27, batch 4500, loss[loss=0.2106, simple_loss=0.2873, pruned_loss=0.06693, over 4752.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2431, pruned_loss=0.04862, over 951499.16 frames. ], batch size: 54, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:11:57,459 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=153421.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:12:03,377 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153430.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:12:12,706 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.19 vs. limit=5.0 +2023-03-27 09:12:29,939 INFO [finetune.py:976] (4/7) Epoch 27, batch 4550, loss[loss=0.1829, simple_loss=0.2546, pruned_loss=0.05557, over 4824.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2445, pruned_loss=0.04914, over 950642.74 frames. ], batch size: 30, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:12:30,512 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.518e+01 1.590e+02 1.867e+02 2.229e+02 3.919e+02, threshold=3.734e+02, percent-clipped=1.0 +2023-03-27 09:12:31,761 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=153473.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:12:37,666 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=153482.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:12:41,916 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6789, 3.6832, 3.3913, 1.4951, 3.7797, 2.7774, 0.7543, 2.4895], + device='cuda:4'), covar=tensor([0.2429, 0.1774, 0.1706, 0.3738, 0.0943, 0.1050, 0.4577, 0.1517], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0181, 0.0161, 0.0131, 0.0162, 0.0125, 0.0150, 0.0126], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 09:12:43,150 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.3231, 2.8932, 3.1278, 3.2689, 3.0840, 2.8886, 3.3652, 0.9314], + device='cuda:4'), covar=tensor([0.1074, 0.1112, 0.1094, 0.1124, 0.1649, 0.2059, 0.1111, 0.5755], + device='cuda:4'), in_proj_covar=tensor([0.0351, 0.0247, 0.0282, 0.0293, 0.0332, 0.0286, 0.0304, 0.0301], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:13:14,346 INFO [finetune.py:976] (4/7) Epoch 27, batch 4600, loss[loss=0.2032, simple_loss=0.2814, pruned_loss=0.06249, over 4718.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.244, pruned_loss=0.04869, over 950687.77 frames. ], batch size: 23, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:13:15,686 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8103, 1.1254, 1.8666, 1.8149, 1.6475, 1.5644, 1.7199, 1.7810], + device='cuda:4'), covar=tensor([0.3184, 0.3298, 0.2557, 0.2954, 0.3931, 0.3215, 0.3520, 0.2438], + device='cuda:4'), in_proj_covar=tensor([0.0266, 0.0247, 0.0268, 0.0297, 0.0295, 0.0273, 0.0302, 0.0253], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:13:56,979 INFO [finetune.py:976] (4/7) Epoch 27, batch 4650, loss[loss=0.1716, simple_loss=0.2399, pruned_loss=0.05166, over 4769.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2428, pruned_loss=0.049, over 951259.90 frames. ], batch size: 26, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:13:57,587 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.034e+02 1.461e+02 1.737e+02 2.165e+02 3.643e+02, threshold=3.475e+02, percent-clipped=0.0 +2023-03-27 09:14:02,489 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8133, 1.5288, 2.0506, 1.3403, 1.7591, 1.9549, 1.4804, 2.0518], + device='cuda:4'), covar=tensor([0.1184, 0.2324, 0.1039, 0.1534, 0.0932, 0.1246, 0.2971, 0.0815], + device='cuda:4'), in_proj_covar=tensor([0.0195, 0.0208, 0.0196, 0.0190, 0.0175, 0.0216, 0.0218, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:14:08,166 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.47 vs. limit=5.0 +2023-03-27 09:14:17,673 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-03-27 09:14:31,814 INFO [finetune.py:976] (4/7) Epoch 27, batch 4700, loss[loss=0.1501, simple_loss=0.2163, pruned_loss=0.04193, over 4817.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2398, pruned_loss=0.04763, over 952891.12 frames. ], batch size: 25, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:14:33,524 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-27 09:14:47,930 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=153634.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:15:12,023 INFO [finetune.py:976] (4/7) Epoch 27, batch 4750, loss[loss=0.1408, simple_loss=0.2142, pruned_loss=0.03371, over 4718.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2375, pruned_loss=0.04709, over 952753.07 frames. ], batch size: 23, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:15:13,081 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.473e+02 1.795e+02 2.173e+02 4.465e+02, threshold=3.590e+02, percent-clipped=3.0 +2023-03-27 09:15:45,876 INFO [finetune.py:976] (4/7) Epoch 27, batch 4800, loss[loss=0.162, simple_loss=0.2368, pruned_loss=0.04354, over 4826.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2402, pruned_loss=0.04809, over 953465.50 frames. ], batch size: 30, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:15:53,611 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6104, 1.5081, 1.3504, 1.6955, 2.0586, 1.6750, 1.4101, 1.3489], + device='cuda:4'), covar=tensor([0.2131, 0.2036, 0.1891, 0.1577, 0.1513, 0.1226, 0.2363, 0.1889], + device='cuda:4'), in_proj_covar=tensor([0.0248, 0.0212, 0.0217, 0.0201, 0.0247, 0.0193, 0.0219, 0.0206], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:16:28,416 INFO [finetune.py:976] (4/7) Epoch 27, batch 4850, loss[loss=0.2068, simple_loss=0.2783, pruned_loss=0.06764, over 4921.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2443, pruned_loss=0.04923, over 954191.33 frames. ], batch size: 38, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:16:28,981 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.100e+02 1.544e+02 1.777e+02 2.223e+02 4.381e+02, threshold=3.554e+02, percent-clipped=2.0 +2023-03-27 09:16:30,774 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153773.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:16:33,641 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=153777.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:16:51,350 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4467, 2.2706, 1.8418, 0.8537, 1.9825, 1.8787, 1.8141, 2.1637], + device='cuda:4'), covar=tensor([0.0806, 0.0751, 0.1399, 0.1994, 0.1235, 0.2435, 0.2176, 0.0838], + device='cuda:4'), in_proj_covar=tensor([0.0171, 0.0191, 0.0201, 0.0182, 0.0210, 0.0211, 0.0224, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:17:00,918 INFO [finetune.py:976] (4/7) Epoch 27, batch 4900, loss[loss=0.2381, simple_loss=0.3002, pruned_loss=0.08796, over 4905.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2455, pruned_loss=0.05027, over 954001.28 frames. ], batch size: 36, lr: 2.91e-03, grad_scale: 32.0 +2023-03-27 09:17:01,623 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153821.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:17:34,602 INFO [finetune.py:976] (4/7) Epoch 27, batch 4950, loss[loss=0.1924, simple_loss=0.2782, pruned_loss=0.05332, over 4815.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2466, pruned_loss=0.05023, over 956207.00 frames. ], batch size: 38, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:17:35,195 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.496e+02 1.750e+02 2.158e+02 3.393e+02, threshold=3.501e+02, percent-clipped=0.0 +2023-03-27 09:18:10,033 INFO [finetune.py:976] (4/7) Epoch 27, batch 5000, loss[loss=0.161, simple_loss=0.2364, pruned_loss=0.04285, over 4901.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2439, pruned_loss=0.04909, over 956284.52 frames. ], batch size: 37, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:18:17,225 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=153923.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:18:21,290 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8511, 1.8479, 1.5876, 1.9526, 2.4378, 2.0615, 1.7366, 1.5347], + device='cuda:4'), covar=tensor([0.2285, 0.1919, 0.1971, 0.1678, 0.1553, 0.1151, 0.2296, 0.1933], + device='cuda:4'), in_proj_covar=tensor([0.0247, 0.0211, 0.0216, 0.0201, 0.0247, 0.0191, 0.0218, 0.0206], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:18:28,733 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=153934.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:18:36,511 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0440, 1.9907, 2.0949, 1.5834, 1.9908, 2.1432, 2.1555, 1.6676], + device='cuda:4'), covar=tensor([0.0490, 0.0490, 0.0563, 0.0772, 0.0867, 0.0525, 0.0448, 0.1028], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0137, 0.0142, 0.0120, 0.0129, 0.0140, 0.0141, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:18:36,863 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-27 09:18:57,222 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-27 09:19:01,713 INFO [finetune.py:976] (4/7) Epoch 27, batch 5050, loss[loss=0.156, simple_loss=0.2247, pruned_loss=0.04365, over 4830.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2405, pruned_loss=0.04815, over 955949.16 frames. ], batch size: 30, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:19:02,312 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.210e+01 1.435e+02 1.808e+02 2.168e+02 4.775e+02, threshold=3.617e+02, percent-clipped=1.0 +2023-03-27 09:19:10,569 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=153982.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:19:11,857 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=153984.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:19:36,570 INFO [finetune.py:976] (4/7) Epoch 27, batch 5100, loss[loss=0.1606, simple_loss=0.2321, pruned_loss=0.04454, over 4821.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2362, pruned_loss=0.04584, over 956854.94 frames. ], batch size: 41, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:20:06,466 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154049.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 09:20:17,387 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6513, 1.5059, 1.0941, 0.2670, 1.2121, 1.4788, 1.4442, 1.4459], + device='cuda:4'), covar=tensor([0.0864, 0.0824, 0.1372, 0.1917, 0.1431, 0.2511, 0.2380, 0.0875], + device='cuda:4'), in_proj_covar=tensor([0.0171, 0.0192, 0.0202, 0.0182, 0.0211, 0.0211, 0.0224, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:20:19,663 INFO [finetune.py:976] (4/7) Epoch 27, batch 5150, loss[loss=0.2139, simple_loss=0.2814, pruned_loss=0.07317, over 4817.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.2364, pruned_loss=0.04607, over 954214.23 frames. ], batch size: 38, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:20:20,256 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.883e+01 1.500e+02 1.789e+02 2.109e+02 4.792e+02, threshold=3.578e+02, percent-clipped=3.0 +2023-03-27 09:20:23,485 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154076.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:20:24,061 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154077.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:20:28,816 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154084.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:20:28,832 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1662, 1.9129, 1.9542, 1.0321, 2.3030, 2.4250, 2.1559, 1.7394], + device='cuda:4'), covar=tensor([0.1023, 0.0706, 0.0535, 0.0687, 0.0472, 0.0636, 0.0427, 0.0735], + device='cuda:4'), in_proj_covar=tensor([0.0121, 0.0147, 0.0129, 0.0122, 0.0131, 0.0131, 0.0142, 0.0150], + device='cuda:4'), out_proj_covar=tensor([8.8265e-05, 1.0545e-04, 9.1998e-05, 8.5872e-05, 9.1708e-05, 9.2521e-05, + 1.0081e-04, 1.0733e-04], device='cuda:4') +2023-03-27 09:20:46,983 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154110.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 09:20:53,423 INFO [finetune.py:976] (4/7) Epoch 27, batch 5200, loss[loss=0.201, simple_loss=0.2761, pruned_loss=0.06296, over 4913.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2412, pruned_loss=0.04815, over 952308.76 frames. ], batch size: 36, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:20:56,441 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154125.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:20:58,026 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 09:21:04,322 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154137.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 09:21:12,329 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154145.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:21:29,292 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5563, 3.7478, 3.5882, 1.8026, 3.8274, 2.9008, 0.9272, 2.6883], + device='cuda:4'), covar=tensor([0.2152, 0.2108, 0.1501, 0.3210, 0.1015, 0.0980, 0.4238, 0.1357], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0180, 0.0160, 0.0131, 0.0162, 0.0123, 0.0149, 0.0125], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 09:21:30,242 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.64 vs. limit=5.0 +2023-03-27 09:21:34,724 INFO [finetune.py:976] (4/7) Epoch 27, batch 5250, loss[loss=0.1866, simple_loss=0.2608, pruned_loss=0.05619, over 4832.00 frames. ], tot_loss[loss=0.171, simple_loss=0.244, pruned_loss=0.04899, over 953137.86 frames. ], batch size: 49, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:21:35,334 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.122e+02 1.556e+02 1.889e+02 2.346e+02 3.556e+02, threshold=3.778e+02, percent-clipped=0.0 +2023-03-27 09:22:04,955 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4816, 2.3262, 2.0184, 0.9671, 2.0414, 1.8633, 1.7969, 2.1878], + device='cuda:4'), covar=tensor([0.0844, 0.0720, 0.1396, 0.1811, 0.1268, 0.2248, 0.2018, 0.0823], + device='cuda:4'), in_proj_covar=tensor([0.0171, 0.0191, 0.0201, 0.0181, 0.0210, 0.0210, 0.0224, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:22:08,470 INFO [finetune.py:976] (4/7) Epoch 27, batch 5300, loss[loss=0.1924, simple_loss=0.269, pruned_loss=0.05793, over 4907.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2471, pruned_loss=0.05038, over 955462.04 frames. ], batch size: 36, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:22:09,185 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154221.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:22:26,266 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-27 09:22:26,673 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.0591, 2.6038, 2.5201, 1.3129, 2.7385, 2.1624, 1.9322, 2.3421], + device='cuda:4'), covar=tensor([0.0982, 0.1075, 0.1961, 0.2265, 0.1601, 0.2374, 0.2513, 0.1368], + device='cuda:4'), in_proj_covar=tensor([0.0171, 0.0191, 0.0201, 0.0181, 0.0210, 0.0210, 0.0224, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:22:41,918 INFO [finetune.py:976] (4/7) Epoch 27, batch 5350, loss[loss=0.1477, simple_loss=0.2142, pruned_loss=0.0406, over 4811.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2464, pruned_loss=0.04944, over 956162.65 frames. ], batch size: 25, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:22:42,509 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.999e+01 1.513e+02 1.798e+02 2.139e+02 3.270e+02, threshold=3.596e+02, percent-clipped=0.0 +2023-03-27 09:22:47,881 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154279.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:22:49,730 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154282.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 09:22:52,188 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154286.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:22:55,593 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154291.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:23:15,341 INFO [finetune.py:976] (4/7) Epoch 27, batch 5400, loss[loss=0.1219, simple_loss=0.2073, pruned_loss=0.01821, over 4781.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2437, pruned_loss=0.04889, over 955408.89 frames. ], batch size: 29, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:23:33,027 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154337.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 09:23:42,783 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154347.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:23:49,145 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154352.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:23:54,381 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1003, 1.4634, 1.9798, 1.9560, 1.7374, 1.7436, 1.9247, 1.8869], + device='cuda:4'), covar=tensor([0.3709, 0.3752, 0.3191, 0.3623, 0.4848, 0.4088, 0.4208, 0.3021], + device='cuda:4'), in_proj_covar=tensor([0.0266, 0.0248, 0.0269, 0.0298, 0.0296, 0.0274, 0.0302, 0.0253], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:24:08,868 INFO [finetune.py:976] (4/7) Epoch 27, batch 5450, loss[loss=0.1512, simple_loss=0.2188, pruned_loss=0.04179, over 4839.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2409, pruned_loss=0.04808, over 957690.04 frames. ], batch size: 40, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:24:09,462 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.752e+01 1.439e+02 1.730e+02 2.063e+02 4.741e+02, threshold=3.460e+02, percent-clipped=1.0 +2023-03-27 09:24:26,426 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154398.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 09:24:32,220 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154405.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 09:24:36,243 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2263, 1.8729, 2.5113, 1.7303, 2.3376, 2.4991, 1.8250, 2.6277], + device='cuda:4'), covar=tensor([0.1252, 0.2029, 0.1445, 0.1943, 0.0869, 0.1278, 0.2652, 0.0745], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0209, 0.0195, 0.0190, 0.0175, 0.0215, 0.0219, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:24:42,544 INFO [finetune.py:976] (4/7) Epoch 27, batch 5500, loss[loss=0.1758, simple_loss=0.2498, pruned_loss=0.05095, over 4851.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2383, pruned_loss=0.04698, over 958321.32 frames. ], batch size: 44, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:24:49,890 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154432.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 09:24:54,779 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154440.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:24:56,576 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.35 vs. limit=5.0 +2023-03-27 09:25:04,344 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2070, 2.3077, 1.7583, 2.5915, 2.2291, 1.9213, 2.8279, 2.3322], + device='cuda:4'), covar=tensor([0.1480, 0.2175, 0.3137, 0.2417, 0.2805, 0.1786, 0.3111, 0.1769], + device='cuda:4'), in_proj_covar=tensor([0.0189, 0.0189, 0.0236, 0.0253, 0.0249, 0.0205, 0.0214, 0.0201], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:25:27,010 INFO [finetune.py:976] (4/7) Epoch 27, batch 5550, loss[loss=0.2092, simple_loss=0.2836, pruned_loss=0.06743, over 4764.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2405, pruned_loss=0.04765, over 959593.23 frames. ], batch size: 54, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:25:27,601 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.282e+01 1.501e+02 1.802e+02 2.038e+02 5.335e+02, threshold=3.603e+02, percent-clipped=2.0 +2023-03-27 09:25:29,559 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154474.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:25:36,844 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.0012, 4.3954, 4.6382, 4.8817, 4.7634, 4.5081, 5.1472, 1.5328], + device='cuda:4'), covar=tensor([0.0795, 0.0806, 0.0740, 0.0789, 0.1112, 0.1736, 0.0605, 0.5955], + device='cuda:4'), in_proj_covar=tensor([0.0349, 0.0245, 0.0281, 0.0293, 0.0331, 0.0285, 0.0302, 0.0299], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:25:51,000 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-27 09:25:57,466 INFO [finetune.py:976] (4/7) Epoch 27, batch 5600, loss[loss=0.1879, simple_loss=0.265, pruned_loss=0.05538, over 4841.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2437, pruned_loss=0.04842, over 955723.60 frames. ], batch size: 49, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:26:07,314 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154535.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:26:11,672 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-03-27 09:26:27,429 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4690, 1.3292, 1.2800, 1.2786, 0.8932, 2.2895, 0.7857, 1.2620], + device='cuda:4'), covar=tensor([0.3442, 0.2774, 0.2337, 0.2687, 0.1903, 0.0353, 0.2899, 0.1382], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0117, 0.0121, 0.0124, 0.0113, 0.0095, 0.0094, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 09:26:30,308 INFO [finetune.py:976] (4/7) Epoch 27, batch 5650, loss[loss=0.1487, simple_loss=0.2152, pruned_loss=0.04114, over 3972.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2455, pruned_loss=0.04862, over 956113.75 frames. ], batch size: 17, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:26:30,866 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.301e+01 1.398e+02 1.740e+02 2.119e+02 4.723e+02, threshold=3.480e+02, percent-clipped=2.0 +2023-03-27 09:26:39,109 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154577.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 09:26:40,313 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154579.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:27:07,855 INFO [finetune.py:976] (4/7) Epoch 27, batch 5700, loss[loss=0.1506, simple_loss=0.2092, pruned_loss=0.04596, over 3984.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2421, pruned_loss=0.04887, over 935840.54 frames. ], batch size: 17, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:27:07,940 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.8392, 3.6052, 3.2530, 1.9535, 3.3440, 2.8990, 2.9046, 3.2018], + device='cuda:4'), covar=tensor([0.0725, 0.0618, 0.1309, 0.1801, 0.1451, 0.1565, 0.1306, 0.0812], + device='cuda:4'), in_proj_covar=tensor([0.0171, 0.0190, 0.0201, 0.0181, 0.0210, 0.0210, 0.0223, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:27:12,038 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154627.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:27:12,699 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154628.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:27:20,865 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154642.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:27:34,185 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154647.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:27:34,737 INFO [finetune.py:976] (4/7) Epoch 28, batch 0, loss[loss=0.1975, simple_loss=0.2777, pruned_loss=0.05868, over 4920.00 frames. ], tot_loss[loss=0.1975, simple_loss=0.2777, pruned_loss=0.05868, over 4920.00 frames. ], batch size: 42, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:27:34,737 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 09:27:44,409 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1633, 1.4670, 1.4478, 0.7660, 1.4449, 1.6465, 1.7319, 1.3789], + device='cuda:4'), covar=tensor([0.0938, 0.0578, 0.0626, 0.0526, 0.0556, 0.0681, 0.0336, 0.0765], + device='cuda:4'), in_proj_covar=tensor([0.0121, 0.0147, 0.0129, 0.0121, 0.0131, 0.0130, 0.0142, 0.0150], + device='cuda:4'), out_proj_covar=tensor([8.8337e-05, 1.0504e-04, 9.1824e-05, 8.5222e-05, 9.1862e-05, 9.1862e-05, + 1.0083e-04, 1.0708e-04], device='cuda:4') +2023-03-27 09:27:54,286 INFO [finetune.py:1010] (4/7) Epoch 28, validation: loss=0.1583, simple_loss=0.2265, pruned_loss=0.04511, over 2265189.00 frames. +2023-03-27 09:27:54,286 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6557MB +2023-03-27 09:27:55,412 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7090, 1.2722, 0.8854, 1.6336, 2.0324, 1.4784, 1.5622, 1.6263], + device='cuda:4'), covar=tensor([0.1331, 0.1955, 0.1764, 0.1079, 0.1982, 0.1923, 0.1351, 0.1730], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0093, 0.0109, 0.0091, 0.0119, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 09:28:08,716 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.011e+02 1.488e+02 1.773e+02 2.221e+02 3.199e+02, threshold=3.546e+02, percent-clipped=0.0 +2023-03-27 09:28:21,200 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154689.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:28:24,050 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154693.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 09:28:27,013 INFO [finetune.py:976] (4/7) Epoch 28, batch 50, loss[loss=0.1716, simple_loss=0.248, pruned_loss=0.04762, over 4883.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2475, pruned_loss=0.05019, over 215647.03 frames. ], batch size: 35, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:28:32,685 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154705.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 09:28:44,246 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154720.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:28:52,125 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154732.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:28:57,924 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154740.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:29:03,130 INFO [finetune.py:976] (4/7) Epoch 28, batch 100, loss[loss=0.1577, simple_loss=0.2335, pruned_loss=0.04092, over 4825.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2388, pruned_loss=0.04697, over 380737.60 frames. ], batch size: 33, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:29:11,916 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154753.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 09:29:26,906 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.078e+02 1.413e+02 1.713e+02 2.093e+02 4.180e+02, threshold=3.426e+02, percent-clipped=2.0 +2023-03-27 09:29:32,451 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154780.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:29:33,129 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=154781.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:29:37,832 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154788.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:29:44,901 INFO [finetune.py:976] (4/7) Epoch 28, batch 150, loss[loss=0.1366, simple_loss=0.2148, pruned_loss=0.02918, over 4828.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.2355, pruned_loss=0.04603, over 509919.01 frames. ], batch size: 40, lr: 2.90e-03, grad_scale: 32.0 +2023-03-27 09:29:45,681 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-03-27 09:30:06,037 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154830.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:30:18,390 INFO [finetune.py:976] (4/7) Epoch 28, batch 200, loss[loss=0.1808, simple_loss=0.2452, pruned_loss=0.05825, over 4926.00 frames. ], tot_loss[loss=0.1631, simple_loss=0.2346, pruned_loss=0.04584, over 608758.51 frames. ], batch size: 38, lr: 2.89e-03, grad_scale: 32.0 +2023-03-27 09:30:40,612 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.190e+01 1.565e+02 1.831e+02 2.234e+02 3.641e+02, threshold=3.662e+02, percent-clipped=1.0 +2023-03-27 09:30:48,050 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154877.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 09:30:49,247 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4948, 1.5069, 1.8150, 1.7164, 1.5971, 3.2943, 1.3502, 1.5670], + device='cuda:4'), covar=tensor([0.1029, 0.1958, 0.1186, 0.0974, 0.1670, 0.0271, 0.1569, 0.1819], + device='cuda:4'), in_proj_covar=tensor([0.0074, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0079], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 09:31:02,680 INFO [finetune.py:976] (4/7) Epoch 28, batch 250, loss[loss=0.1704, simple_loss=0.2461, pruned_loss=0.04737, over 4816.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.238, pruned_loss=0.04678, over 687730.81 frames. ], batch size: 39, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:31:20,576 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154925.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:31:31,433 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154942.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:31:34,987 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154947.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:31:35,502 INFO [finetune.py:976] (4/7) Epoch 28, batch 300, loss[loss=0.1756, simple_loss=0.2614, pruned_loss=0.04494, over 4862.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2432, pruned_loss=0.04882, over 747759.06 frames. ], batch size: 34, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:31:51,471 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.026e+02 1.523e+02 1.869e+02 2.212e+02 3.864e+02, threshold=3.739e+02, percent-clipped=1.0 +2023-03-27 09:31:59,668 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=154976.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:32:07,917 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=154984.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:32:11,535 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154990.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:32:13,387 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=154993.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 09:32:15,017 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=154995.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:32:16,795 INFO [finetune.py:976] (4/7) Epoch 28, batch 350, loss[loss=0.1607, simple_loss=0.2449, pruned_loss=0.03829, over 4886.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.243, pruned_loss=0.04902, over 791340.09 frames. ], batch size: 35, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:32:17,096 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-27 09:32:43,060 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155037.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:32:45,385 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155041.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 09:32:49,880 INFO [finetune.py:976] (4/7) Epoch 28, batch 400, loss[loss=0.2244, simple_loss=0.2803, pruned_loss=0.08424, over 4818.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2443, pruned_loss=0.04903, over 828636.54 frames. ], batch size: 38, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:33:12,960 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-03-27 09:33:13,501 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155069.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:33:15,062 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.906e+01 1.559e+02 1.879e+02 2.352e+02 4.263e+02, threshold=3.758e+02, percent-clipped=3.0 +2023-03-27 09:33:18,285 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155076.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:33:20,116 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.3913, 3.7693, 4.0171, 4.1857, 4.1922, 3.9553, 4.4654, 1.3891], + device='cuda:4'), covar=tensor([0.0738, 0.0886, 0.0809, 0.0935, 0.1105, 0.1510, 0.0726, 0.6032], + device='cuda:4'), in_proj_covar=tensor([0.0354, 0.0248, 0.0285, 0.0298, 0.0336, 0.0288, 0.0307, 0.0304], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:33:31,490 INFO [finetune.py:976] (4/7) Epoch 28, batch 450, loss[loss=0.124, simple_loss=0.2029, pruned_loss=0.02258, over 4018.00 frames. ], tot_loss[loss=0.17, simple_loss=0.243, pruned_loss=0.04851, over 857227.67 frames. ], batch size: 17, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:33:54,285 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155130.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:33:54,314 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155130.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:34:03,764 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.81 vs. limit=5.0 +2023-03-27 09:34:05,150 INFO [finetune.py:976] (4/7) Epoch 28, batch 500, loss[loss=0.1674, simple_loss=0.2253, pruned_loss=0.0548, over 4861.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2412, pruned_loss=0.04819, over 881502.41 frames. ], batch size: 44, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:34:28,558 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.687e+01 1.475e+02 1.683e+02 2.204e+02 4.497e+02, threshold=3.366e+02, percent-clipped=1.0 +2023-03-27 09:34:37,155 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155178.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:34:49,245 INFO [finetune.py:976] (4/7) Epoch 28, batch 550, loss[loss=0.1466, simple_loss=0.2284, pruned_loss=0.03239, over 4818.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2385, pruned_loss=0.04758, over 897676.77 frames. ], batch size: 25, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:35:01,230 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3304, 2.2403, 2.0170, 1.0139, 2.1235, 1.8378, 1.7406, 2.1790], + device='cuda:4'), covar=tensor([0.1031, 0.0763, 0.1415, 0.1777, 0.1233, 0.2271, 0.2041, 0.0794], + device='cuda:4'), in_proj_covar=tensor([0.0172, 0.0192, 0.0202, 0.0182, 0.0210, 0.0210, 0.0224, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:35:15,247 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6941, 1.4506, 1.9083, 1.3869, 1.7987, 1.8131, 1.3255, 2.0852], + device='cuda:4'), covar=tensor([0.1203, 0.2285, 0.1286, 0.1542, 0.0868, 0.1244, 0.3043, 0.0750], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0208, 0.0194, 0.0189, 0.0174, 0.0213, 0.0219, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:35:23,083 INFO [finetune.py:976] (4/7) Epoch 28, batch 600, loss[loss=0.2258, simple_loss=0.3009, pruned_loss=0.07539, over 4811.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2398, pruned_loss=0.04827, over 910178.63 frames. ], batch size: 45, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:35:23,835 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1493, 2.2733, 1.8532, 2.3144, 2.1133, 2.0767, 2.1293, 2.9284], + device='cuda:4'), covar=tensor([0.3639, 0.4230, 0.3309, 0.3977, 0.4399, 0.2433, 0.4013, 0.1633], + device='cuda:4'), in_proj_covar=tensor([0.0291, 0.0265, 0.0238, 0.0276, 0.0262, 0.0231, 0.0260, 0.0240], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:35:32,569 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4357, 2.1444, 2.7883, 1.7823, 2.4072, 2.6253, 1.9000, 2.7436], + device='cuda:4'), covar=tensor([0.1295, 0.1845, 0.1427, 0.1980, 0.0904, 0.1234, 0.2715, 0.0752], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0207, 0.0193, 0.0189, 0.0174, 0.0213, 0.0219, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:35:38,979 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.640e+01 1.454e+02 1.702e+02 1.998e+02 4.828e+02, threshold=3.403e+02, percent-clipped=2.0 +2023-03-27 09:35:53,268 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155284.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:36:05,106 INFO [finetune.py:976] (4/7) Epoch 28, batch 650, loss[loss=0.1455, simple_loss=0.2263, pruned_loss=0.03234, over 4802.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2439, pruned_loss=0.04972, over 920458.51 frames. ], batch size: 45, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:36:29,093 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155332.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:36:29,113 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155332.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:36:38,735 INFO [finetune.py:976] (4/7) Epoch 28, batch 700, loss[loss=0.1676, simple_loss=0.2554, pruned_loss=0.03991, over 4917.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2452, pruned_loss=0.05038, over 927721.14 frames. ], batch size: 38, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:36:44,110 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-03-27 09:36:54,659 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.124e+02 1.563e+02 1.812e+02 2.261e+02 4.160e+02, threshold=3.625e+02, percent-clipped=3.0 +2023-03-27 09:36:57,813 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155376.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:37:19,500 INFO [finetune.py:976] (4/7) Epoch 28, batch 750, loss[loss=0.1918, simple_loss=0.2526, pruned_loss=0.06555, over 4767.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2453, pruned_loss=0.05018, over 933751.72 frames. ], batch size: 28, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:37:40,159 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155424.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:37:41,310 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155425.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:37:56,715 INFO [finetune.py:976] (4/7) Epoch 28, batch 800, loss[loss=0.1845, simple_loss=0.26, pruned_loss=0.05448, over 4902.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2447, pruned_loss=0.04969, over 939676.95 frames. ], batch size: 37, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:38:02,328 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155457.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:38:11,935 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.111e+02 1.495e+02 1.724e+02 1.967e+02 3.002e+02, threshold=3.447e+02, percent-clipped=0.0 +2023-03-27 09:38:39,882 INFO [finetune.py:976] (4/7) Epoch 28, batch 850, loss[loss=0.1405, simple_loss=0.218, pruned_loss=0.03156, over 4773.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2422, pruned_loss=0.04859, over 939701.39 frames. ], batch size: 28, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:38:52,643 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155518.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:39:13,741 INFO [finetune.py:976] (4/7) Epoch 28, batch 900, loss[loss=0.1709, simple_loss=0.2321, pruned_loss=0.05482, over 4876.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2398, pruned_loss=0.04792, over 943633.39 frames. ], batch size: 34, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:39:28,252 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.783e+01 1.413e+02 1.787e+02 2.288e+02 4.282e+02, threshold=3.575e+02, percent-clipped=3.0 +2023-03-27 09:39:38,356 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1476, 2.0836, 1.7401, 2.0764, 2.1198, 1.8874, 2.3852, 2.1461], + device='cuda:4'), covar=tensor([0.1372, 0.2064, 0.3019, 0.2637, 0.2704, 0.1701, 0.3132, 0.1748], + device='cuda:4'), in_proj_covar=tensor([0.0190, 0.0190, 0.0237, 0.0255, 0.0250, 0.0208, 0.0215, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:39:54,474 INFO [finetune.py:976] (4/7) Epoch 28, batch 950, loss[loss=0.1561, simple_loss=0.2244, pruned_loss=0.04384, over 4829.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2374, pruned_loss=0.04699, over 947881.57 frames. ], batch size: 25, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:40:02,634 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155605.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:40:11,184 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3409, 1.4661, 1.8095, 1.7258, 1.6171, 3.2434, 1.3345, 1.5578], + device='cuda:4'), covar=tensor([0.1006, 0.1813, 0.1044, 0.0912, 0.1527, 0.0218, 0.1518, 0.1715], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 09:40:20,549 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155632.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:40:31,713 INFO [finetune.py:976] (4/7) Epoch 28, batch 1000, loss[loss=0.1597, simple_loss=0.2239, pruned_loss=0.04772, over 4887.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2407, pruned_loss=0.04841, over 950595.08 frames. ], batch size: 32, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:40:37,329 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155657.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:40:42,788 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155666.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:40:45,714 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.619e+01 1.519e+02 1.814e+02 2.190e+02 3.109e+02, threshold=3.628e+02, percent-clipped=0.0 +2023-03-27 09:40:50,465 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0838, 1.0298, 0.9377, 0.5754, 0.9971, 1.1933, 1.1971, 0.9831], + device='cuda:4'), covar=tensor([0.0856, 0.0612, 0.0603, 0.0506, 0.0566, 0.0685, 0.0409, 0.0749], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0147, 0.0130, 0.0123, 0.0132, 0.0130, 0.0143, 0.0150], + device='cuda:4'), out_proj_covar=tensor([8.8758e-05, 1.0555e-04, 9.2581e-05, 8.6096e-05, 9.2301e-05, 9.2359e-05, + 1.0139e-04, 1.0743e-04], device='cuda:4') +2023-03-27 09:40:54,192 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155680.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:41:13,977 INFO [finetune.py:976] (4/7) Epoch 28, batch 1050, loss[loss=0.1761, simple_loss=0.2499, pruned_loss=0.0511, over 4827.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2438, pruned_loss=0.04914, over 952865.08 frames. ], batch size: 30, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:41:26,785 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155718.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:41:31,001 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=155725.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:41:46,713 INFO [finetune.py:976] (4/7) Epoch 28, batch 1100, loss[loss=0.1799, simple_loss=0.2559, pruned_loss=0.052, over 4701.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2451, pruned_loss=0.04946, over 954307.63 frames. ], batch size: 59, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:41:49,117 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2614, 3.7081, 3.9695, 4.0885, 4.0088, 3.7874, 4.3599, 1.5374], + device='cuda:4'), covar=tensor([0.0803, 0.0866, 0.0796, 0.1070, 0.1280, 0.1473, 0.0661, 0.5647], + device='cuda:4'), in_proj_covar=tensor([0.0352, 0.0246, 0.0283, 0.0297, 0.0335, 0.0287, 0.0305, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:42:01,622 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.104e+02 1.629e+02 1.915e+02 2.256e+02 9.973e+02, threshold=3.830e+02, percent-clipped=2.0 +2023-03-27 09:42:02,898 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=155773.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:42:02,959 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4672, 1.4883, 1.1936, 1.4907, 1.8446, 1.6643, 1.5152, 1.3967], + device='cuda:4'), covar=tensor([0.0340, 0.0328, 0.0593, 0.0308, 0.0187, 0.0490, 0.0310, 0.0403], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0106, 0.0148, 0.0111, 0.0101, 0.0116, 0.0103, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.8543e-05, 8.1141e-05, 1.1533e-04, 8.4781e-05, 7.8351e-05, 8.5137e-05, + 7.6640e-05, 8.6105e-05], device='cuda:4') +2023-03-27 09:42:10,637 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=155784.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:42:21,325 INFO [finetune.py:976] (4/7) Epoch 28, batch 1150, loss[loss=0.1437, simple_loss=0.2151, pruned_loss=0.03612, over 4766.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2451, pruned_loss=0.04903, over 955773.05 frames. ], batch size: 26, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:42:39,762 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155813.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:43:01,212 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=155845.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 09:43:02,858 INFO [finetune.py:976] (4/7) Epoch 28, batch 1200, loss[loss=0.1756, simple_loss=0.2503, pruned_loss=0.05042, over 4929.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2435, pruned_loss=0.04831, over 955784.96 frames. ], batch size: 38, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:43:18,198 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.526e+02 1.746e+02 2.167e+02 3.236e+02, threshold=3.492e+02, percent-clipped=0.0 +2023-03-27 09:43:45,614 INFO [finetune.py:976] (4/7) Epoch 28, batch 1250, loss[loss=0.1868, simple_loss=0.2607, pruned_loss=0.05647, over 4910.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2415, pruned_loss=0.04821, over 956515.34 frames. ], batch size: 32, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:44:01,999 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.12 vs. limit=2.0 +2023-03-27 09:44:15,801 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.90 vs. limit=5.0 +2023-03-27 09:44:19,205 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5497, 1.4407, 1.2964, 1.6443, 1.6233, 1.6622, 1.0275, 1.3482], + device='cuda:4'), covar=tensor([0.2419, 0.2196, 0.2147, 0.1744, 0.1787, 0.1352, 0.2828, 0.2027], + device='cuda:4'), in_proj_covar=tensor([0.0245, 0.0212, 0.0215, 0.0199, 0.0246, 0.0190, 0.0217, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:44:22,108 INFO [finetune.py:976] (4/7) Epoch 28, batch 1300, loss[loss=0.1231, simple_loss=0.196, pruned_loss=0.02509, over 4761.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2379, pruned_loss=0.04711, over 957447.26 frames. ], batch size: 27, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:44:32,015 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=155961.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:44:38,006 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.818e+01 1.476e+02 1.729e+02 2.197e+02 4.050e+02, threshold=3.458e+02, percent-clipped=1.0 +2023-03-27 09:44:55,325 INFO [finetune.py:976] (4/7) Epoch 28, batch 1350, loss[loss=0.1661, simple_loss=0.2426, pruned_loss=0.04479, over 4825.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.239, pruned_loss=0.04728, over 958556.63 frames. ], batch size: 51, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:45:10,152 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156013.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:45:32,753 INFO [finetune.py:976] (4/7) Epoch 28, batch 1400, loss[loss=0.2033, simple_loss=0.2695, pruned_loss=0.06854, over 4886.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2416, pruned_loss=0.04762, over 958074.70 frames. ], batch size: 32, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:45:45,199 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8501, 1.8825, 1.5264, 1.9754, 2.4382, 2.0290, 1.7517, 1.4743], + device='cuda:4'), covar=tensor([0.2094, 0.1739, 0.1810, 0.1509, 0.1501, 0.1103, 0.2204, 0.1886], + device='cuda:4'), in_proj_covar=tensor([0.0245, 0.0212, 0.0216, 0.0199, 0.0246, 0.0190, 0.0217, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:45:48,213 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156070.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:45:48,708 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.532e+02 1.806e+02 2.221e+02 4.474e+02, threshold=3.612e+02, percent-clipped=3.0 +2023-03-27 09:46:01,395 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5702, 1.1326, 0.8256, 1.4377, 2.0297, 0.7476, 1.3883, 1.4064], + device='cuda:4'), covar=tensor([0.1473, 0.2073, 0.1603, 0.1154, 0.1840, 0.1980, 0.1458, 0.1901], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0109, 0.0092, 0.0120, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 09:46:06,169 INFO [finetune.py:976] (4/7) Epoch 28, batch 1450, loss[loss=0.1692, simple_loss=0.2396, pruned_loss=0.0494, over 4790.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2429, pruned_loss=0.04786, over 956549.34 frames. ], batch size: 25, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:46:23,136 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156113.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:46:26,611 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1254, 2.1444, 1.7662, 2.2377, 2.0579, 2.0569, 2.0197, 2.8276], + device='cuda:4'), covar=tensor([0.3847, 0.4565, 0.3388, 0.4009, 0.4363, 0.2443, 0.4386, 0.1601], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0264, 0.0237, 0.0275, 0.0260, 0.0230, 0.0259, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:46:35,155 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=156131.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 09:46:40,536 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156140.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 09:46:45,776 INFO [finetune.py:976] (4/7) Epoch 28, batch 1500, loss[loss=0.1701, simple_loss=0.238, pruned_loss=0.05106, over 4820.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2445, pruned_loss=0.04907, over 957822.29 frames. ], batch size: 30, lr: 2.89e-03, grad_scale: 64.0 +2023-03-27 09:46:54,177 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156161.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:47:02,075 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.065e+02 1.585e+02 1.899e+02 2.331e+02 3.577e+02, threshold=3.799e+02, percent-clipped=0.0 +2023-03-27 09:47:18,929 INFO [finetune.py:976] (4/7) Epoch 28, batch 1550, loss[loss=0.1585, simple_loss=0.2321, pruned_loss=0.04241, over 4902.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2429, pruned_loss=0.04877, over 956835.66 frames. ], batch size: 43, lr: 2.89e-03, grad_scale: 32.0 +2023-03-27 09:47:30,566 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156216.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:47:56,396 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0109, 1.6678, 2.3743, 1.6061, 2.0434, 2.1786, 1.6332, 2.3586], + device='cuda:4'), covar=tensor([0.1155, 0.2083, 0.1278, 0.1909, 0.0914, 0.1367, 0.2765, 0.0780], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0208, 0.0193, 0.0190, 0.0175, 0.0213, 0.0219, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:47:59,290 INFO [finetune.py:976] (4/7) Epoch 28, batch 1600, loss[loss=0.1411, simple_loss=0.209, pruned_loss=0.03661, over 4764.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2398, pruned_loss=0.04747, over 958046.38 frames. ], batch size: 26, lr: 2.89e-03, grad_scale: 32.0 +2023-03-27 09:48:08,658 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156261.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:48:15,839 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.043e+02 1.435e+02 1.767e+02 2.111e+02 3.704e+02, threshold=3.535e+02, percent-clipped=0.0 +2023-03-27 09:48:19,970 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=156277.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:48:27,883 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156290.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:48:32,610 INFO [finetune.py:976] (4/7) Epoch 28, batch 1650, loss[loss=0.1614, simple_loss=0.2342, pruned_loss=0.04432, over 4753.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.236, pruned_loss=0.04582, over 957235.46 frames. ], batch size: 54, lr: 2.89e-03, grad_scale: 32.0 +2023-03-27 09:48:40,818 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156309.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:48:43,249 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156313.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:49:18,368 INFO [finetune.py:976] (4/7) Epoch 28, batch 1700, loss[loss=0.185, simple_loss=0.2593, pruned_loss=0.05534, over 4920.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2354, pruned_loss=0.04579, over 958041.39 frames. ], batch size: 37, lr: 2.88e-03, grad_scale: 32.0 +2023-03-27 09:49:20,343 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=156351.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:49:26,384 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7908, 1.3622, 0.8762, 1.7553, 2.3165, 1.3262, 1.8476, 1.5727], + device='cuda:4'), covar=tensor([0.1509, 0.2031, 0.1781, 0.1177, 0.1828, 0.1829, 0.1251, 0.2115], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0110, 0.0092, 0.0120, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 09:49:27,454 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156361.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:49:34,437 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.260e+01 1.525e+02 1.796e+02 2.204e+02 4.546e+02, threshold=3.593e+02, percent-clipped=3.0 +2023-03-27 09:49:51,285 INFO [finetune.py:976] (4/7) Epoch 28, batch 1750, loss[loss=0.1563, simple_loss=0.2406, pruned_loss=0.03597, over 4845.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2385, pruned_loss=0.04703, over 957692.25 frames. ], batch size: 47, lr: 2.88e-03, grad_scale: 32.0 +2023-03-27 09:50:09,511 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156426.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 09:50:14,090 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6713, 3.6414, 3.4186, 1.7778, 3.7780, 2.8858, 1.0626, 2.5257], + device='cuda:4'), covar=tensor([0.2743, 0.1934, 0.1547, 0.3462, 0.1138, 0.1012, 0.4185, 0.1631], + device='cuda:4'), in_proj_covar=tensor([0.0149, 0.0178, 0.0158, 0.0129, 0.0162, 0.0122, 0.0147, 0.0124], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 09:50:19,439 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156440.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 09:50:24,159 INFO [finetune.py:976] (4/7) Epoch 28, batch 1800, loss[loss=0.1788, simple_loss=0.2245, pruned_loss=0.06652, over 4295.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2415, pruned_loss=0.04818, over 957370.69 frames. ], batch size: 18, lr: 2.88e-03, grad_scale: 32.0 +2023-03-27 09:50:39,988 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.522e+02 1.851e+02 2.291e+02 4.651e+02, threshold=3.702e+02, percent-clipped=5.0 +2023-03-27 09:50:51,504 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156488.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:50:57,545 INFO [finetune.py:976] (4/7) Epoch 28, batch 1850, loss[loss=0.1517, simple_loss=0.2217, pruned_loss=0.04085, over 4782.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.243, pruned_loss=0.04885, over 955018.80 frames. ], batch size: 26, lr: 2.88e-03, grad_scale: 32.0 +2023-03-27 09:50:59,517 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7844, 2.5041, 2.0875, 1.0534, 2.3060, 2.2019, 2.0002, 2.2788], + device='cuda:4'), covar=tensor([0.0842, 0.0920, 0.1496, 0.1955, 0.1378, 0.2025, 0.2060, 0.0982], + device='cuda:4'), in_proj_covar=tensor([0.0172, 0.0191, 0.0201, 0.0181, 0.0210, 0.0210, 0.0223, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:51:40,565 INFO [finetune.py:976] (4/7) Epoch 28, batch 1900, loss[loss=0.1563, simple_loss=0.2288, pruned_loss=0.04191, over 4780.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2451, pruned_loss=0.04943, over 955802.98 frames. ], batch size: 26, lr: 2.88e-03, grad_scale: 32.0 +2023-03-27 09:51:47,675 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.03 vs. limit=5.0 +2023-03-27 09:51:56,029 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.831e+01 1.540e+02 1.871e+02 2.242e+02 4.934e+02, threshold=3.741e+02, percent-clipped=1.0 +2023-03-27 09:51:56,112 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156572.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:52:13,684 INFO [finetune.py:976] (4/7) Epoch 28, batch 1950, loss[loss=0.2063, simple_loss=0.2599, pruned_loss=0.07634, over 4845.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.244, pruned_loss=0.04943, over 955208.77 frames. ], batch size: 49, lr: 2.88e-03, grad_scale: 32.0 +2023-03-27 09:52:20,499 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-27 09:52:46,374 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=156646.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:52:47,510 INFO [finetune.py:976] (4/7) Epoch 28, batch 2000, loss[loss=0.1942, simple_loss=0.2494, pruned_loss=0.0695, over 4872.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2413, pruned_loss=0.04871, over 954877.10 frames. ], batch size: 31, lr: 2.88e-03, grad_scale: 32.0 +2023-03-27 09:53:04,320 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.919e+01 1.529e+02 1.760e+02 2.169e+02 4.761e+02, threshold=3.520e+02, percent-clipped=1.0 +2023-03-27 09:53:29,982 INFO [finetune.py:976] (4/7) Epoch 28, batch 2050, loss[loss=0.1513, simple_loss=0.2227, pruned_loss=0.04, over 4912.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2374, pruned_loss=0.04724, over 955811.21 frames. ], batch size: 36, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:53:30,052 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5873, 3.9013, 3.6957, 1.8729, 4.0418, 2.9254, 0.8250, 2.7548], + device='cuda:4'), covar=tensor([0.2661, 0.2238, 0.1556, 0.3426, 0.1028, 0.1011, 0.4761, 0.1628], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0179, 0.0159, 0.0129, 0.0163, 0.0123, 0.0148, 0.0125], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 09:53:31,841 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9502, 1.8333, 1.6167, 2.0275, 2.4978, 2.1162, 1.9666, 1.6048], + device='cuda:4'), covar=tensor([0.2075, 0.1868, 0.1866, 0.1577, 0.1574, 0.1092, 0.1999, 0.1885], + device='cuda:4'), in_proj_covar=tensor([0.0245, 0.0211, 0.0215, 0.0199, 0.0244, 0.0190, 0.0215, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:53:47,944 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156726.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 09:54:08,957 INFO [finetune.py:976] (4/7) Epoch 28, batch 2100, loss[loss=0.1788, simple_loss=0.2586, pruned_loss=0.04953, over 4818.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2394, pruned_loss=0.04858, over 954228.50 frames. ], batch size: 45, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:54:09,691 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156749.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:54:37,772 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.070e+02 1.521e+02 1.844e+02 2.179e+02 3.224e+02, threshold=3.687e+02, percent-clipped=0.0 +2023-03-27 09:54:38,476 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156774.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:54:47,436 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.34 vs. limit=5.0 +2023-03-27 09:54:54,965 INFO [finetune.py:976] (4/7) Epoch 28, batch 2150, loss[loss=0.1918, simple_loss=0.2686, pruned_loss=0.05754, over 4835.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2425, pruned_loss=0.04941, over 954984.66 frames. ], batch size: 47, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:55:02,854 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2691, 1.5292, 0.8501, 2.0706, 2.4499, 1.9062, 1.8581, 1.9567], + device='cuda:4'), covar=tensor([0.1461, 0.2013, 0.2036, 0.1190, 0.1961, 0.1746, 0.1409, 0.1968], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0109, 0.0092, 0.0120, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 09:55:03,481 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=156810.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:55:23,295 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.61 vs. limit=5.0 +2023-03-27 09:55:27,788 INFO [finetune.py:976] (4/7) Epoch 28, batch 2200, loss[loss=0.1905, simple_loss=0.2638, pruned_loss=0.05859, over 4803.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2447, pruned_loss=0.04994, over 954066.63 frames. ], batch size: 41, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:55:36,800 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9896, 1.6711, 2.1174, 1.4974, 1.9705, 2.1585, 2.1491, 1.3712], + device='cuda:4'), covar=tensor([0.0751, 0.1003, 0.0705, 0.0909, 0.0851, 0.0706, 0.0657, 0.1854], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0138, 0.0140, 0.0118, 0.0128, 0.0138, 0.0139, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:55:44,124 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.0183, 4.7190, 4.4364, 2.7107, 4.8445, 3.5723, 1.1416, 3.4194], + device='cuda:4'), covar=tensor([0.2073, 0.1541, 0.1235, 0.2613, 0.0733, 0.0767, 0.3970, 0.1174], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0178, 0.0159, 0.0129, 0.0162, 0.0123, 0.0148, 0.0125], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 09:55:44,156 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156872.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:55:44,654 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.040e+02 1.543e+02 1.740e+02 2.127e+02 4.555e+02, threshold=3.480e+02, percent-clipped=1.0 +2023-03-27 09:55:52,563 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=156885.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:56:01,289 INFO [finetune.py:976] (4/7) Epoch 28, batch 2250, loss[loss=0.159, simple_loss=0.2433, pruned_loss=0.03732, over 4866.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2468, pruned_loss=0.05062, over 954263.12 frames. ], batch size: 34, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:56:15,019 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6032, 3.6000, 3.3844, 1.6026, 3.7479, 2.7102, 0.8082, 2.4907], + device='cuda:4'), covar=tensor([0.2376, 0.1725, 0.1592, 0.3254, 0.1051, 0.0996, 0.4169, 0.1474], + device='cuda:4'), in_proj_covar=tensor([0.0150, 0.0178, 0.0159, 0.0129, 0.0162, 0.0123, 0.0148, 0.0125], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 09:56:16,207 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156920.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:56:31,027 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.7339, 4.1387, 4.3087, 4.5271, 4.4871, 4.2189, 4.8334, 1.7192], + device='cuda:4'), covar=tensor([0.0737, 0.0848, 0.0804, 0.0812, 0.1138, 0.1612, 0.0614, 0.5598], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0244, 0.0282, 0.0293, 0.0333, 0.0285, 0.0304, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:56:32,863 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=156946.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:56:32,881 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=156946.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:56:34,000 INFO [finetune.py:976] (4/7) Epoch 28, batch 2300, loss[loss=0.1507, simple_loss=0.2403, pruned_loss=0.03055, over 4845.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2466, pruned_loss=0.05008, over 955389.01 frames. ], batch size: 49, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:57:00,120 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.437e+02 1.655e+02 2.039e+02 3.893e+02, threshold=3.311e+02, percent-clipped=1.0 +2023-03-27 09:57:10,963 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-27 09:57:17,549 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=156994.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:57:20,397 INFO [finetune.py:976] (4/7) Epoch 28, batch 2350, loss[loss=0.1467, simple_loss=0.211, pruned_loss=0.04113, over 4780.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2432, pruned_loss=0.04878, over 954117.21 frames. ], batch size: 29, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:57:20,569 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-27 09:57:28,243 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157010.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:57:45,924 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.3266, 2.9710, 3.0587, 3.2676, 3.0944, 2.9093, 3.3581, 0.8866], + device='cuda:4'), covar=tensor([0.1152, 0.1009, 0.1186, 0.1119, 0.1758, 0.1888, 0.1129, 0.6004], + device='cuda:4'), in_proj_covar=tensor([0.0352, 0.0245, 0.0284, 0.0294, 0.0335, 0.0287, 0.0305, 0.0304], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:57:52,968 INFO [finetune.py:976] (4/7) Epoch 28, batch 2400, loss[loss=0.1604, simple_loss=0.2347, pruned_loss=0.04302, over 4906.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2404, pruned_loss=0.04775, over 954871.72 frames. ], batch size: 36, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:58:08,924 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157071.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:58:12,823 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.019e+02 1.490e+02 1.799e+02 2.218e+02 3.254e+02, threshold=3.597e+02, percent-clipped=0.0 +2023-03-27 09:58:23,940 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5319, 1.5035, 1.3339, 1.4789, 1.8022, 1.7592, 1.5784, 1.3498], + device='cuda:4'), covar=tensor([0.0348, 0.0326, 0.0629, 0.0327, 0.0212, 0.0446, 0.0357, 0.0421], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0106, 0.0149, 0.0112, 0.0102, 0.0116, 0.0104, 0.0114], + device='cuda:4'), out_proj_covar=tensor([7.8704e-05, 8.1352e-05, 1.1575e-04, 8.5421e-05, 7.8705e-05, 8.5549e-05, + 7.7365e-05, 8.6344e-05], device='cuda:4') +2023-03-27 09:58:28,590 INFO [finetune.py:976] (4/7) Epoch 28, batch 2450, loss[loss=0.1574, simple_loss=0.2397, pruned_loss=0.03757, over 4910.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2379, pruned_loss=0.04695, over 956313.88 frames. ], batch size: 35, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:58:33,537 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157105.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:58:34,805 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6798, 1.5499, 1.4957, 1.6246, 1.2728, 3.6600, 1.4423, 1.9006], + device='cuda:4'), covar=tensor([0.3409, 0.2634, 0.2247, 0.2435, 0.1718, 0.0179, 0.2662, 0.1239], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0124, 0.0114, 0.0096, 0.0094, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 09:58:57,712 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157142.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 09:59:01,795 INFO [finetune.py:976] (4/7) Epoch 28, batch 2500, loss[loss=0.1561, simple_loss=0.2393, pruned_loss=0.03645, over 4826.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2393, pruned_loss=0.04719, over 954098.59 frames. ], batch size: 39, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:59:23,962 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.586e+02 1.871e+02 2.257e+02 5.817e+02, threshold=3.742e+02, percent-clipped=3.0 +2023-03-27 09:59:35,830 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2708, 3.6886, 3.9190, 4.1137, 4.0463, 3.7801, 4.3429, 1.3282], + device='cuda:4'), covar=tensor([0.0816, 0.0865, 0.0898, 0.0953, 0.1228, 0.1531, 0.0666, 0.6211], + device='cuda:4'), in_proj_covar=tensor([0.0352, 0.0246, 0.0284, 0.0294, 0.0335, 0.0286, 0.0305, 0.0303], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 09:59:51,558 INFO [finetune.py:976] (4/7) Epoch 28, batch 2550, loss[loss=0.1415, simple_loss=0.2348, pruned_loss=0.02407, over 4778.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2417, pruned_loss=0.0474, over 954697.95 frames. ], batch size: 28, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 09:59:55,365 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157203.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:00:20,725 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157241.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:00:24,904 INFO [finetune.py:976] (4/7) Epoch 28, batch 2600, loss[loss=0.1701, simple_loss=0.2461, pruned_loss=0.04705, over 4824.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.244, pruned_loss=0.04864, over 956416.10 frames. ], batch size: 39, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:00:34,239 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3337, 1.3855, 1.8654, 1.7312, 1.4843, 3.1656, 1.1901, 1.5342], + device='cuda:4'), covar=tensor([0.1020, 0.1873, 0.1275, 0.0959, 0.1622, 0.0275, 0.1604, 0.1846], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0074, 0.0076, 0.0091, 0.0080, 0.0086, 0.0081], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 10:00:41,901 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.034e+02 1.561e+02 1.846e+02 2.212e+02 4.271e+02, threshold=3.692e+02, percent-clipped=1.0 +2023-03-27 10:00:57,757 INFO [finetune.py:976] (4/7) Epoch 28, batch 2650, loss[loss=0.1702, simple_loss=0.2528, pruned_loss=0.04384, over 4728.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2452, pruned_loss=0.04921, over 957142.07 frames. ], batch size: 59, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:01:14,925 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9260, 1.2939, 1.9829, 1.9266, 1.7520, 1.7313, 1.8081, 1.8909], + device='cuda:4'), covar=tensor([0.4276, 0.4052, 0.3404, 0.3812, 0.5324, 0.4022, 0.5059, 0.3228], + device='cuda:4'), in_proj_covar=tensor([0.0269, 0.0250, 0.0270, 0.0299, 0.0299, 0.0274, 0.0304, 0.0254], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:01:30,664 INFO [finetune.py:976] (4/7) Epoch 28, batch 2700, loss[loss=0.1472, simple_loss=0.2066, pruned_loss=0.04385, over 4227.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2439, pruned_loss=0.04839, over 956022.97 frames. ], batch size: 18, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:01:35,644 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157356.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:01:42,597 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157366.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:01:47,169 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.080e+02 1.455e+02 1.749e+02 2.146e+02 4.370e+02, threshold=3.498e+02, percent-clipped=1.0 +2023-03-27 10:01:53,267 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157382.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:02:12,808 INFO [finetune.py:976] (4/7) Epoch 28, batch 2750, loss[loss=0.2253, simple_loss=0.2792, pruned_loss=0.08576, over 4806.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2422, pruned_loss=0.04832, over 954313.06 frames. ], batch size: 45, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:02:20,322 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157404.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 10:02:20,951 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=157405.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:02:21,572 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8955, 1.7845, 1.5999, 2.0033, 2.1396, 1.9899, 1.5190, 1.6092], + device='cuda:4'), covar=tensor([0.2002, 0.1918, 0.1876, 0.1533, 0.1641, 0.1197, 0.2434, 0.1759], + device='cuda:4'), in_proj_covar=tensor([0.0247, 0.0212, 0.0216, 0.0200, 0.0246, 0.0191, 0.0217, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:02:22,854 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.88 vs. limit=5.0 +2023-03-27 10:02:29,175 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157417.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:02:33,288 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9507, 1.8349, 1.6184, 1.5040, 1.9603, 1.6862, 1.8753, 1.9689], + device='cuda:4'), covar=tensor([0.1314, 0.1979, 0.2958, 0.2426, 0.2660, 0.1711, 0.2691, 0.1665], + device='cuda:4'), in_proj_covar=tensor([0.0190, 0.0191, 0.0238, 0.0255, 0.0251, 0.0208, 0.0215, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:02:37,516 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7181, 3.8131, 3.5428, 1.5383, 3.8126, 2.9583, 1.1718, 2.6326], + device='cuda:4'), covar=tensor([0.2009, 0.1785, 0.1621, 0.3592, 0.1091, 0.0907, 0.3942, 0.1616], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0180, 0.0160, 0.0130, 0.0164, 0.0124, 0.0149, 0.0126], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 10:02:46,869 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157443.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:02:50,279 INFO [finetune.py:976] (4/7) Epoch 28, batch 2800, loss[loss=0.149, simple_loss=0.2208, pruned_loss=0.03862, over 4934.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2389, pruned_loss=0.0474, over 953972.53 frames. ], batch size: 38, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:02:53,300 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=157453.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:02:53,353 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6127, 1.5321, 1.4798, 1.5226, 1.0715, 2.9395, 1.0961, 1.5601], + device='cuda:4'), covar=tensor([0.3205, 0.2616, 0.2171, 0.2353, 0.1738, 0.0286, 0.2649, 0.1249], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0116, 0.0121, 0.0124, 0.0114, 0.0095, 0.0094, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 10:02:57,187 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-03-27 10:03:01,076 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157465.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 10:03:06,316 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.288e+01 1.520e+02 1.688e+02 2.071e+02 7.416e+02, threshold=3.376e+02, percent-clipped=4.0 +2023-03-27 10:03:18,783 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9593, 2.1452, 1.7310, 1.9602, 2.5152, 2.5099, 2.0999, 1.9701], + device='cuda:4'), covar=tensor([0.0384, 0.0336, 0.0617, 0.0349, 0.0276, 0.0657, 0.0457, 0.0438], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0106, 0.0148, 0.0111, 0.0102, 0.0116, 0.0104, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.8687e-05, 8.0964e-05, 1.1521e-04, 8.5012e-05, 7.8758e-05, 8.5260e-05, + 7.7045e-05, 8.5938e-05], device='cuda:4') +2023-03-27 10:03:23,414 INFO [finetune.py:976] (4/7) Epoch 28, batch 2850, loss[loss=0.163, simple_loss=0.2434, pruned_loss=0.04133, over 4778.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2371, pruned_loss=0.04709, over 954130.66 frames. ], batch size: 29, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:03:23,482 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157498.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:03:28,872 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157506.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 10:03:52,480 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=157541.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:03:57,079 INFO [finetune.py:976] (4/7) Epoch 28, batch 2900, loss[loss=0.179, simple_loss=0.2652, pruned_loss=0.04636, over 4840.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2409, pruned_loss=0.04842, over 953914.80 frames. ], batch size: 47, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:04:09,767 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157567.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 10:04:10,386 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157568.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:04:13,241 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.581e+01 1.484e+02 1.768e+02 2.098e+02 4.175e+02, threshold=3.535e+02, percent-clipped=1.0 +2023-03-27 10:04:13,380 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.5087, 1.3435, 1.3987, 0.8126, 1.5349, 1.7143, 1.6650, 1.2937], + device='cuda:4'), covar=tensor([0.0917, 0.0830, 0.0570, 0.0601, 0.0514, 0.0606, 0.0403, 0.0807], + device='cuda:4'), in_proj_covar=tensor([0.0122, 0.0149, 0.0130, 0.0124, 0.0133, 0.0131, 0.0143, 0.0152], + device='cuda:4'), out_proj_covar=tensor([8.9097e-05, 1.0690e-04, 9.2674e-05, 8.6739e-05, 9.3137e-05, 9.2957e-05, + 1.0175e-04, 1.0851e-04], device='cuda:4') +2023-03-27 10:04:24,445 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=157589.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:04:32,418 INFO [finetune.py:976] (4/7) Epoch 28, batch 2950, loss[loss=0.1832, simple_loss=0.2533, pruned_loss=0.05652, over 4817.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2444, pruned_loss=0.04923, over 955619.22 frames. ], batch size: 33, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:05:06,832 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157629.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:05:18,354 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-27 10:05:23,714 INFO [finetune.py:976] (4/7) Epoch 28, batch 3000, loss[loss=0.2111, simple_loss=0.2667, pruned_loss=0.07772, over 4275.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2451, pruned_loss=0.04985, over 956228.42 frames. ], batch size: 65, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:05:23,714 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 10:05:27,153 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7851, 1.5009, 1.1073, 1.8087, 2.0823, 1.3938, 1.7120, 1.7360], + device='cuda:4'), covar=tensor([0.1131, 0.1433, 0.1374, 0.0832, 0.1579, 0.1596, 0.0961, 0.1496], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0093, 0.0108, 0.0092, 0.0119, 0.0092, 0.0097, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 10:05:29,884 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.7259, 3.3178, 3.4495, 3.5907, 3.4999, 3.3038, 3.7559, 1.4795], + device='cuda:4'), covar=tensor([0.0732, 0.0748, 0.0802, 0.0851, 0.1114, 0.1429, 0.0692, 0.4885], + device='cuda:4'), in_proj_covar=tensor([0.0351, 0.0245, 0.0285, 0.0294, 0.0334, 0.0285, 0.0305, 0.0303], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:05:34,508 INFO [finetune.py:1010] (4/7) Epoch 28, validation: loss=0.1567, simple_loss=0.2243, pruned_loss=0.04455, over 2265189.00 frames. +2023-03-27 10:05:34,509 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6557MB +2023-03-27 10:05:46,487 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=157666.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:05:50,618 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.041e+02 1.524e+02 1.841e+02 2.248e+02 4.082e+02, threshold=3.682e+02, percent-clipped=3.0 +2023-03-27 10:06:07,190 INFO [finetune.py:976] (4/7) Epoch 28, batch 3050, loss[loss=0.1813, simple_loss=0.2508, pruned_loss=0.05588, over 4907.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.246, pruned_loss=0.04986, over 957183.11 frames. ], batch size: 36, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:06:16,640 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157712.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:06:17,821 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=157714.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:06:26,138 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2638, 2.0906, 2.8186, 4.2262, 3.0650, 2.9022, 1.0033, 3.6081], + device='cuda:4'), covar=tensor([0.1432, 0.1069, 0.1043, 0.0433, 0.0601, 0.1315, 0.1748, 0.0311], + device='cuda:4'), in_proj_covar=tensor([0.0098, 0.0115, 0.0132, 0.0163, 0.0100, 0.0134, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 10:06:33,278 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157738.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:06:40,235 INFO [finetune.py:976] (4/7) Epoch 28, batch 3100, loss[loss=0.1166, simple_loss=0.193, pruned_loss=0.0201, over 4769.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2437, pruned_loss=0.04895, over 955449.59 frames. ], batch size: 28, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:06:48,818 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157760.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 10:06:57,050 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.389e+02 1.744e+02 2.105e+02 3.209e+02, threshold=3.488e+02, percent-clipped=0.0 +2023-03-27 10:07:19,510 INFO [finetune.py:976] (4/7) Epoch 28, batch 3150, loss[loss=0.1762, simple_loss=0.2425, pruned_loss=0.05498, over 4747.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2409, pruned_loss=0.04783, over 956179.54 frames. ], batch size: 27, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:07:20,085 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=157798.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:07:56,803 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7087, 1.6225, 1.5987, 1.6620, 1.1032, 3.2862, 1.3827, 1.7610], + device='cuda:4'), covar=tensor([0.3254, 0.2660, 0.2103, 0.2402, 0.1894, 0.0249, 0.2629, 0.1246], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0116, 0.0120, 0.0124, 0.0114, 0.0095, 0.0093, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 10:08:04,471 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=157846.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:08:05,669 INFO [finetune.py:976] (4/7) Epoch 28, batch 3200, loss[loss=0.1524, simple_loss=0.2188, pruned_loss=0.04302, over 4860.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2378, pruned_loss=0.04736, over 956949.14 frames. ], batch size: 31, lr: 2.88e-03, grad_scale: 16.0 +2023-03-27 10:08:09,372 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=157853.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:08:15,278 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157862.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 10:08:22,855 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.817e+01 1.555e+02 1.831e+02 2.254e+02 7.078e+02, threshold=3.662e+02, percent-clipped=7.0 +2023-03-27 10:08:38,489 INFO [finetune.py:976] (4/7) Epoch 28, batch 3250, loss[loss=0.1874, simple_loss=0.249, pruned_loss=0.06288, over 4823.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2392, pruned_loss=0.04817, over 956578.05 frames. ], batch size: 33, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:08:49,790 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9890, 4.6827, 4.3831, 2.3959, 4.8654, 3.6598, 1.0140, 3.3529], + device='cuda:4'), covar=tensor([0.2148, 0.1864, 0.1411, 0.2825, 0.0748, 0.0766, 0.4352, 0.1224], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0181, 0.0161, 0.0131, 0.0165, 0.0125, 0.0151, 0.0127], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 10:08:49,835 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=157914.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:08:56,862 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=157924.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:09:11,706 INFO [finetune.py:976] (4/7) Epoch 28, batch 3300, loss[loss=0.1853, simple_loss=0.2582, pruned_loss=0.05621, over 4175.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2432, pruned_loss=0.04942, over 955827.12 frames. ], batch size: 65, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:09:12,304 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5482, 1.3573, 1.2498, 1.5033, 1.7277, 1.5774, 1.1807, 1.3267], + device='cuda:4'), covar=tensor([0.1992, 0.1927, 0.1864, 0.1600, 0.1512, 0.1157, 0.2217, 0.1736], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0212, 0.0215, 0.0199, 0.0245, 0.0190, 0.0217, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:09:29,137 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.081e+02 1.521e+02 1.758e+02 2.140e+02 4.599e+02, threshold=3.515e+02, percent-clipped=1.0 +2023-03-27 10:09:44,716 INFO [finetune.py:976] (4/7) Epoch 28, batch 3350, loss[loss=0.1427, simple_loss=0.2184, pruned_loss=0.03354, over 4853.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2457, pruned_loss=0.0503, over 957271.93 frames. ], batch size: 31, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:09:58,153 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158012.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:10:08,968 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5581, 1.4541, 1.2804, 1.6227, 1.6718, 1.6438, 1.0904, 1.3438], + device='cuda:4'), covar=tensor([0.2289, 0.2039, 0.2067, 0.1769, 0.1645, 0.1252, 0.2317, 0.1966], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0212, 0.0215, 0.0199, 0.0245, 0.0190, 0.0218, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:10:18,852 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7236, 3.9322, 3.6581, 1.8568, 4.1163, 2.9912, 0.8157, 2.8000], + device='cuda:4'), covar=tensor([0.2161, 0.1876, 0.1504, 0.3303, 0.0934, 0.0942, 0.4583, 0.1380], + device='cuda:4'), in_proj_covar=tensor([0.0153, 0.0181, 0.0161, 0.0131, 0.0165, 0.0125, 0.0151, 0.0127], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 10:10:32,988 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158038.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:10:38,997 INFO [finetune.py:976] (4/7) Epoch 28, batch 3400, loss[loss=0.184, simple_loss=0.2575, pruned_loss=0.05523, over 4902.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2468, pruned_loss=0.05038, over 957039.99 frames. ], batch size: 36, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:10:46,888 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158060.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:10:46,929 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158060.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 10:10:55,549 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.103e+02 1.550e+02 1.897e+02 2.350e+02 3.360e+02, threshold=3.793e+02, percent-clipped=0.0 +2023-03-27 10:11:04,976 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158086.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:11:06,885 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6331, 2.4811, 1.9960, 2.8192, 2.6004, 2.1753, 3.0585, 2.6456], + device='cuda:4'), covar=tensor([0.1366, 0.2259, 0.3091, 0.2490, 0.2541, 0.1830, 0.2938, 0.1821], + device='cuda:4'), in_proj_covar=tensor([0.0189, 0.0189, 0.0236, 0.0253, 0.0250, 0.0206, 0.0213, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:11:12,078 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6681, 1.5391, 2.2548, 3.5035, 2.2901, 2.4865, 1.1122, 2.9250], + device='cuda:4'), covar=tensor([0.1678, 0.1295, 0.1212, 0.0519, 0.0804, 0.1305, 0.1735, 0.0425], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0115, 0.0132, 0.0163, 0.0100, 0.0135, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 10:11:12,582 INFO [finetune.py:976] (4/7) Epoch 28, batch 3450, loss[loss=0.1755, simple_loss=0.2448, pruned_loss=0.05304, over 4925.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2465, pruned_loss=0.04994, over 956800.97 frames. ], batch size: 33, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:11:18,612 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158108.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 10:11:28,437 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158122.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:11:40,119 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7184, 1.5013, 1.4503, 1.5829, 1.9173, 1.9047, 1.6967, 1.4409], + device='cuda:4'), covar=tensor([0.0421, 0.0377, 0.0597, 0.0355, 0.0265, 0.0450, 0.0340, 0.0429], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0105, 0.0147, 0.0111, 0.0101, 0.0115, 0.0103, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.8087e-05, 8.0550e-05, 1.1438e-04, 8.4481e-05, 7.8204e-05, 8.4921e-05, + 7.6585e-05, 8.5550e-05], device='cuda:4') +2023-03-27 10:11:46,009 INFO [finetune.py:976] (4/7) Epoch 28, batch 3500, loss[loss=0.1589, simple_loss=0.2239, pruned_loss=0.047, over 4754.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2435, pruned_loss=0.04888, over 956045.37 frames. ], batch size: 23, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:11:55,066 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158162.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 10:12:02,510 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.432e+02 1.800e+02 2.074e+02 3.411e+02, threshold=3.600e+02, percent-clipped=0.0 +2023-03-27 10:12:09,607 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158183.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:12:15,736 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-27 10:12:19,503 INFO [finetune.py:976] (4/7) Epoch 28, batch 3550, loss[loss=0.1316, simple_loss=0.2034, pruned_loss=0.02993, over 4749.00 frames. ], tot_loss[loss=0.169, simple_loss=0.241, pruned_loss=0.04853, over 955134.00 frames. ], batch size: 54, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:12:28,585 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=158209.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:12:29,210 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158210.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 10:12:38,728 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158224.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:12:47,365 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5459, 1.5554, 1.3797, 1.6369, 1.9965, 1.8507, 1.6950, 1.4488], + device='cuda:4'), covar=tensor([0.0384, 0.0340, 0.0594, 0.0280, 0.0196, 0.0516, 0.0320, 0.0401], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0106, 0.0147, 0.0111, 0.0101, 0.0116, 0.0103, 0.0113], + device='cuda:4'), out_proj_covar=tensor([7.8456e-05, 8.0790e-05, 1.1450e-04, 8.4616e-05, 7.8321e-05, 8.5153e-05, + 7.6449e-05, 8.5715e-05], device='cuda:4') +2023-03-27 10:12:52,855 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.33 vs. limit=5.0 +2023-03-27 10:13:02,525 INFO [finetune.py:976] (4/7) Epoch 28, batch 3600, loss[loss=0.132, simple_loss=0.2078, pruned_loss=0.02811, over 4765.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2381, pruned_loss=0.04733, over 955609.41 frames. ], batch size: 28, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:13:18,164 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158272.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:13:18,717 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.065e+02 1.425e+02 1.679e+02 2.016e+02 3.584e+02, threshold=3.358e+02, percent-clipped=0.0 +2023-03-27 10:13:24,065 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6304, 1.6591, 1.5111, 1.6472, 1.5559, 4.2591, 1.5974, 1.8589], + device='cuda:4'), covar=tensor([0.3624, 0.2631, 0.2301, 0.2504, 0.1621, 0.0148, 0.2761, 0.1308], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0117, 0.0121, 0.0124, 0.0114, 0.0095, 0.0094, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 10:13:36,306 INFO [finetune.py:976] (4/7) Epoch 28, batch 3650, loss[loss=0.2094, simple_loss=0.2734, pruned_loss=0.0727, over 4804.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2405, pruned_loss=0.04862, over 953893.86 frames. ], batch size: 41, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:14:10,063 INFO [finetune.py:976] (4/7) Epoch 28, batch 3700, loss[loss=0.1324, simple_loss=0.2184, pruned_loss=0.0232, over 4784.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2432, pruned_loss=0.04909, over 951289.83 frames. ], batch size: 25, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:14:17,898 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.6321, 3.3079, 2.9880, 1.4763, 3.2029, 2.6245, 2.5772, 3.0248], + device='cuda:4'), covar=tensor([0.0757, 0.0746, 0.1402, 0.2055, 0.1329, 0.1966, 0.1797, 0.0895], + device='cuda:4'), in_proj_covar=tensor([0.0172, 0.0191, 0.0201, 0.0181, 0.0211, 0.0210, 0.0224, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:14:24,555 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4198, 1.5130, 1.2429, 1.5159, 1.7922, 1.7623, 1.5850, 1.4215], + device='cuda:4'), covar=tensor([0.0409, 0.0330, 0.0656, 0.0319, 0.0254, 0.0543, 0.0291, 0.0426], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0107, 0.0148, 0.0112, 0.0102, 0.0117, 0.0104, 0.0114], + device='cuda:4'), out_proj_covar=tensor([7.9032e-05, 8.1460e-05, 1.1550e-04, 8.5209e-05, 7.8981e-05, 8.5817e-05, + 7.7274e-05, 8.6567e-05], device='cuda:4') +2023-03-27 10:14:26,132 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.573e+02 1.954e+02 2.338e+02 5.991e+02, threshold=3.909e+02, percent-clipped=5.0 +2023-03-27 10:14:32,289 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.35 vs. limit=5.0 +2023-03-27 10:14:33,428 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158384.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:14:36,938 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0792, 1.8699, 1.7896, 1.8415, 2.2622, 2.3019, 1.9822, 1.7933], + device='cuda:4'), covar=tensor([0.0310, 0.0337, 0.0547, 0.0329, 0.0205, 0.0417, 0.0287, 0.0417], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0107, 0.0149, 0.0112, 0.0102, 0.0117, 0.0104, 0.0114], + device='cuda:4'), out_proj_covar=tensor([7.9186e-05, 8.1571e-05, 1.1572e-04, 8.5357e-05, 7.9066e-05, 8.5917e-05, + 7.7399e-05, 8.6734e-05], device='cuda:4') +2023-03-27 10:14:43,251 INFO [finetune.py:976] (4/7) Epoch 28, batch 3750, loss[loss=0.1685, simple_loss=0.2407, pruned_loss=0.04809, over 4772.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2439, pruned_loss=0.04949, over 949968.57 frames. ], batch size: 28, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:15:19,906 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158445.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:15:22,039 INFO [finetune.py:976] (4/7) Epoch 28, batch 3800, loss[loss=0.1691, simple_loss=0.2531, pruned_loss=0.04256, over 4818.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2442, pruned_loss=0.04895, over 952530.55 frames. ], batch size: 39, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:15:51,763 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.090e+02 1.580e+02 1.909e+02 2.258e+02 3.504e+02, threshold=3.818e+02, percent-clipped=0.0 +2023-03-27 10:15:55,428 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=158478.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:16:08,884 INFO [finetune.py:976] (4/7) Epoch 28, batch 3850, loss[loss=0.1884, simple_loss=0.2667, pruned_loss=0.05506, over 4908.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2432, pruned_loss=0.04823, over 954926.91 frames. ], batch size: 43, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:16:16,606 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158509.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:16:17,983 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-27 10:16:42,041 INFO [finetune.py:976] (4/7) Epoch 28, batch 3900, loss[loss=0.1448, simple_loss=0.2261, pruned_loss=0.03176, over 4831.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2409, pruned_loss=0.04785, over 955904.45 frames. ], batch size: 33, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:16:42,742 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4007, 2.4684, 1.9981, 2.7483, 2.4135, 2.2065, 2.8150, 2.5188], + device='cuda:4'), covar=tensor([0.1231, 0.2085, 0.2521, 0.2045, 0.2105, 0.1387, 0.2540, 0.1457], + device='cuda:4'), in_proj_covar=tensor([0.0189, 0.0190, 0.0236, 0.0254, 0.0250, 0.0207, 0.0214, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:16:48,974 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158557.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:16:58,923 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.091e+02 1.423e+02 1.682e+02 2.011e+02 3.434e+02, threshold=3.365e+02, percent-clipped=0.0 +2023-03-27 10:17:15,468 INFO [finetune.py:976] (4/7) Epoch 28, batch 3950, loss[loss=0.1622, simple_loss=0.2309, pruned_loss=0.0467, over 4204.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2386, pruned_loss=0.04716, over 953627.80 frames. ], batch size: 65, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:17:33,575 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7036, 1.2784, 0.8486, 1.6009, 2.1445, 1.3328, 1.5982, 1.5780], + device='cuda:4'), covar=tensor([0.1487, 0.2113, 0.1987, 0.1261, 0.1957, 0.1870, 0.1441, 0.1992], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0109, 0.0093, 0.0120, 0.0092, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 10:17:48,879 INFO [finetune.py:976] (4/7) Epoch 28, batch 4000, loss[loss=0.1301, simple_loss=0.1936, pruned_loss=0.03334, over 4234.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2385, pruned_loss=0.04768, over 953923.40 frames. ], batch size: 18, lr: 2.87e-03, grad_scale: 16.0 +2023-03-27 10:18:00,874 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.2118, 4.5820, 4.7402, 4.9825, 4.9761, 4.7981, 5.3598, 1.5525], + device='cuda:4'), covar=tensor([0.0789, 0.0932, 0.0829, 0.1019, 0.1195, 0.1552, 0.0539, 0.5982], + device='cuda:4'), in_proj_covar=tensor([0.0351, 0.0245, 0.0285, 0.0296, 0.0337, 0.0286, 0.0306, 0.0303], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:18:15,604 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.131e+01 1.444e+02 1.852e+02 2.094e+02 7.470e+02, threshold=3.703e+02, percent-clipped=2.0 +2023-03-27 10:18:29,393 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3889, 2.3407, 1.8354, 2.4382, 2.3451, 2.0620, 2.6362, 2.3931], + device='cuda:4'), covar=tensor([0.1340, 0.2006, 0.3020, 0.2461, 0.2502, 0.1702, 0.2975, 0.1582], + device='cuda:4'), in_proj_covar=tensor([0.0190, 0.0190, 0.0236, 0.0254, 0.0250, 0.0207, 0.0214, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:18:32,166 INFO [finetune.py:976] (4/7) Epoch 28, batch 4050, loss[loss=0.1759, simple_loss=0.2581, pruned_loss=0.04681, over 4816.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2424, pruned_loss=0.04954, over 952120.38 frames. ], batch size: 51, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:18:59,956 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=158740.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:19:05,204 INFO [finetune.py:976] (4/7) Epoch 28, batch 4100, loss[loss=0.1436, simple_loss=0.2228, pruned_loss=0.03221, over 4874.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2451, pruned_loss=0.04959, over 952792.76 frames. ], batch size: 34, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:19:18,208 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7267, 1.5098, 2.3470, 3.6475, 2.4220, 2.4401, 1.0434, 2.9992], + device='cuda:4'), covar=tensor([0.1753, 0.1447, 0.1278, 0.0480, 0.0765, 0.1358, 0.1920, 0.0404], + device='cuda:4'), in_proj_covar=tensor([0.0099, 0.0116, 0.0132, 0.0164, 0.0100, 0.0135, 0.0125, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 10:19:18,274 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2732, 2.1132, 1.8915, 2.2243, 2.0719, 2.0548, 2.0987, 2.7978], + device='cuda:4'), covar=tensor([0.3634, 0.4307, 0.3328, 0.3689, 0.3753, 0.2611, 0.3779, 0.1708], + device='cuda:4'), in_proj_covar=tensor([0.0290, 0.0265, 0.0238, 0.0276, 0.0261, 0.0231, 0.0260, 0.0239], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:19:22,719 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.067e+02 1.534e+02 1.832e+02 2.269e+02 3.411e+02, threshold=3.665e+02, percent-clipped=0.0 +2023-03-27 10:19:25,311 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158777.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:19:25,886 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=158778.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:19:29,395 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158783.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:19:38,842 INFO [finetune.py:976] (4/7) Epoch 28, batch 4150, loss[loss=0.205, simple_loss=0.2712, pruned_loss=0.06937, over 4817.00 frames. ], tot_loss[loss=0.1734, simple_loss=0.2466, pruned_loss=0.05008, over 953513.02 frames. ], batch size: 30, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:19:45,667 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-27 10:19:50,674 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3146, 2.2281, 2.3213, 1.6484, 2.2545, 2.4882, 2.5874, 1.9734], + device='cuda:4'), covar=tensor([0.0575, 0.0612, 0.0645, 0.0846, 0.0675, 0.0608, 0.0517, 0.1037], + device='cuda:4'), in_proj_covar=tensor([0.0130, 0.0137, 0.0138, 0.0118, 0.0128, 0.0138, 0.0139, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:19:58,145 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=158826.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:20:03,654 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=158835.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 10:20:05,932 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158838.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:20:06,483 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1426, 3.6652, 3.8186, 4.0122, 3.9046, 3.6096, 4.2270, 1.2973], + device='cuda:4'), covar=tensor([0.0868, 0.0834, 0.0877, 0.1003, 0.1215, 0.1703, 0.0716, 0.5851], + device='cuda:4'), in_proj_covar=tensor([0.0353, 0.0247, 0.0286, 0.0297, 0.0339, 0.0286, 0.0306, 0.0304], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:20:09,571 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158844.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:20:11,342 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2532, 2.8998, 2.7940, 1.2756, 3.0365, 2.2643, 0.6690, 1.9024], + device='cuda:4'), covar=tensor([0.2473, 0.2546, 0.1926, 0.3568, 0.1577, 0.1247, 0.4269, 0.1757], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0181, 0.0160, 0.0130, 0.0163, 0.0124, 0.0150, 0.0126], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 10:20:11,899 INFO [finetune.py:976] (4/7) Epoch 28, batch 4200, loss[loss=0.1803, simple_loss=0.2607, pruned_loss=0.04994, over 4924.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2455, pruned_loss=0.04875, over 952381.08 frames. ], batch size: 33, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:20:18,411 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6360, 1.1751, 0.7769, 1.4461, 2.0623, 0.8039, 1.4185, 1.4280], + device='cuda:4'), covar=tensor([0.1569, 0.2126, 0.1673, 0.1252, 0.1886, 0.1934, 0.1447, 0.2123], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0109, 0.0093, 0.0120, 0.0092, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 10:20:20,272 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.10 vs. limit=2.0 +2023-03-27 10:20:35,368 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.763e+01 1.457e+02 1.627e+02 2.050e+02 3.601e+02, threshold=3.253e+02, percent-clipped=0.0 +2023-03-27 10:20:58,735 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=158896.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 10:21:04,570 INFO [finetune.py:976] (4/7) Epoch 28, batch 4250, loss[loss=0.165, simple_loss=0.2494, pruned_loss=0.04028, over 4903.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2431, pruned_loss=0.04821, over 952027.56 frames. ], batch size: 36, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:21:04,912 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-03-27 10:21:26,980 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1129, 1.6831, 2.3638, 1.5370, 2.0903, 2.3543, 1.6011, 2.4084], + device='cuda:4'), covar=tensor([0.1312, 0.2379, 0.1268, 0.1963, 0.0977, 0.1215, 0.3145, 0.0766], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0207, 0.0194, 0.0190, 0.0174, 0.0214, 0.0219, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:21:46,472 INFO [finetune.py:976] (4/7) Epoch 28, batch 4300, loss[loss=0.1564, simple_loss=0.229, pruned_loss=0.0419, over 4856.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.24, pruned_loss=0.04709, over 952899.82 frames. ], batch size: 49, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:22:03,617 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.096e+02 1.473e+02 1.789e+02 2.045e+02 3.501e+02, threshold=3.577e+02, percent-clipped=2.0 +2023-03-27 10:22:20,211 INFO [finetune.py:976] (4/7) Epoch 28, batch 4350, loss[loss=0.1413, simple_loss=0.2131, pruned_loss=0.03474, over 4849.00 frames. ], tot_loss[loss=0.1641, simple_loss=0.2364, pruned_loss=0.04592, over 952732.30 frames. ], batch size: 44, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:22:48,303 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159040.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:22:53,110 INFO [finetune.py:976] (4/7) Epoch 28, batch 4400, loss[loss=0.1564, simple_loss=0.2096, pruned_loss=0.05163, over 4151.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2367, pruned_loss=0.04635, over 949474.04 frames. ], batch size: 18, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:23:07,430 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159069.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:23:09,715 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.014e+02 1.529e+02 1.726e+02 2.218e+02 4.795e+02, threshold=3.452e+02, percent-clipped=2.0 +2023-03-27 10:23:24,711 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159088.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:23:35,375 INFO [finetune.py:976] (4/7) Epoch 28, batch 4450, loss[loss=0.1699, simple_loss=0.2411, pruned_loss=0.04939, over 4934.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2418, pruned_loss=0.0479, over 951667.89 frames. ], batch size: 33, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:23:56,084 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5310, 1.4851, 1.3097, 1.4543, 1.8801, 1.7407, 1.6094, 1.3872], + device='cuda:4'), covar=tensor([0.0385, 0.0360, 0.0658, 0.0357, 0.0218, 0.0563, 0.0313, 0.0434], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0107, 0.0148, 0.0112, 0.0102, 0.0117, 0.0105, 0.0115], + device='cuda:4'), out_proj_covar=tensor([7.9118e-05, 8.1683e-05, 1.1545e-04, 8.5419e-05, 7.8984e-05, 8.5994e-05, + 7.7779e-05, 8.7016e-05], device='cuda:4') +2023-03-27 10:24:00,770 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159130.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:24:02,994 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159133.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:24:07,590 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159139.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:24:12,912 INFO [finetune.py:976] (4/7) Epoch 28, batch 4500, loss[loss=0.231, simple_loss=0.2925, pruned_loss=0.08472, over 4810.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2428, pruned_loss=0.04817, over 952013.19 frames. ], batch size: 45, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:24:20,198 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159159.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:24:28,963 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.131e+02 1.553e+02 1.848e+02 2.152e+02 3.966e+02, threshold=3.696e+02, percent-clipped=2.0 +2023-03-27 10:24:41,941 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159191.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 10:24:46,582 INFO [finetune.py:976] (4/7) Epoch 28, batch 4550, loss[loss=0.1771, simple_loss=0.2525, pruned_loss=0.05085, over 4869.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2442, pruned_loss=0.04865, over 952763.13 frames. ], batch size: 34, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:24:51,460 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-27 10:25:00,127 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.74 vs. limit=5.0 +2023-03-27 10:25:00,530 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159220.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:25:20,095 INFO [finetune.py:976] (4/7) Epoch 28, batch 4600, loss[loss=0.1614, simple_loss=0.2346, pruned_loss=0.0441, over 4828.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2455, pruned_loss=0.04931, over 954903.30 frames. ], batch size: 39, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:25:35,686 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.005e+02 1.454e+02 1.668e+02 2.062e+02 3.965e+02, threshold=3.336e+02, percent-clipped=3.0 +2023-03-27 10:25:43,695 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4726, 1.4525, 2.2055, 1.7436, 1.6642, 3.5738, 1.3699, 1.7264], + device='cuda:4'), covar=tensor([0.0944, 0.1683, 0.0909, 0.0899, 0.1502, 0.0208, 0.1484, 0.1707], + device='cuda:4'), in_proj_covar=tensor([0.0076, 0.0083, 0.0074, 0.0077, 0.0092, 0.0081, 0.0086, 0.0081], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 10:25:54,148 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159290.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:26:05,022 INFO [finetune.py:976] (4/7) Epoch 28, batch 4650, loss[loss=0.156, simple_loss=0.2344, pruned_loss=0.03875, over 4822.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2431, pruned_loss=0.049, over 953954.08 frames. ], batch size: 38, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:26:34,545 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3283, 1.9879, 2.6874, 1.7896, 2.3308, 2.5190, 1.8119, 2.5974], + device='cuda:4'), covar=tensor([0.1270, 0.1942, 0.1471, 0.1795, 0.0870, 0.1342, 0.2515, 0.0831], + device='cuda:4'), in_proj_covar=tensor([0.0191, 0.0206, 0.0193, 0.0189, 0.0174, 0.0214, 0.0218, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:26:35,148 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9059, 1.3164, 0.7836, 1.7862, 2.4385, 1.4396, 1.7467, 1.8016], + device='cuda:4'), covar=tensor([0.1376, 0.2019, 0.1902, 0.1138, 0.1694, 0.1868, 0.1269, 0.1968], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0108, 0.0093, 0.0120, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 10:26:55,598 INFO [finetune.py:976] (4/7) Epoch 28, batch 4700, loss[loss=0.1432, simple_loss=0.2117, pruned_loss=0.03735, over 4828.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2402, pruned_loss=0.04812, over 954819.27 frames. ], batch size: 40, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:26:58,023 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159351.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:27:11,662 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.356e+01 1.407e+02 1.592e+02 1.978e+02 3.098e+02, threshold=3.183e+02, percent-clipped=0.0 +2023-03-27 10:27:28,643 INFO [finetune.py:976] (4/7) Epoch 28, batch 4750, loss[loss=0.1899, simple_loss=0.2531, pruned_loss=0.06333, over 4867.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2387, pruned_loss=0.04809, over 953054.60 frames. ], batch size: 34, lr: 2.87e-03, grad_scale: 32.0 +2023-03-27 10:27:46,457 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159425.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:27:51,867 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159433.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:27:55,937 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159439.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:28:02,248 INFO [finetune.py:976] (4/7) Epoch 28, batch 4800, loss[loss=0.2142, simple_loss=0.2909, pruned_loss=0.06872, over 4833.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2414, pruned_loss=0.04892, over 953244.37 frames. ], batch size: 49, lr: 2.86e-03, grad_scale: 32.0 +2023-03-27 10:28:18,796 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.120e+02 1.565e+02 1.779e+02 2.195e+02 3.956e+02, threshold=3.558e+02, percent-clipped=1.0 +2023-03-27 10:28:23,706 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159481.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:28:27,820 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159487.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:28:30,768 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159491.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 10:28:36,823 INFO [finetune.py:976] (4/7) Epoch 28, batch 4850, loss[loss=0.2107, simple_loss=0.2878, pruned_loss=0.06678, over 4831.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.2439, pruned_loss=0.0494, over 953803.64 frames. ], batch size: 30, lr: 2.86e-03, grad_scale: 32.0 +2023-03-27 10:28:36,945 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0580, 2.0068, 2.1140, 1.5694, 1.9978, 2.2029, 2.2626, 1.7372], + device='cuda:4'), covar=tensor([0.0613, 0.0691, 0.0698, 0.0835, 0.0822, 0.0708, 0.0599, 0.1226], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0139, 0.0142, 0.0120, 0.0130, 0.0141, 0.0142, 0.0165], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:28:49,519 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-27 10:28:57,882 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159515.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:29:17,409 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159539.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 10:29:19,014 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-27 10:29:23,231 INFO [finetune.py:976] (4/7) Epoch 28, batch 4900, loss[loss=0.1699, simple_loss=0.2483, pruned_loss=0.04572, over 4907.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2458, pruned_loss=0.04984, over 955791.83 frames. ], batch size: 37, lr: 2.86e-03, grad_scale: 32.0 +2023-03-27 10:29:23,733 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-03-27 10:29:40,326 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.553e+02 1.825e+02 2.271e+02 5.584e+02, threshold=3.651e+02, percent-clipped=3.0 +2023-03-27 10:29:56,962 INFO [finetune.py:976] (4/7) Epoch 28, batch 4950, loss[loss=0.1603, simple_loss=0.2331, pruned_loss=0.04378, over 4726.00 frames. ], tot_loss[loss=0.1735, simple_loss=0.2465, pruned_loss=0.0503, over 954955.62 frames. ], batch size: 54, lr: 2.86e-03, grad_scale: 32.0 +2023-03-27 10:30:09,227 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2807, 2.1346, 1.7558, 2.2442, 2.7299, 2.2521, 2.2618, 1.7498], + device='cuda:4'), covar=tensor([0.2098, 0.1907, 0.1955, 0.1651, 0.1675, 0.1116, 0.1938, 0.1768], + device='cuda:4'), in_proj_covar=tensor([0.0248, 0.0214, 0.0217, 0.0201, 0.0248, 0.0193, 0.0219, 0.0207], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:30:18,764 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159631.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:30:18,814 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5351, 2.4399, 2.0197, 2.7377, 2.4719, 2.0711, 2.9938, 2.6110], + device='cuda:4'), covar=tensor([0.1295, 0.2139, 0.2844, 0.2321, 0.2523, 0.1607, 0.2773, 0.1646], + device='cuda:4'), in_proj_covar=tensor([0.0188, 0.0189, 0.0235, 0.0252, 0.0249, 0.0207, 0.0214, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:30:28,813 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159646.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:30:29,931 INFO [finetune.py:976] (4/7) Epoch 28, batch 5000, loss[loss=0.1313, simple_loss=0.2111, pruned_loss=0.0257, over 4754.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2444, pruned_loss=0.04949, over 955838.88 frames. ], batch size: 27, lr: 2.86e-03, grad_scale: 32.0 +2023-03-27 10:30:41,500 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8425, 1.3981, 1.0340, 1.7623, 2.2579, 1.4779, 1.6274, 1.6712], + device='cuda:4'), covar=tensor([0.1346, 0.1963, 0.1683, 0.1123, 0.1807, 0.1938, 0.1444, 0.1793], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0093, 0.0109, 0.0093, 0.0120, 0.0093, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 10:30:47,442 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.023e+02 1.490e+02 1.734e+02 1.957e+02 3.436e+02, threshold=3.469e+02, percent-clipped=0.0 +2023-03-27 10:30:59,455 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159692.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:31:05,619 INFO [finetune.py:976] (4/7) Epoch 28, batch 5050, loss[loss=0.1461, simple_loss=0.2245, pruned_loss=0.03388, over 4826.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2417, pruned_loss=0.04882, over 955874.13 frames. ], batch size: 25, lr: 2.86e-03, grad_scale: 32.0 +2023-03-27 10:31:14,971 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9122, 1.9899, 1.7074, 2.1570, 2.2829, 2.1455, 1.8138, 1.5973], + device='cuda:4'), covar=tensor([0.2029, 0.1670, 0.1685, 0.1398, 0.1647, 0.1096, 0.2173, 0.1789], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0212, 0.0214, 0.0199, 0.0245, 0.0191, 0.0217, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:31:32,282 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159725.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:31:43,427 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.9503, 4.3449, 4.5190, 4.7188, 4.7410, 4.4246, 5.0517, 1.6146], + device='cuda:4'), covar=tensor([0.0714, 0.0832, 0.0765, 0.0917, 0.1124, 0.1465, 0.0571, 0.5689], + device='cuda:4'), in_proj_covar=tensor([0.0355, 0.0249, 0.0288, 0.0299, 0.0342, 0.0290, 0.0307, 0.0306], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:31:55,160 INFO [finetune.py:976] (4/7) Epoch 28, batch 5100, loss[loss=0.1756, simple_loss=0.2441, pruned_loss=0.05356, over 4866.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2378, pruned_loss=0.04723, over 955939.80 frames. ], batch size: 31, lr: 2.86e-03, grad_scale: 32.0 +2023-03-27 10:32:21,761 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.008e+02 1.553e+02 1.866e+02 2.180e+02 3.771e+02, threshold=3.731e+02, percent-clipped=1.0 +2023-03-27 10:32:21,834 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159773.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:32:37,624 INFO [finetune.py:976] (4/7) Epoch 28, batch 5150, loss[loss=0.174, simple_loss=0.2585, pruned_loss=0.04472, over 4820.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2395, pruned_loss=0.04816, over 954712.84 frames. ], batch size: 39, lr: 2.86e-03, grad_scale: 32.0 +2023-03-27 10:32:44,548 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6555, 1.5581, 1.3812, 1.7563, 1.6614, 1.6901, 1.0682, 1.4608], + device='cuda:4'), covar=tensor([0.2198, 0.1927, 0.1919, 0.1575, 0.1588, 0.1228, 0.2516, 0.1896], + device='cuda:4'), in_proj_covar=tensor([0.0247, 0.0212, 0.0215, 0.0200, 0.0246, 0.0192, 0.0218, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:32:49,249 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159815.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:33:11,651 INFO [finetune.py:976] (4/7) Epoch 28, batch 5200, loss[loss=0.158, simple_loss=0.2311, pruned_loss=0.04247, over 4775.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2419, pruned_loss=0.04859, over 953084.05 frames. ], batch size: 26, lr: 2.86e-03, grad_scale: 32.0 +2023-03-27 10:33:16,454 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159855.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:33:18,546 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-27 10:33:21,712 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159863.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:33:28,751 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.072e+02 1.499e+02 1.871e+02 2.174e+02 4.313e+02, threshold=3.743e+02, percent-clipped=1.0 +2023-03-27 10:33:40,664 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-03-27 10:33:44,910 INFO [finetune.py:976] (4/7) Epoch 28, batch 5250, loss[loss=0.1757, simple_loss=0.2559, pruned_loss=0.04774, over 4865.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2427, pruned_loss=0.04799, over 954351.18 frames. ], batch size: 34, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:33:52,098 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0501, 1.8977, 1.6987, 2.0464, 2.6861, 2.0725, 2.1274, 1.6224], + device='cuda:4'), covar=tensor([0.2053, 0.1888, 0.1840, 0.1621, 0.1581, 0.1138, 0.1874, 0.1765], + device='cuda:4'), in_proj_covar=tensor([0.0246, 0.0212, 0.0215, 0.0199, 0.0245, 0.0191, 0.0217, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:33:53,299 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=159910.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:33:59,359 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159916.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:34:14,942 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7405, 1.1403, 0.8082, 1.6196, 2.0906, 1.5096, 1.4678, 1.6151], + device='cuda:4'), covar=tensor([0.1532, 0.2166, 0.1911, 0.1268, 0.1868, 0.1883, 0.1490, 0.1942], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0093, 0.0108, 0.0093, 0.0120, 0.0092, 0.0098, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 10:34:26,208 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=159946.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:34:27,324 INFO [finetune.py:976] (4/7) Epoch 28, batch 5300, loss[loss=0.1962, simple_loss=0.2781, pruned_loss=0.05718, over 4928.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2452, pruned_loss=0.04883, over 955802.96 frames. ], batch size: 42, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:34:50,623 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=159971.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:34:52,306 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.066e+02 1.498e+02 1.765e+02 2.107e+02 3.764e+02, threshold=3.530e+02, percent-clipped=1.0 +2023-03-27 10:34:55,872 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4256, 2.6509, 2.4941, 1.9791, 2.4687, 2.8290, 2.8927, 2.3533], + device='cuda:4'), covar=tensor([0.0637, 0.0589, 0.0756, 0.0791, 0.0744, 0.0664, 0.0554, 0.1013], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0139, 0.0141, 0.0120, 0.0130, 0.0140, 0.0141, 0.0165], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:35:01,757 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=159987.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:35:06,016 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=159994.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:35:08,413 INFO [finetune.py:976] (4/7) Epoch 28, batch 5350, loss[loss=0.1718, simple_loss=0.2511, pruned_loss=0.04624, over 4864.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2448, pruned_loss=0.0487, over 952809.82 frames. ], batch size: 34, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:35:43,143 INFO [finetune.py:976] (4/7) Epoch 28, batch 5400, loss[loss=0.1465, simple_loss=0.2298, pruned_loss=0.03162, over 4770.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2427, pruned_loss=0.04847, over 953745.84 frames. ], batch size: 27, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:36:00,195 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.280e+01 1.494e+02 1.877e+02 2.215e+02 3.734e+02, threshold=3.755e+02, percent-clipped=1.0 +2023-03-27 10:36:16,670 INFO [finetune.py:976] (4/7) Epoch 28, batch 5450, loss[loss=0.1444, simple_loss=0.2196, pruned_loss=0.03463, over 4855.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2402, pruned_loss=0.04817, over 955284.30 frames. ], batch size: 44, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:37:05,409 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4217, 3.1743, 3.0692, 1.5101, 3.3062, 2.4757, 1.0001, 2.2753], + device='cuda:4'), covar=tensor([0.2164, 0.1749, 0.1501, 0.3393, 0.1026, 0.1046, 0.3852, 0.1403], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0179, 0.0159, 0.0129, 0.0163, 0.0124, 0.0148, 0.0126], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 10:37:07,796 INFO [finetune.py:976] (4/7) Epoch 28, batch 5500, loss[loss=0.1553, simple_loss=0.232, pruned_loss=0.03933, over 4787.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2375, pruned_loss=0.04699, over 955164.73 frames. ], batch size: 29, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:37:38,236 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.035e+02 1.497e+02 1.808e+02 2.088e+02 3.346e+02, threshold=3.616e+02, percent-clipped=0.0 +2023-03-27 10:37:55,669 INFO [finetune.py:976] (4/7) Epoch 28, batch 5550, loss[loss=0.1817, simple_loss=0.265, pruned_loss=0.04926, over 4931.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2387, pruned_loss=0.04765, over 953338.20 frames. ], batch size: 38, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:38:03,852 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=160211.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:38:27,206 INFO [finetune.py:976] (4/7) Epoch 28, batch 5600, loss[loss=0.1829, simple_loss=0.2604, pruned_loss=0.05272, over 4871.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2415, pruned_loss=0.04756, over 953961.31 frames. ], batch size: 34, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:38:37,623 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=160266.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:38:41,743 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4277, 2.5946, 2.4792, 1.8703, 2.3570, 2.7741, 2.9485, 2.2232], + device='cuda:4'), covar=tensor([0.0624, 0.0627, 0.0724, 0.0826, 0.0888, 0.0670, 0.0488, 0.1006], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0139, 0.0141, 0.0120, 0.0130, 0.0141, 0.0141, 0.0164], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:38:42,223 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.097e+02 1.448e+02 1.891e+02 2.366e+02 4.690e+02, threshold=3.782e+02, percent-clipped=2.0 +2023-03-27 10:38:49,852 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=160287.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:38:52,192 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7857, 1.6840, 1.6106, 1.7566, 1.2784, 3.7307, 1.5326, 2.1542], + device='cuda:4'), covar=tensor([0.3288, 0.2507, 0.2127, 0.2327, 0.1727, 0.0168, 0.2484, 0.1115], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0117, 0.0121, 0.0124, 0.0114, 0.0096, 0.0094, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 10:38:56,599 INFO [finetune.py:976] (4/7) Epoch 28, batch 5650, loss[loss=0.1482, simple_loss=0.229, pruned_loss=0.03373, over 4856.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2433, pruned_loss=0.04749, over 952117.55 frames. ], batch size: 49, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:39:01,990 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=160307.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 10:39:04,338 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0423, 0.9860, 0.9500, 0.3791, 0.9570, 1.1609, 1.1668, 0.9779], + device='cuda:4'), covar=tensor([0.0906, 0.0648, 0.0616, 0.0576, 0.0556, 0.0687, 0.0393, 0.0729], + device='cuda:4'), in_proj_covar=tensor([0.0120, 0.0147, 0.0130, 0.0122, 0.0130, 0.0129, 0.0142, 0.0150], + device='cuda:4'), out_proj_covar=tensor([8.7481e-05, 1.0549e-04, 9.2170e-05, 8.5213e-05, 9.1351e-05, 9.1622e-05, + 1.0084e-04, 1.0734e-04], device='cuda:4') +2023-03-27 10:39:24,455 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.80 vs. limit=2.0 +2023-03-27 10:39:24,564 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=160335.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:39:32,243 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6951, 2.5489, 2.6376, 1.9866, 2.4675, 2.7920, 2.8653, 2.2122], + device='cuda:4'), covar=tensor([0.0474, 0.0563, 0.0557, 0.0754, 0.0870, 0.0588, 0.0464, 0.1052], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0139, 0.0141, 0.0120, 0.0129, 0.0141, 0.0141, 0.0164], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:39:35,784 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6230, 1.5846, 1.3961, 1.7548, 1.6169, 1.6380, 1.1048, 1.4365], + device='cuda:4'), covar=tensor([0.2021, 0.1824, 0.1824, 0.1460, 0.1495, 0.1152, 0.2126, 0.1805], + device='cuda:4'), in_proj_covar=tensor([0.0245, 0.0210, 0.0214, 0.0198, 0.0244, 0.0190, 0.0215, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:39:36,256 INFO [finetune.py:976] (4/7) Epoch 28, batch 5700, loss[loss=0.1298, simple_loss=0.2033, pruned_loss=0.02821, over 4200.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2392, pruned_loss=0.0471, over 933308.17 frames. ], batch size: 18, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:39:36,940 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.0021, 2.8561, 2.5435, 3.2587, 2.8489, 2.6338, 3.4032, 2.9887], + device='cuda:4'), covar=tensor([0.1248, 0.1942, 0.2632, 0.2046, 0.2648, 0.1732, 0.2347, 0.1759], + device='cuda:4'), in_proj_covar=tensor([0.0188, 0.0189, 0.0235, 0.0252, 0.0249, 0.0207, 0.0215, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:39:48,027 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=160368.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 10:39:53,238 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.932e+01 1.353e+02 1.695e+02 2.046e+02 5.031e+02, threshold=3.390e+02, percent-clipped=3.0 +2023-03-27 10:40:10,867 INFO [finetune.py:976] (4/7) Epoch 29, batch 0, loss[loss=0.1938, simple_loss=0.2636, pruned_loss=0.062, over 4919.00 frames. ], tot_loss[loss=0.1938, simple_loss=0.2636, pruned_loss=0.062, over 4919.00 frames. ], batch size: 33, lr: 2.86e-03, grad_scale: 16.0 +2023-03-27 10:40:10,867 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 10:40:21,879 INFO [finetune.py:1010] (4/7) Epoch 29, validation: loss=0.1588, simple_loss=0.2262, pruned_loss=0.04569, over 2265189.00 frames. +2023-03-27 10:40:21,879 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6557MB +2023-03-27 10:40:31,944 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.80 vs. limit=2.0 +2023-03-27 10:40:57,961 INFO [finetune.py:976] (4/7) Epoch 29, batch 50, loss[loss=0.1386, simple_loss=0.2131, pruned_loss=0.03208, over 4880.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.2485, pruned_loss=0.05006, over 217000.93 frames. ], batch size: 35, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:41:38,858 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.219e+02 1.505e+02 1.852e+02 2.145e+02 8.823e+02, threshold=3.704e+02, percent-clipped=1.0 +2023-03-27 10:41:39,935 INFO [finetune.py:976] (4/7) Epoch 29, batch 100, loss[loss=0.1551, simple_loss=0.2281, pruned_loss=0.04106, over 4911.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2391, pruned_loss=0.04744, over 380418.14 frames. ], batch size: 46, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:42:12,556 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4633, 1.4260, 1.9935, 1.7208, 1.5724, 3.2820, 1.3725, 1.5922], + device='cuda:4'), covar=tensor([0.1028, 0.1810, 0.1298, 0.0991, 0.1539, 0.0266, 0.1450, 0.1722], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0083, 0.0073, 0.0076, 0.0091, 0.0080, 0.0086, 0.0081], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 10:42:14,767 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=160511.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:42:34,274 INFO [finetune.py:976] (4/7) Epoch 29, batch 150, loss[loss=0.1873, simple_loss=0.2549, pruned_loss=0.05987, over 4926.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2356, pruned_loss=0.04693, over 506524.48 frames. ], batch size: 33, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:42:55,950 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=160559.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:42:56,609 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6669, 1.6379, 1.9061, 1.2074, 1.6688, 1.9068, 1.5199, 2.0731], + device='cuda:4'), covar=tensor([0.1298, 0.2066, 0.1338, 0.1723, 0.0949, 0.1281, 0.3024, 0.0877], + device='cuda:4'), in_proj_covar=tensor([0.0191, 0.0205, 0.0193, 0.0189, 0.0174, 0.0213, 0.0218, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:43:00,196 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9914, 1.8024, 1.6571, 1.4359, 1.7829, 1.7968, 1.7798, 2.2595], + device='cuda:4'), covar=tensor([0.3641, 0.4033, 0.3155, 0.3586, 0.3547, 0.2386, 0.3196, 0.1885], + device='cuda:4'), in_proj_covar=tensor([0.0290, 0.0266, 0.0238, 0.0277, 0.0262, 0.0233, 0.0261, 0.0239], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:43:00,755 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=160566.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:43:06,536 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.027e+02 1.446e+02 1.774e+02 2.135e+02 3.412e+02, threshold=3.547e+02, percent-clipped=0.0 +2023-03-27 10:43:07,129 INFO [finetune.py:976] (4/7) Epoch 29, batch 200, loss[loss=0.1841, simple_loss=0.2539, pruned_loss=0.05721, over 4801.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2335, pruned_loss=0.04626, over 607926.10 frames. ], batch size: 51, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:43:18,703 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-03-27 10:43:30,743 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0570, 0.8788, 0.9443, 1.0643, 1.2318, 1.1681, 1.0515, 0.9731], + device='cuda:4'), covar=tensor([0.0397, 0.0339, 0.0694, 0.0333, 0.0297, 0.0409, 0.0373, 0.0444], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0106, 0.0148, 0.0111, 0.0101, 0.0116, 0.0104, 0.0114], + device='cuda:4'), out_proj_covar=tensor([7.8404e-05, 8.0991e-05, 1.1512e-04, 8.4740e-05, 7.8258e-05, 8.5371e-05, + 7.7249e-05, 8.6834e-05], device='cuda:4') +2023-03-27 10:43:32,501 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=160614.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:43:39,909 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9534, 1.4155, 1.9912, 2.0006, 1.7730, 1.7723, 1.8627, 1.9166], + device='cuda:4'), covar=tensor([0.3821, 0.3966, 0.3389, 0.3694, 0.5059, 0.4106, 0.4527, 0.3030], + device='cuda:4'), in_proj_covar=tensor([0.0268, 0.0248, 0.0269, 0.0298, 0.0298, 0.0274, 0.0303, 0.0253], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:43:40,981 INFO [finetune.py:976] (4/7) Epoch 29, batch 250, loss[loss=0.1654, simple_loss=0.2376, pruned_loss=0.04654, over 4826.00 frames. ], tot_loss[loss=0.1641, simple_loss=0.2348, pruned_loss=0.04664, over 686691.54 frames. ], batch size: 33, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:44:05,537 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=160663.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 10:44:12,918 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.082e+02 1.455e+02 1.868e+02 2.134e+02 3.558e+02, threshold=3.736e+02, percent-clipped=1.0 +2023-03-27 10:44:13,967 INFO [finetune.py:976] (4/7) Epoch 29, batch 300, loss[loss=0.1478, simple_loss=0.2244, pruned_loss=0.03557, over 4801.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2392, pruned_loss=0.04781, over 747436.99 frames. ], batch size: 25, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:44:16,384 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2410, 2.2654, 2.3611, 1.6017, 2.2410, 2.4272, 2.4696, 1.9331], + device='cuda:4'), covar=tensor([0.0654, 0.0690, 0.0676, 0.0889, 0.0721, 0.0713, 0.0617, 0.1100], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0138, 0.0140, 0.0119, 0.0128, 0.0140, 0.0140, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:44:42,904 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2993, 1.1143, 1.0572, 1.1133, 1.5504, 1.4520, 1.2534, 1.1355], + device='cuda:4'), covar=tensor([0.0375, 0.0388, 0.0936, 0.0442, 0.0280, 0.0528, 0.0372, 0.0514], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0106, 0.0148, 0.0111, 0.0101, 0.0116, 0.0104, 0.0114], + device='cuda:4'), out_proj_covar=tensor([7.8271e-05, 8.0882e-05, 1.1512e-04, 8.4626e-05, 7.8226e-05, 8.5415e-05, + 7.7299e-05, 8.6607e-05], device='cuda:4') +2023-03-27 10:44:57,148 INFO [finetune.py:976] (4/7) Epoch 29, batch 350, loss[loss=0.1215, simple_loss=0.195, pruned_loss=0.02395, over 4724.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2412, pruned_loss=0.04832, over 794349.12 frames. ], batch size: 23, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:45:37,468 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.928e+01 1.581e+02 1.864e+02 2.210e+02 3.251e+02, threshold=3.728e+02, percent-clipped=0.0 +2023-03-27 10:45:38,107 INFO [finetune.py:976] (4/7) Epoch 29, batch 400, loss[loss=0.1966, simple_loss=0.2719, pruned_loss=0.06061, over 4878.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2433, pruned_loss=0.04899, over 828098.91 frames. ], batch size: 35, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:46:04,902 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2041, 2.0639, 2.1727, 1.0364, 2.4695, 2.6094, 2.2609, 1.9679], + device='cuda:4'), covar=tensor([0.1130, 0.0934, 0.0586, 0.0797, 0.0698, 0.0837, 0.0556, 0.0962], + device='cuda:4'), in_proj_covar=tensor([0.0120, 0.0148, 0.0130, 0.0122, 0.0131, 0.0130, 0.0142, 0.0151], + device='cuda:4'), out_proj_covar=tensor([8.7865e-05, 1.0578e-04, 9.2259e-05, 8.5376e-05, 9.1669e-05, 9.1751e-05, + 1.0126e-04, 1.0756e-04], device='cuda:4') +2023-03-27 10:46:11,858 INFO [finetune.py:976] (4/7) Epoch 29, batch 450, loss[loss=0.197, simple_loss=0.2646, pruned_loss=0.06473, over 4686.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.243, pruned_loss=0.04869, over 853774.37 frames. ], batch size: 23, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:46:55,069 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.473e+02 1.674e+02 2.082e+02 4.783e+02, threshold=3.348e+02, percent-clipped=2.0 +2023-03-27 10:46:55,687 INFO [finetune.py:976] (4/7) Epoch 29, batch 500, loss[loss=0.1706, simple_loss=0.2341, pruned_loss=0.0535, over 4837.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2392, pruned_loss=0.04754, over 874948.78 frames. ], batch size: 33, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:47:03,971 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1246, 2.0546, 2.1878, 1.4926, 2.1171, 2.3175, 2.3454, 1.8442], + device='cuda:4'), covar=tensor([0.0532, 0.0600, 0.0634, 0.0870, 0.0740, 0.0664, 0.0511, 0.1032], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0138, 0.0140, 0.0119, 0.0128, 0.0140, 0.0140, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:47:15,098 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=160901.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:47:36,926 INFO [finetune.py:976] (4/7) Epoch 29, batch 550, loss[loss=0.1441, simple_loss=0.2189, pruned_loss=0.0347, over 4817.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2365, pruned_loss=0.04657, over 894816.69 frames. ], batch size: 40, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:47:56,851 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.7742, 3.5784, 3.4975, 1.7690, 3.8186, 2.9123, 0.9248, 2.5299], + device='cuda:4'), covar=tensor([0.2847, 0.2700, 0.1738, 0.3712, 0.1118, 0.0981, 0.4713, 0.1812], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0180, 0.0161, 0.0130, 0.0164, 0.0124, 0.0150, 0.0127], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 10:48:08,224 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=160962.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:48:08,817 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=160963.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 10:48:10,040 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=160965.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:48:15,384 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.000e+01 1.387e+02 1.769e+02 2.082e+02 3.377e+02, threshold=3.538e+02, percent-clipped=1.0 +2023-03-27 10:48:16,011 INFO [finetune.py:976] (4/7) Epoch 29, batch 600, loss[loss=0.1817, simple_loss=0.2658, pruned_loss=0.0488, over 4819.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2373, pruned_loss=0.04736, over 907963.99 frames. ], batch size: 40, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:48:27,318 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7257, 1.6004, 1.4224, 1.7853, 1.7044, 1.7572, 1.0952, 1.4865], + device='cuda:4'), covar=tensor([0.2024, 0.1868, 0.1928, 0.1542, 0.1469, 0.1177, 0.2343, 0.1824], + device='cuda:4'), in_proj_covar=tensor([0.0249, 0.0214, 0.0217, 0.0201, 0.0248, 0.0193, 0.0218, 0.0207], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:48:34,186 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.87 vs. limit=5.0 +2023-03-27 10:48:36,297 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.95 vs. limit=5.0 +2023-03-27 10:48:41,019 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161011.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 10:48:48,394 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-27 10:48:49,431 INFO [finetune.py:976] (4/7) Epoch 29, batch 650, loss[loss=0.1881, simple_loss=0.2597, pruned_loss=0.05822, over 4905.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2394, pruned_loss=0.0475, over 919978.70 frames. ], batch size: 36, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:48:50,201 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161026.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:49:22,503 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.055e+02 1.587e+02 1.904e+02 2.267e+02 4.706e+02, threshold=3.807e+02, percent-clipped=2.0 +2023-03-27 10:49:23,106 INFO [finetune.py:976] (4/7) Epoch 29, batch 700, loss[loss=0.1929, simple_loss=0.2703, pruned_loss=0.05778, over 4929.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2415, pruned_loss=0.04816, over 926155.16 frames. ], batch size: 38, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:50:03,475 INFO [finetune.py:976] (4/7) Epoch 29, batch 750, loss[loss=0.168, simple_loss=0.2421, pruned_loss=0.04696, over 4733.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2437, pruned_loss=0.04873, over 932176.78 frames. ], batch size: 54, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:50:46,355 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.613e+01 1.645e+02 1.965e+02 2.346e+02 5.118e+02, threshold=3.931e+02, percent-clipped=1.0 +2023-03-27 10:50:46,989 INFO [finetune.py:976] (4/7) Epoch 29, batch 800, loss[loss=0.1454, simple_loss=0.2155, pruned_loss=0.03769, over 4835.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.243, pruned_loss=0.04803, over 936444.93 frames. ], batch size: 49, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:50:53,812 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6200, 2.8325, 2.6667, 2.0572, 2.7258, 3.1638, 3.1596, 2.4504], + device='cuda:4'), covar=tensor([0.0569, 0.0557, 0.0670, 0.0773, 0.0626, 0.0557, 0.0473, 0.1014], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0139, 0.0141, 0.0119, 0.0129, 0.0141, 0.0141, 0.0164], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:51:12,469 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1303, 2.0386, 1.7978, 1.9117, 1.9973, 1.9391, 1.9647, 2.6570], + device='cuda:4'), covar=tensor([0.3233, 0.3925, 0.2899, 0.3510, 0.3514, 0.2366, 0.3136, 0.1471], + device='cuda:4'), in_proj_covar=tensor([0.0288, 0.0264, 0.0238, 0.0275, 0.0260, 0.0231, 0.0259, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:51:15,289 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=161216.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:51:20,580 INFO [finetune.py:976] (4/7) Epoch 29, batch 850, loss[loss=0.1537, simple_loss=0.2292, pruned_loss=0.03903, over 4856.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2407, pruned_loss=0.04729, over 941225.27 frames. ], batch size: 49, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:51:24,350 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=161231.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:51:27,349 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.8345, 3.3356, 3.5456, 3.6767, 3.5859, 3.3115, 3.8977, 1.1903], + device='cuda:4'), covar=tensor([0.0833, 0.1021, 0.0967, 0.1080, 0.1422, 0.1765, 0.0852, 0.5922], + device='cuda:4'), in_proj_covar=tensor([0.0356, 0.0250, 0.0289, 0.0301, 0.0343, 0.0290, 0.0309, 0.0308], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:51:48,943 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161257.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:52:02,288 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-27 10:52:04,266 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.030e+02 1.396e+02 1.676e+02 2.040e+02 3.102e+02, threshold=3.353e+02, percent-clipped=0.0 +2023-03-27 10:52:04,923 INFO [finetune.py:976] (4/7) Epoch 29, batch 900, loss[loss=0.1481, simple_loss=0.2268, pruned_loss=0.03467, over 4790.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2388, pruned_loss=0.04651, over 945628.48 frames. ], batch size: 29, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:52:06,250 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161277.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:52:15,355 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161292.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:52:35,803 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161321.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:52:38,180 INFO [finetune.py:976] (4/7) Epoch 29, batch 950, loss[loss=0.1798, simple_loss=0.2485, pruned_loss=0.05559, over 4870.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2376, pruned_loss=0.04612, over 949190.35 frames. ], batch size: 34, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:52:56,964 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0696, 1.3657, 0.7840, 1.9356, 2.2595, 1.7247, 1.5594, 1.9410], + device='cuda:4'), covar=tensor([0.1412, 0.2112, 0.2135, 0.1166, 0.1920, 0.1924, 0.1414, 0.1899], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0109, 0.0093, 0.0120, 0.0093, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 10:53:00,265 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.80 vs. limit=2.0 +2023-03-27 10:53:28,291 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.815e+01 1.511e+02 1.765e+02 2.170e+02 3.612e+02, threshold=3.529e+02, percent-clipped=1.0 +2023-03-27 10:53:28,921 INFO [finetune.py:976] (4/7) Epoch 29, batch 1000, loss[loss=0.1662, simple_loss=0.2496, pruned_loss=0.04137, over 4900.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2402, pruned_loss=0.04729, over 950455.57 frames. ], batch size: 35, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:53:48,062 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.4429, 1.5688, 1.6272, 0.9047, 1.6950, 1.9713, 1.8657, 1.4431], + device='cuda:4'), covar=tensor([0.0931, 0.0672, 0.0585, 0.0591, 0.0486, 0.0575, 0.0373, 0.0808], + device='cuda:4'), in_proj_covar=tensor([0.0121, 0.0148, 0.0131, 0.0123, 0.0132, 0.0130, 0.0143, 0.0152], + device='cuda:4'), out_proj_covar=tensor([8.8578e-05, 1.0634e-04, 9.3134e-05, 8.6019e-05, 9.2506e-05, 9.2271e-05, + 1.0194e-04, 1.0874e-04], device='cuda:4') +2023-03-27 10:53:49,443 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=161401.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:54:06,942 INFO [finetune.py:976] (4/7) Epoch 29, batch 1050, loss[loss=0.174, simple_loss=0.2531, pruned_loss=0.04749, over 4814.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2436, pruned_loss=0.04807, over 952221.31 frames. ], batch size: 39, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:54:13,089 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5837, 1.4729, 1.3641, 1.4758, 1.9048, 1.8552, 1.7147, 1.3968], + device='cuda:4'), covar=tensor([0.0330, 0.0365, 0.0707, 0.0350, 0.0218, 0.0437, 0.0286, 0.0422], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0107, 0.0149, 0.0113, 0.0102, 0.0118, 0.0105, 0.0115], + device='cuda:4'), out_proj_covar=tensor([7.9509e-05, 8.1921e-05, 1.1587e-04, 8.5688e-05, 7.9166e-05, 8.6512e-05, + 7.8162e-05, 8.7466e-05], device='cuda:4') +2023-03-27 10:54:28,547 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=161460.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:54:30,827 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161462.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:54:39,302 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.696e+01 1.443e+02 1.825e+02 2.108e+02 3.139e+02, threshold=3.650e+02, percent-clipped=0.0 +2023-03-27 10:54:39,917 INFO [finetune.py:976] (4/7) Epoch 29, batch 1100, loss[loss=0.1531, simple_loss=0.2352, pruned_loss=0.03547, over 4743.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2447, pruned_loss=0.04856, over 953490.59 frames. ], batch size: 27, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:55:11,594 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161521.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:55:14,359 INFO [finetune.py:976] (4/7) Epoch 29, batch 1150, loss[loss=0.1613, simple_loss=0.2343, pruned_loss=0.04412, over 4897.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2443, pruned_loss=0.04787, over 954155.82 frames. ], batch size: 46, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:55:15,171 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.63 vs. limit=2.0 +2023-03-27 10:55:33,557 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 10:55:39,776 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=161557.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:55:50,718 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161572.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:55:51,850 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.769e+01 1.420e+02 1.779e+02 2.194e+02 3.095e+02, threshold=3.558e+02, percent-clipped=0.0 +2023-03-27 10:55:52,954 INFO [finetune.py:976] (4/7) Epoch 29, batch 1200, loss[loss=0.1646, simple_loss=0.2382, pruned_loss=0.0455, over 4802.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2439, pruned_loss=0.04815, over 953093.08 frames. ], batch size: 29, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:56:03,030 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161587.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:56:21,545 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-27 10:56:21,955 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-03-27 10:56:22,444 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161605.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:56:33,596 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=161621.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:56:36,409 INFO [finetune.py:976] (4/7) Epoch 29, batch 1250, loss[loss=0.1583, simple_loss=0.2241, pruned_loss=0.04621, over 4818.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2412, pruned_loss=0.04771, over 952677.18 frames. ], batch size: 25, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:56:40,736 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-27 10:56:51,967 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7894, 1.0802, 1.8739, 1.8532, 1.6655, 1.5917, 1.7522, 1.8471], + device='cuda:4'), covar=tensor([0.3771, 0.3700, 0.2896, 0.3171, 0.4102, 0.3491, 0.3781, 0.2733], + device='cuda:4'), in_proj_covar=tensor([0.0271, 0.0251, 0.0271, 0.0300, 0.0300, 0.0277, 0.0307, 0.0255], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:57:07,718 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161669.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:57:07,785 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6434, 2.5655, 2.1871, 1.1075, 2.2929, 1.9358, 1.8332, 2.3791], + device='cuda:4'), covar=tensor([0.0838, 0.0677, 0.1458, 0.1993, 0.1415, 0.2375, 0.2164, 0.0853], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0188, 0.0200, 0.0179, 0.0208, 0.0209, 0.0222, 0.0194], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:57:07,796 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.6044, 1.6061, 1.5884, 0.8908, 1.7938, 1.9899, 1.9226, 1.5050], + device='cuda:4'), covar=tensor([0.0951, 0.0620, 0.0490, 0.0586, 0.0475, 0.0549, 0.0327, 0.0747], + device='cuda:4'), in_proj_covar=tensor([0.0121, 0.0148, 0.0131, 0.0122, 0.0132, 0.0130, 0.0143, 0.0152], + device='cuda:4'), out_proj_covar=tensor([8.8020e-05, 1.0589e-04, 9.3013e-05, 8.5697e-05, 9.2468e-05, 9.2013e-05, + 1.0156e-04, 1.0853e-04], device='cuda:4') +2023-03-27 10:57:10,131 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6689, 1.6115, 1.4949, 1.6025, 1.9687, 1.9856, 1.7683, 1.4878], + device='cuda:4'), covar=tensor([0.0381, 0.0330, 0.0643, 0.0347, 0.0232, 0.0480, 0.0332, 0.0473], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0107, 0.0149, 0.0112, 0.0102, 0.0118, 0.0105, 0.0115], + device='cuda:4'), out_proj_covar=tensor([7.9511e-05, 8.1784e-05, 1.1567e-04, 8.5501e-05, 7.9208e-05, 8.6742e-05, + 7.8247e-05, 8.7668e-05], device='cuda:4') +2023-03-27 10:57:11,726 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.453e+02 1.676e+02 2.058e+02 3.499e+02, threshold=3.351e+02, percent-clipped=0.0 +2023-03-27 10:57:12,869 INFO [finetune.py:976] (4/7) Epoch 29, batch 1300, loss[loss=0.1462, simple_loss=0.2207, pruned_loss=0.03587, over 4895.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2378, pruned_loss=0.04645, over 953669.77 frames. ], batch size: 32, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:57:54,740 INFO [finetune.py:976] (4/7) Epoch 29, batch 1350, loss[loss=0.1717, simple_loss=0.2445, pruned_loss=0.04946, over 4866.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2381, pruned_loss=0.0463, over 955578.88 frames. ], batch size: 34, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:58:29,781 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161757.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:58:40,926 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.1541, 1.3643, 1.3880, 0.7470, 1.3666, 1.5720, 1.5881, 1.2812], + device='cuda:4'), covar=tensor([0.0948, 0.0597, 0.0579, 0.0506, 0.0600, 0.0639, 0.0371, 0.0876], + device='cuda:4'), in_proj_covar=tensor([0.0120, 0.0147, 0.0131, 0.0122, 0.0132, 0.0130, 0.0142, 0.0152], + device='cuda:4'), out_proj_covar=tensor([8.7740e-05, 1.0549e-04, 9.2820e-05, 8.5420e-05, 9.2262e-05, 9.2014e-05, + 1.0131e-04, 1.0832e-04], device='cuda:4') +2023-03-27 10:58:47,851 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.11 vs. limit=2.0 +2023-03-27 10:58:49,497 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.001e+02 1.561e+02 1.829e+02 2.338e+02 6.236e+02, threshold=3.658e+02, percent-clipped=4.0 +2023-03-27 10:58:50,571 INFO [finetune.py:976] (4/7) Epoch 29, batch 1400, loss[loss=0.1229, simple_loss=0.1946, pruned_loss=0.02557, over 4770.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2414, pruned_loss=0.04724, over 956362.24 frames. ], batch size: 26, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:59:17,529 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3993, 2.4505, 2.4894, 1.7441, 2.3413, 2.7249, 2.6574, 2.1499], + device='cuda:4'), covar=tensor([0.0629, 0.0653, 0.0747, 0.0873, 0.0880, 0.0713, 0.0617, 0.1065], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0139, 0.0141, 0.0119, 0.0129, 0.0141, 0.0141, 0.0164], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 10:59:17,540 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=161815.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:59:18,092 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=161816.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:59:23,991 INFO [finetune.py:976] (4/7) Epoch 29, batch 1450, loss[loss=0.1602, simple_loss=0.2508, pruned_loss=0.03481, over 4768.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.242, pruned_loss=0.04715, over 955421.34 frames. ], batch size: 28, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:59:39,011 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-03-27 10:59:55,400 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=161872.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 10:59:56,503 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.509e+02 1.810e+02 2.171e+02 4.074e+02, threshold=3.620e+02, percent-clipped=1.0 +2023-03-27 10:59:57,117 INFO [finetune.py:976] (4/7) Epoch 29, batch 1500, loss[loss=0.1991, simple_loss=0.2625, pruned_loss=0.06785, over 4900.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2446, pruned_loss=0.0487, over 957003.21 frames. ], batch size: 36, lr: 2.85e-03, grad_scale: 16.0 +2023-03-27 10:59:58,316 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=161876.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:00:05,870 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=161887.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:00:17,536 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3086, 2.2486, 1.6836, 2.2908, 2.1705, 1.8901, 2.5790, 2.2946], + device='cuda:4'), covar=tensor([0.1260, 0.2001, 0.2961, 0.2460, 0.2432, 0.1673, 0.2923, 0.1576], + device='cuda:4'), in_proj_covar=tensor([0.0188, 0.0188, 0.0235, 0.0251, 0.0248, 0.0207, 0.0214, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:00:27,815 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161920.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:00:32,886 INFO [finetune.py:976] (4/7) Epoch 29, batch 1550, loss[loss=0.1865, simple_loss=0.2525, pruned_loss=0.06025, over 4899.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.246, pruned_loss=0.04947, over 955898.69 frames. ], batch size: 43, lr: 2.85e-03, grad_scale: 32.0 +2023-03-27 11:00:43,265 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4149, 1.2678, 1.7594, 2.3078, 1.5809, 2.0863, 1.1771, 2.0772], + device='cuda:4'), covar=tensor([0.1542, 0.1331, 0.0936, 0.0742, 0.0895, 0.1699, 0.1288, 0.0509], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0115, 0.0132, 0.0164, 0.0100, 0.0136, 0.0125, 0.0101], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 11:00:44,438 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=161935.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:01:10,618 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1584, 2.0513, 1.8412, 1.7633, 1.9944, 1.9805, 1.9784, 2.6279], + device='cuda:4'), covar=tensor([0.3511, 0.3520, 0.3019, 0.3255, 0.3347, 0.2342, 0.3115, 0.1571], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0264, 0.0238, 0.0276, 0.0261, 0.0232, 0.0259, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:01:16,694 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.413e+02 1.667e+02 1.984e+02 3.261e+02, threshold=3.334e+02, percent-clipped=0.0 +2023-03-27 11:01:16,864 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3950, 2.3100, 1.8525, 2.4528, 2.2985, 2.0293, 2.6742, 2.3880], + device='cuda:4'), covar=tensor([0.1310, 0.2124, 0.3092, 0.2460, 0.2613, 0.1751, 0.2894, 0.1751], + device='cuda:4'), in_proj_covar=tensor([0.0189, 0.0189, 0.0235, 0.0251, 0.0249, 0.0207, 0.0214, 0.0202], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:01:17,327 INFO [finetune.py:976] (4/7) Epoch 29, batch 1600, loss[loss=0.148, simple_loss=0.2229, pruned_loss=0.03656, over 4888.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2443, pruned_loss=0.04908, over 957289.96 frames. ], batch size: 32, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:01:46,197 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5527, 1.3941, 1.8464, 1.6521, 1.6414, 3.2712, 1.4905, 1.5803], + device='cuda:4'), covar=tensor([0.0940, 0.1883, 0.1056, 0.1023, 0.1557, 0.0249, 0.1483, 0.1805], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 11:01:59,455 INFO [finetune.py:976] (4/7) Epoch 29, batch 1650, loss[loss=0.1548, simple_loss=0.2306, pruned_loss=0.03956, over 4904.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2407, pruned_loss=0.04788, over 955615.12 frames. ], batch size: 32, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:02:01,064 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-27 11:02:22,370 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162057.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:02:34,950 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.499e+01 1.468e+02 1.758e+02 2.171e+02 4.899e+02, threshold=3.516e+02, percent-clipped=3.0 +2023-03-27 11:02:35,582 INFO [finetune.py:976] (4/7) Epoch 29, batch 1700, loss[loss=0.1338, simple_loss=0.2128, pruned_loss=0.0274, over 4748.00 frames. ], tot_loss[loss=0.166, simple_loss=0.238, pruned_loss=0.04703, over 953141.07 frames. ], batch size: 59, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:02:36,906 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=162077.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:02:55,085 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3647, 2.3686, 2.5443, 2.0011, 2.6486, 2.9544, 2.8530, 1.7815], + device='cuda:4'), covar=tensor([0.0809, 0.0895, 0.0890, 0.0927, 0.0744, 0.0755, 0.0730, 0.1921], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0138, 0.0140, 0.0118, 0.0128, 0.0140, 0.0140, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:03:01,315 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162105.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:03:08,006 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162116.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:03:16,128 INFO [finetune.py:976] (4/7) Epoch 29, batch 1750, loss[loss=0.1422, simple_loss=0.2182, pruned_loss=0.03305, over 4757.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2394, pruned_loss=0.04746, over 953110.14 frames. ], batch size: 23, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:03:16,234 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0812, 2.0994, 2.2292, 1.7634, 2.1187, 2.5196, 2.4216, 1.7967], + device='cuda:4'), covar=tensor([0.0754, 0.0793, 0.0829, 0.0873, 0.1139, 0.0747, 0.0653, 0.1357], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0137, 0.0140, 0.0118, 0.0128, 0.0140, 0.0140, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:03:24,615 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=162138.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:03:26,059 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.83 vs. limit=5.0 +2023-03-27 11:03:41,214 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3205, 2.2090, 1.7376, 2.0661, 2.1511, 1.8583, 2.4380, 2.2451], + device='cuda:4'), covar=tensor([0.1221, 0.1800, 0.2820, 0.2443, 0.2453, 0.1763, 0.2698, 0.1518], + device='cuda:4'), in_proj_covar=tensor([0.0190, 0.0191, 0.0237, 0.0253, 0.0250, 0.0209, 0.0216, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:03:59,950 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162164.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:04:04,284 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=162171.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:04:10,045 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.681e+01 1.626e+02 1.856e+02 2.284e+02 4.131e+02, threshold=3.712e+02, percent-clipped=2.0 +2023-03-27 11:04:10,647 INFO [finetune.py:976] (4/7) Epoch 29, batch 1800, loss[loss=0.22, simple_loss=0.295, pruned_loss=0.07249, over 4793.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2426, pruned_loss=0.04781, over 954562.21 frames. ], batch size: 45, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:04:30,953 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5790, 1.0806, 0.7356, 1.3898, 2.0155, 0.8266, 1.2567, 1.2969], + device='cuda:4'), covar=tensor([0.1403, 0.2126, 0.1635, 0.1180, 0.1731, 0.1814, 0.1403, 0.2061], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0093, 0.0108, 0.0093, 0.0119, 0.0092, 0.0097, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 11:04:40,283 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3541, 1.4194, 1.4752, 1.1384, 1.3504, 1.5169, 1.3636, 1.6725], + device='cuda:4'), covar=tensor([0.1183, 0.2065, 0.1441, 0.1484, 0.1011, 0.1243, 0.2857, 0.0961], + device='cuda:4'), in_proj_covar=tensor([0.0190, 0.0204, 0.0191, 0.0187, 0.0173, 0.0211, 0.0216, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:04:44,452 INFO [finetune.py:976] (4/7) Epoch 29, batch 1850, loss[loss=0.1883, simple_loss=0.2707, pruned_loss=0.05297, over 4911.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2434, pruned_loss=0.04806, over 954653.25 frames. ], batch size: 36, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:04:48,179 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=162231.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:05:17,383 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.145e+02 1.604e+02 1.802e+02 2.267e+02 3.875e+02, threshold=3.605e+02, percent-clipped=1.0 +2023-03-27 11:05:17,854 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-27 11:05:17,995 INFO [finetune.py:976] (4/7) Epoch 29, batch 1900, loss[loss=0.157, simple_loss=0.2334, pruned_loss=0.04036, over 4872.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.245, pruned_loss=0.04908, over 954770.52 frames. ], batch size: 35, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:05:28,434 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=162292.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:05:39,668 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 11:05:51,672 INFO [finetune.py:976] (4/7) Epoch 29, batch 1950, loss[loss=0.1591, simple_loss=0.2304, pruned_loss=0.04395, over 4889.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2442, pruned_loss=0.04877, over 956096.41 frames. ], batch size: 32, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:06:28,892 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8320, 1.7676, 1.8699, 1.2035, 1.8601, 1.9197, 1.8683, 1.5488], + device='cuda:4'), covar=tensor([0.0493, 0.0625, 0.0597, 0.0862, 0.0812, 0.0560, 0.0540, 0.1111], + device='cuda:4'), in_proj_covar=tensor([0.0129, 0.0136, 0.0138, 0.0117, 0.0126, 0.0138, 0.0138, 0.0161], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:06:36,622 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.724e+01 1.430e+02 1.855e+02 2.130e+02 3.474e+02, threshold=3.709e+02, percent-clipped=0.0 +2023-03-27 11:06:37,251 INFO [finetune.py:976] (4/7) Epoch 29, batch 2000, loss[loss=0.1673, simple_loss=0.2327, pruned_loss=0.05094, over 4826.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2411, pruned_loss=0.04752, over 954932.99 frames. ], batch size: 30, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:07:03,737 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.96 vs. limit=5.0 +2023-03-27 11:07:14,707 INFO [finetune.py:976] (4/7) Epoch 29, batch 2050, loss[loss=0.1572, simple_loss=0.2286, pruned_loss=0.04293, over 4836.00 frames. ], tot_loss[loss=0.1641, simple_loss=0.2367, pruned_loss=0.0457, over 957176.25 frames. ], batch size: 30, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:07:19,649 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=162433.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:07:45,861 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162471.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:07:47,599 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.116e+01 1.601e+02 1.964e+02 2.330e+02 4.874e+02, threshold=3.928e+02, percent-clipped=3.0 +2023-03-27 11:07:48,229 INFO [finetune.py:976] (4/7) Epoch 29, batch 2100, loss[loss=0.1719, simple_loss=0.2452, pruned_loss=0.04925, over 4718.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2375, pruned_loss=0.04641, over 954844.48 frames. ], batch size: 23, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:08:22,565 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=162511.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:08:27,343 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.51 vs. limit=2.0 +2023-03-27 11:08:28,428 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162519.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:08:34,946 INFO [finetune.py:976] (4/7) Epoch 29, batch 2150, loss[loss=0.1759, simple_loss=0.2636, pruned_loss=0.04411, over 4785.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2404, pruned_loss=0.04723, over 952502.72 frames. ], batch size: 29, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:08:36,160 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7264, 1.6065, 2.3224, 2.0582, 1.9954, 4.4313, 1.6418, 1.8820], + device='cuda:4'), covar=tensor([0.1025, 0.1979, 0.1099, 0.0982, 0.1621, 0.0159, 0.1594, 0.1890], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 11:09:17,088 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=162572.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 11:09:18,163 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.053e+02 1.475e+02 1.779e+02 2.116e+02 3.329e+02, threshold=3.558e+02, percent-clipped=0.0 +2023-03-27 11:09:18,786 INFO [finetune.py:976] (4/7) Epoch 29, batch 2200, loss[loss=0.1827, simple_loss=0.2564, pruned_loss=0.05454, over 4899.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2423, pruned_loss=0.04806, over 953897.47 frames. ], batch size: 36, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:09:27,182 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=162587.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:09:29,122 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-27 11:10:00,797 INFO [finetune.py:976] (4/7) Epoch 29, batch 2250, loss[loss=0.1425, simple_loss=0.2144, pruned_loss=0.03525, over 4682.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2433, pruned_loss=0.04813, over 954619.93 frames. ], batch size: 23, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:10:33,516 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.222e+01 1.558e+02 1.861e+02 2.054e+02 5.864e+02, threshold=3.723e+02, percent-clipped=1.0 +2023-03-27 11:10:34,114 INFO [finetune.py:976] (4/7) Epoch 29, batch 2300, loss[loss=0.184, simple_loss=0.2532, pruned_loss=0.05735, over 4907.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2438, pruned_loss=0.0485, over 953829.09 frames. ], batch size: 37, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:10:36,995 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7685, 1.8551, 1.6725, 1.6817, 2.3961, 2.4656, 2.1551, 1.9145], + device='cuda:4'), covar=tensor([0.0508, 0.0372, 0.0593, 0.0377, 0.0243, 0.0488, 0.0301, 0.0427], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0106, 0.0149, 0.0112, 0.0102, 0.0117, 0.0105, 0.0115], + device='cuda:4'), out_proj_covar=tensor([7.9099e-05, 8.1069e-05, 1.1560e-04, 8.4941e-05, 7.8937e-05, 8.6108e-05, + 7.7721e-05, 8.7255e-05], device='cuda:4') +2023-03-27 11:10:42,245 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4479, 1.3723, 1.9147, 1.7713, 1.5707, 3.3421, 1.3459, 1.6191], + device='cuda:4'), covar=tensor([0.1038, 0.1813, 0.1077, 0.0941, 0.1638, 0.0241, 0.1556, 0.1727], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0080, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 11:10:48,267 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1883, 2.0804, 2.1496, 1.4725, 2.0919, 2.2468, 2.2024, 1.7606], + device='cuda:4'), covar=tensor([0.0516, 0.0575, 0.0613, 0.0849, 0.0707, 0.0568, 0.0532, 0.1130], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0137, 0.0140, 0.0119, 0.0128, 0.0140, 0.0140, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:11:09,593 INFO [finetune.py:976] (4/7) Epoch 29, batch 2350, loss[loss=0.1349, simple_loss=0.2033, pruned_loss=0.03328, over 4913.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2424, pruned_loss=0.04862, over 955934.92 frames. ], batch size: 46, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:11:20,033 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162733.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:11:50,493 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.467e+01 1.478e+02 1.866e+02 2.325e+02 4.091e+02, threshold=3.732e+02, percent-clipped=2.0 +2023-03-27 11:11:51,107 INFO [finetune.py:976] (4/7) Epoch 29, batch 2400, loss[loss=0.1471, simple_loss=0.2091, pruned_loss=0.04257, over 4861.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2402, pruned_loss=0.04825, over 957524.65 frames. ], batch size: 31, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:11:56,643 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162781.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:12:33,622 INFO [finetune.py:976] (4/7) Epoch 29, batch 2450, loss[loss=0.1761, simple_loss=0.2381, pruned_loss=0.05704, over 4750.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2367, pruned_loss=0.04694, over 957937.37 frames. ], batch size: 28, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:12:35,959 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=162827.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:12:47,139 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6355, 3.6644, 3.5234, 1.8175, 3.7584, 2.8904, 1.0297, 2.6104], + device='cuda:4'), covar=tensor([0.2341, 0.2382, 0.1494, 0.3163, 0.1030, 0.1013, 0.4315, 0.1494], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0181, 0.0160, 0.0129, 0.0163, 0.0124, 0.0150, 0.0126], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 11:13:02,270 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=162867.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 11:13:09,319 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.738e+01 1.481e+02 1.739e+02 2.016e+02 4.521e+02, threshold=3.479e+02, percent-clipped=1.0 +2023-03-27 11:13:09,951 INFO [finetune.py:976] (4/7) Epoch 29, batch 2500, loss[loss=0.1338, simple_loss=0.2116, pruned_loss=0.02803, over 4761.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2376, pruned_loss=0.04704, over 957851.16 frames. ], batch size: 27, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:13:27,484 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=162887.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:13:28,107 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=162888.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:13:40,483 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5424, 2.1154, 2.7884, 1.5611, 2.3874, 2.6420, 1.9398, 2.7047], + device='cuda:4'), covar=tensor([0.1120, 0.1918, 0.1473, 0.2263, 0.0813, 0.1418, 0.2518, 0.0779], + device='cuda:4'), in_proj_covar=tensor([0.0190, 0.0206, 0.0193, 0.0189, 0.0174, 0.0212, 0.0218, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:13:52,411 INFO [finetune.py:976] (4/7) Epoch 29, batch 2550, loss[loss=0.1595, simple_loss=0.2334, pruned_loss=0.04286, over 4779.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2413, pruned_loss=0.04757, over 955984.58 frames. ], batch size: 29, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:14:01,596 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=162935.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:14:16,850 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.61 vs. limit=5.0 +2023-03-27 11:14:29,284 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1480, 3.5848, 3.7795, 3.9012, 3.9326, 3.6595, 4.2223, 1.5708], + device='cuda:4'), covar=tensor([0.0732, 0.0886, 0.0787, 0.0939, 0.1096, 0.1541, 0.0655, 0.5305], + device='cuda:4'), in_proj_covar=tensor([0.0350, 0.0244, 0.0284, 0.0295, 0.0337, 0.0284, 0.0305, 0.0300], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:14:36,962 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.986e+01 1.516e+02 1.791e+02 2.248e+02 3.208e+02, threshold=3.582e+02, percent-clipped=0.0 +2023-03-27 11:14:37,596 INFO [finetune.py:976] (4/7) Epoch 29, batch 2600, loss[loss=0.1532, simple_loss=0.2246, pruned_loss=0.04093, over 4871.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2438, pruned_loss=0.04888, over 955711.34 frames. ], batch size: 34, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:14:52,698 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-27 11:15:17,741 INFO [finetune.py:976] (4/7) Epoch 29, batch 2650, loss[loss=0.1633, simple_loss=0.2514, pruned_loss=0.0376, over 4800.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2445, pruned_loss=0.04886, over 956422.88 frames. ], batch size: 41, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:15:51,094 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.228e+01 1.484e+02 1.662e+02 2.061e+02 3.704e+02, threshold=3.324e+02, percent-clipped=1.0 +2023-03-27 11:15:51,724 INFO [finetune.py:976] (4/7) Epoch 29, batch 2700, loss[loss=0.1603, simple_loss=0.2212, pruned_loss=0.04974, over 4776.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2437, pruned_loss=0.04838, over 956164.82 frames. ], batch size: 23, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:16:02,256 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=163090.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:16:21,022 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9884, 1.4438, 2.0153, 2.0007, 1.7982, 1.7656, 1.9866, 1.8932], + device='cuda:4'), covar=tensor([0.3879, 0.3727, 0.3149, 0.3438, 0.4550, 0.3695, 0.3996, 0.2863], + device='cuda:4'), in_proj_covar=tensor([0.0269, 0.0249, 0.0269, 0.0299, 0.0298, 0.0276, 0.0304, 0.0254], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:16:25,140 INFO [finetune.py:976] (4/7) Epoch 29, batch 2750, loss[loss=0.1536, simple_loss=0.2176, pruned_loss=0.04474, over 4064.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2418, pruned_loss=0.04799, over 955017.81 frames. ], batch size: 17, lr: 2.84e-03, grad_scale: 32.0 +2023-03-27 11:16:38,213 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9525, 1.7045, 2.0622, 1.3708, 1.8589, 2.1030, 1.6099, 2.2755], + device='cuda:4'), covar=tensor([0.1144, 0.1849, 0.1545, 0.1824, 0.0888, 0.1246, 0.2681, 0.0784], + device='cuda:4'), in_proj_covar=tensor([0.0193, 0.0208, 0.0195, 0.0191, 0.0176, 0.0214, 0.0220, 0.0199], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:16:52,785 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=163151.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:17:03,843 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=163167.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 11:17:10,458 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.013e+02 1.483e+02 1.714e+02 1.995e+02 3.279e+02, threshold=3.429e+02, percent-clipped=0.0 +2023-03-27 11:17:10,474 INFO [finetune.py:976] (4/7) Epoch 29, batch 2800, loss[loss=0.1651, simple_loss=0.238, pruned_loss=0.04608, over 4923.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2391, pruned_loss=0.04702, over 955045.32 frames. ], batch size: 28, lr: 2.84e-03, grad_scale: 16.0 +2023-03-27 11:17:15,477 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=163183.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:17:37,945 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=163215.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:17:44,448 INFO [finetune.py:976] (4/7) Epoch 29, batch 2850, loss[loss=0.135, simple_loss=0.2054, pruned_loss=0.03233, over 4780.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2362, pruned_loss=0.04537, over 955366.67 frames. ], batch size: 26, lr: 2.84e-03, grad_scale: 16.0 +2023-03-27 11:17:47,622 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5628, 3.6222, 3.4290, 1.8981, 3.6680, 3.0265, 0.9231, 2.5809], + device='cuda:4'), covar=tensor([0.2958, 0.2290, 0.1646, 0.3283, 0.1134, 0.0992, 0.4472, 0.1693], + device='cuda:4'), in_proj_covar=tensor([0.0152, 0.0181, 0.0160, 0.0130, 0.0163, 0.0124, 0.0150, 0.0126], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 11:17:59,041 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=163240.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:18:17,896 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5071, 1.5622, 1.3392, 1.3963, 1.8767, 1.7941, 1.6029, 1.3880], + device='cuda:4'), covar=tensor([0.0355, 0.0318, 0.0640, 0.0366, 0.0212, 0.0436, 0.0353, 0.0433], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0106, 0.0149, 0.0112, 0.0102, 0.0117, 0.0105, 0.0115], + device='cuda:4'), out_proj_covar=tensor([7.9207e-05, 8.1235e-05, 1.1557e-04, 8.5055e-05, 7.9148e-05, 8.6331e-05, + 7.7956e-05, 8.7190e-05], device='cuda:4') +2023-03-27 11:18:31,674 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.921e+01 1.528e+02 1.890e+02 2.266e+02 3.566e+02, threshold=3.779e+02, percent-clipped=1.0 +2023-03-27 11:18:31,690 INFO [finetune.py:976] (4/7) Epoch 29, batch 2900, loss[loss=0.1556, simple_loss=0.2358, pruned_loss=0.03774, over 4762.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2393, pruned_loss=0.04696, over 956569.17 frames. ], batch size: 27, lr: 2.84e-03, grad_scale: 16.0 +2023-03-27 11:18:42,083 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-27 11:18:52,056 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=163301.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:19:08,248 INFO [finetune.py:976] (4/7) Epoch 29, batch 2950, loss[loss=0.1673, simple_loss=0.2396, pruned_loss=0.04749, over 4819.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2416, pruned_loss=0.04783, over 954988.10 frames. ], batch size: 39, lr: 2.84e-03, grad_scale: 16.0 +2023-03-27 11:19:13,784 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.5373, 3.9431, 4.1872, 4.3725, 4.2844, 4.0666, 4.6342, 1.3702], + device='cuda:4'), covar=tensor([0.0728, 0.0904, 0.0824, 0.0879, 0.1141, 0.1547, 0.0610, 0.5775], + device='cuda:4'), in_proj_covar=tensor([0.0353, 0.0246, 0.0287, 0.0297, 0.0340, 0.0287, 0.0306, 0.0303], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:19:45,706 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-27 11:19:49,296 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.216e+02 1.530e+02 1.927e+02 2.309e+02 4.929e+02, threshold=3.855e+02, percent-clipped=3.0 +2023-03-27 11:19:49,311 INFO [finetune.py:976] (4/7) Epoch 29, batch 3000, loss[loss=0.1486, simple_loss=0.2374, pruned_loss=0.0299, over 4819.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2432, pruned_loss=0.04822, over 956073.45 frames. ], batch size: 38, lr: 2.84e-03, grad_scale: 16.0 +2023-03-27 11:19:49,312 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 11:19:56,043 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.2214, 1.2401, 1.1731, 1.1943, 1.4451, 1.4826, 1.2883, 1.2031], + device='cuda:4'), covar=tensor([0.0536, 0.0278, 0.0677, 0.0305, 0.0276, 0.0390, 0.0354, 0.0383], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0107, 0.0150, 0.0112, 0.0103, 0.0118, 0.0106, 0.0116], + device='cuda:4'), out_proj_covar=tensor([7.9539e-05, 8.1654e-05, 1.1625e-04, 8.5512e-05, 7.9793e-05, 8.6870e-05, + 7.8495e-05, 8.7811e-05], device='cuda:4') +2023-03-27 11:20:05,062 INFO [finetune.py:1010] (4/7) Epoch 29, validation: loss=0.158, simple_loss=0.2251, pruned_loss=0.04545, over 2265189.00 frames. +2023-03-27 11:20:05,062 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6557MB +2023-03-27 11:20:43,039 INFO [finetune.py:976] (4/7) Epoch 29, batch 3050, loss[loss=0.1499, simple_loss=0.2368, pruned_loss=0.03153, over 4820.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2445, pruned_loss=0.04824, over 956162.08 frames. ], batch size: 41, lr: 2.84e-03, grad_scale: 16.0 +2023-03-27 11:20:57,940 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=163446.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:21:13,971 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5489, 1.6150, 2.2935, 1.8719, 1.8612, 4.3225, 1.6813, 1.8585], + device='cuda:4'), covar=tensor([0.1008, 0.1817, 0.1160, 0.0962, 0.1576, 0.0138, 0.1452, 0.1762], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0083, 0.0073, 0.0077, 0.0092, 0.0081, 0.0086, 0.0081], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 11:21:16,336 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.820e+01 1.385e+02 1.622e+02 1.937e+02 3.745e+02, threshold=3.244e+02, percent-clipped=0.0 +2023-03-27 11:21:16,352 INFO [finetune.py:976] (4/7) Epoch 29, batch 3100, loss[loss=0.1278, simple_loss=0.2034, pruned_loss=0.0261, over 4743.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.242, pruned_loss=0.0474, over 954891.48 frames. ], batch size: 23, lr: 2.84e-03, grad_scale: 16.0 +2023-03-27 11:21:22,351 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=163483.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:21:38,441 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1999, 1.4141, 1.8026, 1.6113, 1.5804, 3.3151, 1.4207, 1.6079], + device='cuda:4'), covar=tensor([0.1057, 0.1832, 0.1062, 0.0936, 0.1615, 0.0210, 0.1426, 0.1709], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0083, 0.0073, 0.0077, 0.0092, 0.0081, 0.0086, 0.0081], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 11:21:51,334 INFO [finetune.py:976] (4/7) Epoch 29, batch 3150, loss[loss=0.1449, simple_loss=0.2156, pruned_loss=0.03714, over 4888.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2382, pruned_loss=0.04613, over 955413.81 frames. ], batch size: 32, lr: 2.84e-03, grad_scale: 16.0 +2023-03-27 11:21:54,475 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.6540, 2.7704, 2.6017, 1.8867, 2.6200, 2.9977, 2.9251, 2.3546], + device='cuda:4'), covar=tensor([0.0562, 0.0479, 0.0669, 0.0798, 0.0826, 0.0590, 0.0495, 0.0962], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0137, 0.0140, 0.0118, 0.0128, 0.0139, 0.0139, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:21:55,513 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=163531.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:22:25,916 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-03-27 11:22:33,520 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.004e+02 1.497e+02 1.704e+02 2.227e+02 4.213e+02, threshold=3.407e+02, percent-clipped=5.0 +2023-03-27 11:22:33,536 INFO [finetune.py:976] (4/7) Epoch 29, batch 3200, loss[loss=0.1752, simple_loss=0.2453, pruned_loss=0.05253, over 4931.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2355, pruned_loss=0.04592, over 957686.54 frames. ], batch size: 33, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:22:49,205 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=163596.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:22:51,076 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1233, 2.0619, 2.1716, 1.3375, 2.1522, 2.2567, 2.2587, 1.8320], + device='cuda:4'), covar=tensor([0.0595, 0.0622, 0.0650, 0.0851, 0.0672, 0.0619, 0.0580, 0.1115], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0138, 0.0141, 0.0119, 0.0129, 0.0139, 0.0140, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:23:07,473 INFO [finetune.py:976] (4/7) Epoch 29, batch 3250, loss[loss=0.1497, simple_loss=0.2299, pruned_loss=0.0347, over 4803.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2366, pruned_loss=0.04674, over 956603.12 frames. ], batch size: 45, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:23:52,621 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.061e+02 1.629e+02 1.981e+02 2.506e+02 4.570e+02, threshold=3.962e+02, percent-clipped=6.0 +2023-03-27 11:23:52,637 INFO [finetune.py:976] (4/7) Epoch 29, batch 3300, loss[loss=0.1842, simple_loss=0.2663, pruned_loss=0.05104, over 4908.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2401, pruned_loss=0.04766, over 958378.35 frames. ], batch size: 35, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:24:32,331 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9847, 1.8853, 1.8082, 1.9371, 1.7292, 4.6217, 1.7858, 2.3629], + device='cuda:4'), covar=tensor([0.3247, 0.2384, 0.1997, 0.2247, 0.1454, 0.0111, 0.2338, 0.1047], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0117, 0.0121, 0.0125, 0.0114, 0.0096, 0.0094, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 11:24:33,176 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-27 11:24:37,006 INFO [finetune.py:976] (4/7) Epoch 29, batch 3350, loss[loss=0.169, simple_loss=0.244, pruned_loss=0.04701, over 4728.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2421, pruned_loss=0.04772, over 958380.80 frames. ], batch size: 54, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:24:55,820 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=163746.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:24:57,444 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-27 11:25:20,408 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.4194, 1.5487, 1.5363, 0.8435, 1.7001, 1.9512, 1.8891, 1.3795], + device='cuda:4'), covar=tensor([0.0929, 0.0741, 0.0636, 0.0598, 0.0548, 0.0711, 0.0342, 0.0859], + device='cuda:4'), in_proj_covar=tensor([0.0120, 0.0148, 0.0131, 0.0122, 0.0131, 0.0130, 0.0142, 0.0152], + device='cuda:4'), out_proj_covar=tensor([8.7900e-05, 1.0609e-04, 9.3270e-05, 8.5627e-05, 9.1978e-05, 9.2227e-05, + 1.0081e-04, 1.0808e-04], device='cuda:4') +2023-03-27 11:25:21,352 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.486e+02 1.801e+02 2.174e+02 4.171e+02, threshold=3.603e+02, percent-clipped=1.0 +2023-03-27 11:25:21,368 INFO [finetune.py:976] (4/7) Epoch 29, batch 3400, loss[loss=0.1535, simple_loss=0.24, pruned_loss=0.03348, over 4800.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2438, pruned_loss=0.04838, over 958514.58 frames. ], batch size: 29, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:25:37,708 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=163794.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:25:52,326 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-27 11:25:58,974 INFO [finetune.py:976] (4/7) Epoch 29, batch 3450, loss[loss=0.1898, simple_loss=0.2639, pruned_loss=0.05783, over 4799.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2437, pruned_loss=0.048, over 956159.10 frames. ], batch size: 29, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:26:00,938 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4427, 2.4288, 1.9114, 2.6140, 2.3804, 2.0213, 2.8915, 2.5294], + device='cuda:4'), covar=tensor([0.1285, 0.2067, 0.2813, 0.2401, 0.2451, 0.1553, 0.2928, 0.1452], + device='cuda:4'), in_proj_covar=tensor([0.0189, 0.0190, 0.0236, 0.0253, 0.0250, 0.0208, 0.0215, 0.0203], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:26:41,242 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.103e+01 1.539e+02 1.877e+02 2.200e+02 3.171e+02, threshold=3.754e+02, percent-clipped=0.0 +2023-03-27 11:26:41,258 INFO [finetune.py:976] (4/7) Epoch 29, batch 3500, loss[loss=0.1611, simple_loss=0.2326, pruned_loss=0.04484, over 4930.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2412, pruned_loss=0.04771, over 954685.99 frames. ], batch size: 38, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:26:42,585 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0574, 0.9347, 0.9727, 0.4691, 0.9031, 1.1828, 1.1451, 0.9472], + device='cuda:4'), covar=tensor([0.0805, 0.0696, 0.0559, 0.0504, 0.0568, 0.0583, 0.0427, 0.0684], + device='cuda:4'), in_proj_covar=tensor([0.0121, 0.0148, 0.0132, 0.0122, 0.0131, 0.0131, 0.0142, 0.0152], + device='cuda:4'), out_proj_covar=tensor([8.8118e-05, 1.0635e-04, 9.3568e-05, 8.5546e-05, 9.2085e-05, 9.2490e-05, + 1.0107e-04, 1.0837e-04], device='cuda:4') +2023-03-27 11:26:55,498 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=163896.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:27:17,138 INFO [finetune.py:976] (4/7) Epoch 29, batch 3550, loss[loss=0.1661, simple_loss=0.2442, pruned_loss=0.04405, over 4829.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2385, pruned_loss=0.04694, over 956058.94 frames. ], batch size: 33, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:27:38,584 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=163944.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:27:52,934 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0462, 1.2664, 1.5292, 1.1979, 1.4273, 2.3447, 1.2182, 1.4231], + device='cuda:4'), covar=tensor([0.0955, 0.1720, 0.0870, 0.0865, 0.1517, 0.0353, 0.1431, 0.1624], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0083, 0.0073, 0.0076, 0.0091, 0.0081, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 11:27:59,308 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.875e+01 1.446e+02 1.734e+02 2.162e+02 4.667e+02, threshold=3.468e+02, percent-clipped=1.0 +2023-03-27 11:27:59,324 INFO [finetune.py:976] (4/7) Epoch 29, batch 3600, loss[loss=0.1924, simple_loss=0.2605, pruned_loss=0.0621, over 4755.00 frames. ], tot_loss[loss=0.1631, simple_loss=0.2352, pruned_loss=0.04549, over 956235.27 frames. ], batch size: 54, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:28:00,642 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.8402, 2.5345, 2.4668, 1.3016, 2.6388, 2.0410, 1.9155, 2.4191], + device='cuda:4'), covar=tensor([0.1228, 0.0809, 0.1626, 0.2156, 0.1532, 0.2454, 0.2200, 0.1111], + device='cuda:4'), in_proj_covar=tensor([0.0172, 0.0189, 0.0202, 0.0182, 0.0210, 0.0211, 0.0225, 0.0196], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:28:15,323 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-27 11:28:26,164 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([5.2221, 4.5431, 4.7711, 5.0130, 4.9531, 4.6298, 5.3087, 1.7301], + device='cuda:4'), covar=tensor([0.0692, 0.0850, 0.0782, 0.0884, 0.1180, 0.1573, 0.0510, 0.5962], + device='cuda:4'), in_proj_covar=tensor([0.0353, 0.0246, 0.0286, 0.0296, 0.0340, 0.0285, 0.0306, 0.0302], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:28:31,995 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4475, 1.3639, 1.7656, 2.4448, 1.6585, 2.2296, 0.9906, 2.1897], + device='cuda:4'), covar=tensor([0.1566, 0.1324, 0.1017, 0.0727, 0.0879, 0.1156, 0.1524, 0.0525], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0114, 0.0132, 0.0163, 0.0100, 0.0135, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 11:28:36,281 INFO [finetune.py:976] (4/7) Epoch 29, batch 3650, loss[loss=0.2133, simple_loss=0.2842, pruned_loss=0.07125, over 4802.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2383, pruned_loss=0.04735, over 953933.43 frames. ], batch size: 51, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:29:19,111 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.932e+01 1.541e+02 1.828e+02 2.331e+02 7.133e+02, threshold=3.656e+02, percent-clipped=4.0 +2023-03-27 11:29:19,127 INFO [finetune.py:976] (4/7) Epoch 29, batch 3700, loss[loss=0.1531, simple_loss=0.2317, pruned_loss=0.03725, over 4829.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2413, pruned_loss=0.04805, over 954608.87 frames. ], batch size: 33, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:29:36,833 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164090.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:30:00,905 INFO [finetune.py:976] (4/7) Epoch 29, batch 3750, loss[loss=0.1851, simple_loss=0.2574, pruned_loss=0.05644, over 4821.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2417, pruned_loss=0.04786, over 955017.53 frames. ], batch size: 39, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:30:17,227 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164151.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:30:20,023 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164155.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:30:37,150 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.292e+01 1.655e+02 1.832e+02 2.179e+02 3.362e+02, threshold=3.664e+02, percent-clipped=0.0 +2023-03-27 11:30:37,166 INFO [finetune.py:976] (4/7) Epoch 29, batch 3800, loss[loss=0.1804, simple_loss=0.2513, pruned_loss=0.05478, over 4786.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2442, pruned_loss=0.04897, over 952869.19 frames. ], batch size: 25, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:30:57,289 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4194, 1.5709, 1.3524, 1.4391, 1.8015, 1.7949, 1.5160, 1.4156], + device='cuda:4'), covar=tensor([0.0393, 0.0301, 0.0573, 0.0319, 0.0240, 0.0438, 0.0310, 0.0376], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0107, 0.0148, 0.0112, 0.0103, 0.0118, 0.0104, 0.0115], + device='cuda:4'), out_proj_covar=tensor([7.9027e-05, 8.1594e-05, 1.1520e-04, 8.5009e-05, 7.9385e-05, 8.6631e-05, + 7.7524e-05, 8.7181e-05], device='cuda:4') +2023-03-27 11:31:03,878 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164216.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:31:09,782 INFO [finetune.py:976] (4/7) Epoch 29, batch 3850, loss[loss=0.1499, simple_loss=0.2171, pruned_loss=0.04137, over 4760.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2418, pruned_loss=0.04783, over 954007.15 frames. ], batch size: 28, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:31:29,008 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7295, 1.6330, 1.5329, 1.6692, 1.2490, 3.6551, 1.3888, 1.7950], + device='cuda:4'), covar=tensor([0.3277, 0.2501, 0.2093, 0.2376, 0.1617, 0.0197, 0.2645, 0.1197], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0116, 0.0120, 0.0124, 0.0113, 0.0095, 0.0093, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 11:31:45,644 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.748e+01 1.394e+02 1.747e+02 2.221e+02 3.425e+02, threshold=3.494e+02, percent-clipped=0.0 +2023-03-27 11:31:45,660 INFO [finetune.py:976] (4/7) Epoch 29, batch 3900, loss[loss=0.1529, simple_loss=0.2111, pruned_loss=0.04731, over 4110.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2392, pruned_loss=0.04717, over 953418.90 frames. ], batch size: 18, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:32:27,502 INFO [finetune.py:976] (4/7) Epoch 29, batch 3950, loss[loss=0.1323, simple_loss=0.2106, pruned_loss=0.02703, over 4812.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2362, pruned_loss=0.04662, over 951375.03 frames. ], batch size: 25, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:32:56,818 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0102, 1.3975, 2.0279, 2.0371, 1.8101, 1.7753, 1.9283, 1.9448], + device='cuda:4'), covar=tensor([0.3873, 0.4061, 0.3629, 0.3720, 0.5183, 0.4191, 0.4905, 0.3044], + device='cuda:4'), in_proj_covar=tensor([0.0270, 0.0250, 0.0270, 0.0300, 0.0299, 0.0277, 0.0305, 0.0254], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:32:58,648 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164360.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:33:11,864 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.819e+01 1.491e+02 1.748e+02 1.970e+02 3.605e+02, threshold=3.496e+02, percent-clipped=1.0 +2023-03-27 11:33:11,880 INFO [finetune.py:976] (4/7) Epoch 29, batch 4000, loss[loss=0.1566, simple_loss=0.2393, pruned_loss=0.03691, over 4930.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2359, pruned_loss=0.04658, over 952467.75 frames. ], batch size: 38, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:33:29,401 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.45 vs. limit=5.0 +2023-03-27 11:33:42,995 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164421.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:33:45,305 INFO [finetune.py:976] (4/7) Epoch 29, batch 4050, loss[loss=0.1779, simple_loss=0.2548, pruned_loss=0.05052, over 4861.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2385, pruned_loss=0.04719, over 953143.53 frames. ], batch size: 31, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:34:06,773 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164446.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:34:27,956 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6596, 1.5846, 1.4047, 1.7244, 1.9580, 1.7505, 1.3726, 1.4078], + device='cuda:4'), covar=tensor([0.2146, 0.1884, 0.1929, 0.1563, 0.1574, 0.1208, 0.2479, 0.1885], + device='cuda:4'), in_proj_covar=tensor([0.0247, 0.0212, 0.0217, 0.0199, 0.0246, 0.0190, 0.0218, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:34:29,030 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.788e+01 1.600e+02 1.814e+02 2.182e+02 3.811e+02, threshold=3.628e+02, percent-clipped=1.0 +2023-03-27 11:34:29,046 INFO [finetune.py:976] (4/7) Epoch 29, batch 4100, loss[loss=0.1822, simple_loss=0.2485, pruned_loss=0.05794, over 4896.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2423, pruned_loss=0.04851, over 952119.46 frames. ], batch size: 37, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:35:04,641 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164511.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:35:13,895 INFO [finetune.py:976] (4/7) Epoch 29, batch 4150, loss[loss=0.159, simple_loss=0.2323, pruned_loss=0.04289, over 4787.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2424, pruned_loss=0.04827, over 950239.92 frames. ], batch size: 29, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:35:42,375 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164568.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:35:46,473 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.025e+02 1.582e+02 1.835e+02 2.360e+02 4.097e+02, threshold=3.670e+02, percent-clipped=1.0 +2023-03-27 11:35:46,489 INFO [finetune.py:976] (4/7) Epoch 29, batch 4200, loss[loss=0.1664, simple_loss=0.2365, pruned_loss=0.04814, over 4838.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2435, pruned_loss=0.04852, over 949574.73 frames. ], batch size: 49, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:35:48,889 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164578.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:35:50,742 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7555, 1.0573, 1.8677, 1.7786, 1.6383, 1.5611, 1.6800, 1.8463], + device='cuda:4'), covar=tensor([0.3833, 0.3691, 0.3009, 0.3378, 0.4519, 0.3569, 0.4172, 0.2716], + device='cuda:4'), in_proj_covar=tensor([0.0269, 0.0250, 0.0270, 0.0300, 0.0300, 0.0277, 0.0305, 0.0254], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:36:13,901 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6347, 1.5387, 1.3911, 1.7330, 1.6582, 1.6792, 1.0694, 1.4062], + device='cuda:4'), covar=tensor([0.2288, 0.1985, 0.2060, 0.1678, 0.1556, 0.1238, 0.2471, 0.1933], + device='cuda:4'), in_proj_covar=tensor([0.0247, 0.0212, 0.0217, 0.0199, 0.0247, 0.0191, 0.0218, 0.0206], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:36:20,292 INFO [finetune.py:976] (4/7) Epoch 29, batch 4250, loss[loss=0.1644, simple_loss=0.2327, pruned_loss=0.04805, over 4932.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2425, pruned_loss=0.04815, over 951454.16 frames. ], batch size: 33, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:36:21,018 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0980, 1.9278, 2.0529, 1.3083, 2.0176, 2.1516, 2.1218, 1.7158], + device='cuda:4'), covar=tensor([0.0573, 0.0735, 0.0673, 0.0871, 0.0748, 0.0646, 0.0563, 0.1178], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0138, 0.0141, 0.0118, 0.0129, 0.0139, 0.0140, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:36:23,253 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164629.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:36:29,286 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164639.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 11:36:41,355 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6938, 1.5743, 1.5141, 1.6340, 1.4475, 3.6863, 1.4747, 1.8429], + device='cuda:4'), covar=tensor([0.3364, 0.2459, 0.2217, 0.2373, 0.1596, 0.0192, 0.2596, 0.1209], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0117, 0.0121, 0.0124, 0.0114, 0.0096, 0.0094, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 11:36:43,173 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164658.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:36:53,779 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.101e+02 1.450e+02 1.663e+02 2.112e+02 3.483e+02, threshold=3.326e+02, percent-clipped=0.0 +2023-03-27 11:36:53,795 INFO [finetune.py:976] (4/7) Epoch 29, batch 4300, loss[loss=0.1726, simple_loss=0.2524, pruned_loss=0.04642, over 4851.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.24, pruned_loss=0.04765, over 955060.09 frames. ], batch size: 44, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:37:12,580 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-03-27 11:37:32,864 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2503, 2.1843, 1.8619, 2.0001, 2.2282, 1.9814, 2.3484, 2.2481], + device='cuda:4'), covar=tensor([0.1336, 0.1919, 0.2845, 0.2332, 0.2507, 0.1786, 0.2754, 0.1761], + device='cuda:4'), in_proj_covar=tensor([0.0191, 0.0192, 0.0238, 0.0255, 0.0252, 0.0210, 0.0216, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:37:39,418 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164716.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:37:41,333 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164719.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:37:44,824 INFO [finetune.py:976] (4/7) Epoch 29, batch 4350, loss[loss=0.1839, simple_loss=0.2428, pruned_loss=0.06246, over 4798.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2366, pruned_loss=0.04653, over 953429.18 frames. ], batch size: 45, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:37:56,919 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8873, 1.3652, 1.9859, 1.9237, 1.7387, 1.6729, 1.9020, 1.9184], + device='cuda:4'), covar=tensor([0.3844, 0.3728, 0.2882, 0.3367, 0.4127, 0.3485, 0.3955, 0.2735], + device='cuda:4'), in_proj_covar=tensor([0.0271, 0.0251, 0.0271, 0.0301, 0.0300, 0.0278, 0.0306, 0.0256], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:37:58,708 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=164746.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:37:58,722 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=164746.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:38:16,045 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.0049, 4.9748, 4.6767, 3.0721, 5.0168, 3.8984, 1.0437, 3.5401], + device='cuda:4'), covar=tensor([0.2378, 0.2210, 0.1332, 0.2939, 0.0786, 0.0778, 0.4828, 0.1519], + device='cuda:4'), in_proj_covar=tensor([0.0151, 0.0180, 0.0160, 0.0130, 0.0163, 0.0124, 0.0149, 0.0126], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:4') +2023-03-27 11:38:20,795 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.196e+01 1.395e+02 1.786e+02 2.060e+02 3.738e+02, threshold=3.572e+02, percent-clipped=2.0 +2023-03-27 11:38:20,811 INFO [finetune.py:976] (4/7) Epoch 29, batch 4400, loss[loss=0.2022, simple_loss=0.2794, pruned_loss=0.06248, over 4751.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.236, pruned_loss=0.04624, over 949834.04 frames. ], batch size: 54, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:38:33,467 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=164794.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:38:43,377 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=164807.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:38:46,288 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=164811.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:38:47,551 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2535, 1.9048, 2.6345, 1.6631, 2.1525, 2.3673, 1.7315, 2.5970], + device='cuda:4'), covar=tensor([0.1357, 0.2249, 0.1756, 0.2492, 0.1083, 0.1686, 0.3153, 0.0964], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0207, 0.0194, 0.0190, 0.0175, 0.0213, 0.0219, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:38:54,769 INFO [finetune.py:976] (4/7) Epoch 29, batch 4450, loss[loss=0.1359, simple_loss=0.2103, pruned_loss=0.03078, over 4748.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2383, pruned_loss=0.04631, over 950968.06 frames. ], batch size: 23, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:39:19,094 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.2186, 3.6480, 3.8447, 4.0322, 3.9759, 3.6877, 4.2866, 1.2202], + device='cuda:4'), covar=tensor([0.0924, 0.0964, 0.1151, 0.1115, 0.1338, 0.1738, 0.0785, 0.6084], + device='cuda:4'), in_proj_covar=tensor([0.0356, 0.0249, 0.0288, 0.0298, 0.0340, 0.0288, 0.0309, 0.0304], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:39:23,087 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=164859.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:39:31,772 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3280, 1.6152, 0.7327, 2.1260, 2.5051, 1.7731, 1.8351, 2.1147], + device='cuda:4'), covar=tensor([0.1391, 0.2008, 0.2130, 0.1138, 0.1771, 0.1829, 0.1380, 0.1888], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0109, 0.0093, 0.0120, 0.0092, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 11:39:37,134 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.009e+02 1.552e+02 1.889e+02 2.217e+02 4.905e+02, threshold=3.778e+02, percent-clipped=1.0 +2023-03-27 11:39:37,150 INFO [finetune.py:976] (4/7) Epoch 29, batch 4500, loss[loss=0.1593, simple_loss=0.2438, pruned_loss=0.0374, over 4825.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2405, pruned_loss=0.04645, over 952936.76 frames. ], batch size: 33, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:39:44,185 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.61 vs. limit=2.0 +2023-03-27 11:40:22,143 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164924.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:40:22,678 INFO [finetune.py:976] (4/7) Epoch 29, batch 4550, loss[loss=0.157, simple_loss=0.2374, pruned_loss=0.03825, over 4883.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2423, pruned_loss=0.04693, over 954820.51 frames. ], batch size: 32, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:40:24,993 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-03-27 11:40:28,158 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=164934.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 11:40:51,828 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4623, 1.3878, 1.6018, 2.4981, 1.6913, 2.1963, 0.9031, 2.2786], + device='cuda:4'), covar=tensor([0.1656, 0.1269, 0.1135, 0.0664, 0.0902, 0.1157, 0.1579, 0.0490], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0115, 0.0132, 0.0164, 0.0100, 0.0135, 0.0124, 0.0100], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 11:40:55,995 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.494e+02 1.764e+02 2.048e+02 3.220e+02, threshold=3.528e+02, percent-clipped=0.0 +2023-03-27 11:40:56,011 INFO [finetune.py:976] (4/7) Epoch 29, batch 4600, loss[loss=0.1404, simple_loss=0.2105, pruned_loss=0.0351, over 4795.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2413, pruned_loss=0.04645, over 955235.15 frames. ], batch size: 29, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:41:17,213 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.54 vs. limit=5.0 +2023-03-27 11:41:22,218 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=165014.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:41:23,386 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165016.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:41:29,267 INFO [finetune.py:976] (4/7) Epoch 29, batch 4650, loss[loss=0.1306, simple_loss=0.2052, pruned_loss=0.02803, over 4690.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2386, pruned_loss=0.04592, over 953551.93 frames. ], batch size: 23, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:41:53,747 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165064.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:42:00,863 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.2477, 2.1680, 1.9542, 2.3550, 2.6916, 2.3157, 2.2817, 1.7234], + device='cuda:4'), covar=tensor([0.2047, 0.1903, 0.1829, 0.1459, 0.1685, 0.1094, 0.1901, 0.1944], + device='cuda:4'), in_proj_covar=tensor([0.0247, 0.0212, 0.0217, 0.0199, 0.0246, 0.0191, 0.0218, 0.0207], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:42:01,328 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.062e+02 1.478e+02 1.798e+02 2.114e+02 1.110e+03, threshold=3.597e+02, percent-clipped=3.0 +2023-03-27 11:42:01,344 INFO [finetune.py:976] (4/7) Epoch 29, batch 4700, loss[loss=0.1369, simple_loss=0.2062, pruned_loss=0.03381, over 4819.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2362, pruned_loss=0.04542, over 955507.98 frames. ], batch size: 38, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:42:27,142 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=165102.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:42:42,542 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.9010, 2.3458, 3.2404, 2.1989, 2.6529, 3.2572, 2.3194, 3.1598], + device='cuda:4'), covar=tensor([0.1111, 0.1961, 0.1245, 0.1692, 0.0971, 0.1006, 0.2487, 0.0716], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0207, 0.0194, 0.0190, 0.0175, 0.0214, 0.0220, 0.0198], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:42:43,770 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=165123.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:42:44,886 INFO [finetune.py:976] (4/7) Epoch 29, batch 4750, loss[loss=0.1939, simple_loss=0.2601, pruned_loss=0.06386, over 4896.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2351, pruned_loss=0.04546, over 952971.98 frames. ], batch size: 32, lr: 2.83e-03, grad_scale: 16.0 +2023-03-27 11:42:45,671 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1233, 2.0305, 1.7452, 1.9708, 1.9249, 1.8915, 1.9544, 2.6098], + device='cuda:4'), covar=tensor([0.3697, 0.3960, 0.3316, 0.3457, 0.4023, 0.2481, 0.3476, 0.1673], + device='cuda:4'), in_proj_covar=tensor([0.0288, 0.0263, 0.0239, 0.0273, 0.0261, 0.0232, 0.0259, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:43:21,527 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.139e+02 1.579e+02 1.786e+02 2.031e+02 3.721e+02, threshold=3.572e+02, percent-clipped=1.0 +2023-03-27 11:43:21,543 INFO [finetune.py:976] (4/7) Epoch 29, batch 4800, loss[loss=0.188, simple_loss=0.2683, pruned_loss=0.05385, over 4832.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2392, pruned_loss=0.04749, over 952008.94 frames. ], batch size: 47, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:43:27,585 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=165184.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:43:44,989 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6432, 1.2302, 0.8049, 1.5221, 2.0618, 1.0660, 1.4248, 1.4834], + device='cuda:4'), covar=tensor([0.1519, 0.2183, 0.1930, 0.1285, 0.2012, 0.1926, 0.1475, 0.2033], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0093, 0.0108, 0.0092, 0.0119, 0.0092, 0.0097, 0.0088], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 11:43:53,910 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165224.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:43:54,433 INFO [finetune.py:976] (4/7) Epoch 29, batch 4850, loss[loss=0.1544, simple_loss=0.2367, pruned_loss=0.03601, over 4823.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2416, pruned_loss=0.04726, over 952979.42 frames. ], batch size: 39, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:44:00,506 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165234.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 11:44:09,608 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.5049, 1.4948, 1.4755, 0.9850, 1.5750, 1.8505, 1.7497, 1.4207], + device='cuda:4'), covar=tensor([0.0966, 0.0713, 0.0505, 0.0485, 0.0439, 0.0536, 0.0352, 0.0709], + device='cuda:4'), in_proj_covar=tensor([0.0121, 0.0147, 0.0131, 0.0121, 0.0131, 0.0130, 0.0141, 0.0151], + device='cuda:4'), out_proj_covar=tensor([8.8115e-05, 1.0507e-04, 9.3020e-05, 8.4911e-05, 9.1732e-05, 9.2130e-05, + 1.0050e-04, 1.0798e-04], device='cuda:4') +2023-03-27 11:44:24,480 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165272.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:44:26,732 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.028e+02 1.742e+02 1.982e+02 2.317e+02 4.079e+02, threshold=3.965e+02, percent-clipped=3.0 +2023-03-27 11:44:26,747 INFO [finetune.py:976] (4/7) Epoch 29, batch 4900, loss[loss=0.1704, simple_loss=0.2536, pruned_loss=0.04355, over 4826.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2421, pruned_loss=0.04722, over 952057.18 frames. ], batch size: 33, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:44:41,265 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165282.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:45:03,725 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165314.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:45:16,148 INFO [finetune.py:976] (4/7) Epoch 29, batch 4950, loss[loss=0.1781, simple_loss=0.2541, pruned_loss=0.051, over 4914.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2428, pruned_loss=0.04771, over 952225.79 frames. ], batch size: 42, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:45:48,524 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165362.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:45:56,866 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.088e+01 1.449e+02 1.743e+02 2.087e+02 3.629e+02, threshold=3.486e+02, percent-clipped=0.0 +2023-03-27 11:45:56,882 INFO [finetune.py:976] (4/7) Epoch 29, batch 5000, loss[loss=0.2059, simple_loss=0.2705, pruned_loss=0.07069, over 4918.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2415, pruned_loss=0.04754, over 953338.54 frames. ], batch size: 33, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:46:12,869 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.4285, 1.6062, 1.6647, 0.9574, 1.7640, 1.9685, 1.8796, 1.5179], + device='cuda:4'), covar=tensor([0.1004, 0.0737, 0.0539, 0.0529, 0.0467, 0.0751, 0.0358, 0.0840], + device='cuda:4'), in_proj_covar=tensor([0.0120, 0.0146, 0.0130, 0.0121, 0.0131, 0.0130, 0.0141, 0.0150], + device='cuda:4'), out_proj_covar=tensor([8.7816e-05, 1.0434e-04, 9.2506e-05, 8.4544e-05, 9.1594e-05, 9.1757e-05, + 1.0006e-04, 1.0746e-04], device='cuda:4') +2023-03-27 11:46:15,402 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165402.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:46:20,905 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.5785, 2.4548, 2.1487, 2.6922, 2.4776, 2.3606, 2.3436, 3.3389], + device='cuda:4'), covar=tensor([0.3630, 0.4738, 0.3279, 0.3857, 0.3850, 0.2546, 0.4214, 0.1546], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0264, 0.0238, 0.0273, 0.0261, 0.0232, 0.0259, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:46:26,293 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.57 vs. limit=2.0 +2023-03-27 11:46:30,165 INFO [finetune.py:976] (4/7) Epoch 29, batch 5050, loss[loss=0.1609, simple_loss=0.2386, pruned_loss=0.04158, over 4786.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2401, pruned_loss=0.04767, over 954186.83 frames. ], batch size: 25, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:46:33,358 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4243, 2.2870, 1.8411, 2.3545, 2.2935, 1.9975, 2.5348, 2.3663], + device='cuda:4'), covar=tensor([0.1265, 0.1911, 0.2878, 0.2296, 0.2340, 0.1721, 0.3078, 0.1669], + device='cuda:4'), in_proj_covar=tensor([0.0191, 0.0191, 0.0237, 0.0254, 0.0251, 0.0209, 0.0216, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:46:39,566 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.0705, 0.9749, 0.9391, 0.4035, 0.8758, 1.1679, 1.1628, 1.0075], + device='cuda:4'), covar=tensor([0.0883, 0.0681, 0.0651, 0.0567, 0.0641, 0.0625, 0.0429, 0.0726], + device='cuda:4'), in_proj_covar=tensor([0.0120, 0.0146, 0.0130, 0.0120, 0.0131, 0.0130, 0.0141, 0.0150], + device='cuda:4'), out_proj_covar=tensor([8.7792e-05, 1.0426e-04, 9.2521e-05, 8.4401e-05, 9.1708e-05, 9.1836e-05, + 1.0011e-04, 1.0742e-04], device='cuda:4') +2023-03-27 11:46:47,815 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165450.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:47:03,405 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.397e+02 1.700e+02 2.095e+02 3.650e+02, threshold=3.400e+02, percent-clipped=1.0 +2023-03-27 11:47:03,421 INFO [finetune.py:976] (4/7) Epoch 29, batch 5100, loss[loss=0.1877, simple_loss=0.2464, pruned_loss=0.06453, over 4895.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2378, pruned_loss=0.04737, over 954435.34 frames. ], batch size: 35, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:47:06,335 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=165479.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:47:23,650 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-27 11:47:37,023 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-27 11:47:46,455 INFO [finetune.py:976] (4/7) Epoch 29, batch 5150, loss[loss=0.1689, simple_loss=0.2349, pruned_loss=0.05145, over 4892.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2379, pruned_loss=0.0475, over 951935.15 frames. ], batch size: 32, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:47:52,163 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-27 11:47:57,772 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.82 vs. limit=5.0 +2023-03-27 11:48:04,763 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=165551.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:48:10,222 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1331, 2.1169, 1.7314, 2.0760, 2.0179, 1.9050, 1.9991, 2.7337], + device='cuda:4'), covar=tensor([0.3738, 0.4090, 0.3465, 0.4081, 0.4108, 0.2635, 0.3879, 0.1666], + device='cuda:4'), in_proj_covar=tensor([0.0289, 0.0264, 0.0239, 0.0274, 0.0261, 0.0232, 0.0259, 0.0238], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:48:20,242 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.818e+01 1.606e+02 1.856e+02 2.249e+02 3.990e+02, threshold=3.713e+02, percent-clipped=3.0 +2023-03-27 11:48:20,258 INFO [finetune.py:976] (4/7) Epoch 29, batch 5200, loss[loss=0.2032, simple_loss=0.2714, pruned_loss=0.06751, over 4901.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2419, pruned_loss=0.04898, over 950746.06 frames. ], batch size: 43, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:48:32,983 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-27 11:48:35,433 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.1230, 3.5588, 3.8131, 3.9904, 3.9101, 3.6309, 4.1719, 1.2764], + device='cuda:4'), covar=tensor([0.0825, 0.1025, 0.1019, 0.1016, 0.1188, 0.1679, 0.0752, 0.6062], + device='cuda:4'), in_proj_covar=tensor([0.0357, 0.0249, 0.0289, 0.0299, 0.0340, 0.0288, 0.0308, 0.0304], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:48:45,762 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=165612.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:48:53,937 INFO [finetune.py:976] (4/7) Epoch 29, batch 5250, loss[loss=0.1926, simple_loss=0.2768, pruned_loss=0.05419, over 4803.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.243, pruned_loss=0.04878, over 951025.42 frames. ], batch size: 45, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:49:06,484 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-27 11:49:15,673 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8837, 1.8754, 1.6053, 2.1205, 2.5057, 2.0920, 1.8240, 1.5676], + device='cuda:4'), covar=tensor([0.2202, 0.1812, 0.1813, 0.1486, 0.1496, 0.1095, 0.2069, 0.1897], + device='cuda:4'), in_proj_covar=tensor([0.0247, 0.0213, 0.0217, 0.0200, 0.0247, 0.0192, 0.0218, 0.0208], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:49:26,774 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.319e+01 1.511e+02 1.772e+02 2.250e+02 3.554e+02, threshold=3.544e+02, percent-clipped=0.0 +2023-03-27 11:49:26,789 INFO [finetune.py:976] (4/7) Epoch 29, batch 5300, loss[loss=0.2351, simple_loss=0.2981, pruned_loss=0.08601, over 4839.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2441, pruned_loss=0.04898, over 952146.35 frames. ], batch size: 47, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:50:10,070 INFO [finetune.py:976] (4/7) Epoch 29, batch 5350, loss[loss=0.214, simple_loss=0.2789, pruned_loss=0.07452, over 4883.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2446, pruned_loss=0.04887, over 952468.75 frames. ], batch size: 35, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:50:11,457 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9170, 1.4723, 1.9638, 2.0007, 1.7662, 1.7121, 1.9343, 1.8837], + device='cuda:4'), covar=tensor([0.3690, 0.3794, 0.3132, 0.3391, 0.4924, 0.3813, 0.4227, 0.2848], + device='cuda:4'), in_proj_covar=tensor([0.0270, 0.0250, 0.0270, 0.0301, 0.0301, 0.0278, 0.0306, 0.0256], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:50:14,345 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5359, 1.5904, 1.3836, 1.5816, 1.8765, 1.8175, 1.6388, 1.4059], + device='cuda:4'), covar=tensor([0.0407, 0.0291, 0.0640, 0.0308, 0.0235, 0.0488, 0.0297, 0.0461], + device='cuda:4'), in_proj_covar=tensor([0.0104, 0.0107, 0.0149, 0.0112, 0.0104, 0.0119, 0.0105, 0.0117], + device='cuda:4'), out_proj_covar=tensor([8.0285e-05, 8.2035e-05, 1.1608e-04, 8.5210e-05, 8.0218e-05, 8.7550e-05, + 7.8278e-05, 8.8689e-05], device='cuda:4') +2023-03-27 11:50:33,772 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=165750.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:50:52,346 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-27 11:50:59,727 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.046e+02 1.485e+02 1.788e+02 2.126e+02 3.231e+02, threshold=3.576e+02, percent-clipped=0.0 +2023-03-27 11:50:59,743 INFO [finetune.py:976] (4/7) Epoch 29, batch 5400, loss[loss=0.1724, simple_loss=0.2381, pruned_loss=0.05335, over 4815.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2431, pruned_loss=0.04874, over 952004.07 frames. ], batch size: 30, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:51:02,243 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=165779.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:51:18,372 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=165803.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:51:24,170 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=165811.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:51:33,003 INFO [finetune.py:976] (4/7) Epoch 29, batch 5450, loss[loss=0.1474, simple_loss=0.2206, pruned_loss=0.03708, over 4840.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2399, pruned_loss=0.04782, over 950489.89 frames. ], batch size: 47, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:51:34,301 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=165827.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:51:59,789 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=165864.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:52:06,351 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.128e+01 1.456e+02 1.721e+02 2.001e+02 5.868e+02, threshold=3.442e+02, percent-clipped=2.0 +2023-03-27 11:52:06,367 INFO [finetune.py:976] (4/7) Epoch 29, batch 5500, loss[loss=0.1385, simple_loss=0.2109, pruned_loss=0.03306, over 4858.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2365, pruned_loss=0.04625, over 951085.39 frames. ], batch size: 31, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:52:27,324 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=165907.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:52:40,105 INFO [finetune.py:976] (4/7) Epoch 29, batch 5550, loss[loss=0.1973, simple_loss=0.2794, pruned_loss=0.05759, over 4833.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2393, pruned_loss=0.04725, over 951562.57 frames. ], batch size: 47, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:53:03,002 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8822, 1.2620, 1.9690, 1.9517, 1.7445, 1.6856, 1.8281, 1.8765], + device='cuda:4'), covar=tensor([0.4312, 0.4141, 0.3337, 0.3538, 0.4841, 0.3863, 0.4263, 0.3013], + device='cuda:4'), in_proj_covar=tensor([0.0269, 0.0250, 0.0268, 0.0300, 0.0300, 0.0277, 0.0305, 0.0255], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:53:22,715 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.003e+02 1.570e+02 1.926e+02 2.203e+02 4.633e+02, threshold=3.853e+02, percent-clipped=1.0 +2023-03-27 11:53:22,731 INFO [finetune.py:976] (4/7) Epoch 29, batch 5600, loss[loss=0.2054, simple_loss=0.274, pruned_loss=0.06841, over 4871.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.243, pruned_loss=0.04792, over 954446.82 frames. ], batch size: 34, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:53:25,749 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3686, 1.3016, 1.7594, 1.6204, 1.4370, 3.1981, 1.2138, 1.3689], + device='cuda:4'), covar=tensor([0.1020, 0.1974, 0.1139, 0.0977, 0.1779, 0.0225, 0.1616, 0.1906], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0083, 0.0073, 0.0076, 0.0091, 0.0081, 0.0086, 0.0081], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 11:53:54,010 INFO [finetune.py:976] (4/7) Epoch 29, batch 5650, loss[loss=0.1356, simple_loss=0.2032, pruned_loss=0.03402, over 4710.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2445, pruned_loss=0.04778, over 954084.34 frames. ], batch size: 23, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:53:57,030 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1474, 2.0300, 1.6476, 2.1817, 2.0514, 1.8193, 2.3613, 2.1247], + device='cuda:4'), covar=tensor([0.1323, 0.1875, 0.2903, 0.2196, 0.2566, 0.1687, 0.2371, 0.1767], + device='cuda:4'), in_proj_covar=tensor([0.0191, 0.0191, 0.0238, 0.0253, 0.0251, 0.0210, 0.0216, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:54:11,182 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166054.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 11:54:23,584 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.094e+01 1.435e+02 1.719e+02 2.053e+02 3.744e+02, threshold=3.437e+02, percent-clipped=0.0 +2023-03-27 11:54:23,600 INFO [finetune.py:976] (4/7) Epoch 29, batch 5700, loss[loss=0.1354, simple_loss=0.2047, pruned_loss=0.0331, over 4296.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2405, pruned_loss=0.04666, over 936018.52 frames. ], batch size: 18, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:54:50,337 INFO [finetune.py:976] (4/7) Epoch 30, batch 0, loss[loss=0.1388, simple_loss=0.2179, pruned_loss=0.02986, over 4782.00 frames. ], tot_loss[loss=0.1388, simple_loss=0.2179, pruned_loss=0.02986, over 4782.00 frames. ], batch size: 25, lr: 2.82e-03, grad_scale: 32.0 +2023-03-27 11:54:50,337 INFO [finetune.py:1001] (4/7) Computing validation loss +2023-03-27 11:54:52,273 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8115, 1.1168, 1.9736, 1.8821, 1.7336, 1.6568, 1.7490, 1.9231], + device='cuda:4'), covar=tensor([0.4355, 0.4266, 0.3387, 0.3703, 0.4919, 0.3872, 0.4548, 0.3188], + device='cuda:4'), in_proj_covar=tensor([0.0270, 0.0250, 0.0269, 0.0301, 0.0300, 0.0277, 0.0306, 0.0255], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:54:57,645 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1819, 1.9067, 1.8067, 1.7772, 1.8800, 1.8856, 1.8845, 2.5805], + device='cuda:4'), covar=tensor([0.3572, 0.4370, 0.3164, 0.3788, 0.4199, 0.2596, 0.3998, 0.1671], + device='cuda:4'), in_proj_covar=tensor([0.0290, 0.0264, 0.0239, 0.0274, 0.0261, 0.0232, 0.0259, 0.0239], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:55:06,737 INFO [finetune.py:1010] (4/7) Epoch 30, validation: loss=0.1598, simple_loss=0.2264, pruned_loss=0.04658, over 2265189.00 frames. +2023-03-27 11:55:06,738 INFO [finetune.py:1011] (4/7) Maximum memory allocated so far is 6557MB +2023-03-27 11:55:11,874 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166106.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:55:21,140 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166115.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 11:55:41,977 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-03-27 11:55:43,739 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166147.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:55:52,370 INFO [finetune.py:976] (4/7) Epoch 30, batch 50, loss[loss=0.1694, simple_loss=0.248, pruned_loss=0.04538, over 4781.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2413, pruned_loss=0.04624, over 214384.16 frames. ], batch size: 29, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 11:55:54,012 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-27 11:55:55,000 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-03-27 11:56:02,575 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166159.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:56:16,049 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.060e+01 1.403e+02 1.686e+02 1.983e+02 3.736e+02, threshold=3.372e+02, percent-clipped=1.0 +2023-03-27 11:56:32,528 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1876, 2.0374, 2.2180, 1.3878, 2.1105, 2.2672, 2.1358, 1.7758], + device='cuda:4'), covar=tensor([0.0595, 0.0701, 0.0611, 0.0859, 0.0719, 0.0635, 0.0622, 0.1115], + device='cuda:4'), in_proj_covar=tensor([0.0133, 0.0139, 0.0142, 0.0119, 0.0130, 0.0140, 0.0141, 0.0164], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 11:56:34,675 INFO [finetune.py:976] (4/7) Epoch 30, batch 100, loss[loss=0.164, simple_loss=0.2224, pruned_loss=0.05281, over 4221.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2353, pruned_loss=0.04555, over 377699.18 frames. ], batch size: 65, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 11:56:38,211 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166207.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:56:39,337 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166208.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:57:01,540 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-27 11:57:07,203 INFO [finetune.py:976] (4/7) Epoch 30, batch 150, loss[loss=0.1558, simple_loss=0.2284, pruned_loss=0.04164, over 4816.00 frames. ], tot_loss[loss=0.1605, simple_loss=0.232, pruned_loss=0.04453, over 507968.04 frames. ], batch size: 33, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 11:57:08,955 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166255.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:57:21,825 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.042e+02 1.442e+02 1.804e+02 2.095e+02 4.016e+02, threshold=3.609e+02, percent-clipped=1.0 +2023-03-27 11:57:39,806 INFO [finetune.py:976] (4/7) Epoch 30, batch 200, loss[loss=0.1677, simple_loss=0.2418, pruned_loss=0.0468, over 4767.00 frames. ], tot_loss[loss=0.162, simple_loss=0.2327, pruned_loss=0.04561, over 609712.81 frames. ], batch size: 26, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 11:57:42,836 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166307.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:58:00,976 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-27 11:58:14,802 INFO [finetune.py:976] (4/7) Epoch 30, batch 250, loss[loss=0.1869, simple_loss=0.2583, pruned_loss=0.05776, over 4825.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2374, pruned_loss=0.04691, over 686191.54 frames. ], batch size: 40, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 11:58:19,111 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.5826, 1.6833, 1.5929, 0.8560, 1.8116, 2.0142, 1.8971, 1.4777], + device='cuda:4'), covar=tensor([0.0991, 0.0752, 0.0603, 0.0665, 0.0480, 0.0598, 0.0389, 0.0759], + device='cuda:4'), in_proj_covar=tensor([0.0121, 0.0147, 0.0131, 0.0121, 0.0131, 0.0130, 0.0141, 0.0151], + device='cuda:4'), out_proj_covar=tensor([8.8058e-05, 1.0484e-04, 9.2994e-05, 8.4829e-05, 9.1985e-05, 9.2202e-05, + 1.0038e-04, 1.0789e-04], device='cuda:4') +2023-03-27 11:58:26,045 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166368.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:58:26,671 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.8553, 2.0718, 1.7557, 1.7088, 2.4600, 2.5870, 2.0753, 2.0793], + device='cuda:4'), covar=tensor([0.0551, 0.0366, 0.0598, 0.0418, 0.0250, 0.0499, 0.0386, 0.0431], + device='cuda:4'), in_proj_covar=tensor([0.0104, 0.0107, 0.0149, 0.0112, 0.0104, 0.0119, 0.0105, 0.0117], + device='cuda:4'), out_proj_covar=tensor([8.0245e-05, 8.2049e-05, 1.1591e-04, 8.4978e-05, 8.0031e-05, 8.7710e-05, + 7.7878e-05, 8.8393e-05], device='cuda:4') +2023-03-27 11:58:30,118 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.080e+02 1.508e+02 1.830e+02 2.341e+02 4.348e+02, threshold=3.661e+02, percent-clipped=2.0 +2023-03-27 11:58:48,139 INFO [finetune.py:976] (4/7) Epoch 30, batch 300, loss[loss=0.1729, simple_loss=0.2496, pruned_loss=0.0481, over 4899.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2391, pruned_loss=0.04681, over 745242.77 frames. ], batch size: 32, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 11:58:50,065 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166406.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:58:52,502 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166410.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 11:59:02,391 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-03-27 11:59:21,478 INFO [finetune.py:976] (4/7) Epoch 30, batch 350, loss[loss=0.1832, simple_loss=0.2631, pruned_loss=0.05165, over 4907.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2422, pruned_loss=0.04774, over 793628.15 frames. ], batch size: 37, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 11:59:22,660 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166454.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:59:25,726 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166459.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:59:37,687 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.476e+02 1.813e+02 2.161e+02 3.890e+02, threshold=3.626e+02, percent-clipped=2.0 +2023-03-27 11:59:41,432 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166481.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:59:55,112 INFO [finetune.py:976] (4/7) Epoch 30, batch 400, loss[loss=0.1569, simple_loss=0.2302, pruned_loss=0.04183, over 4783.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2429, pruned_loss=0.04748, over 830423.04 frames. ], batch size: 51, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 11:59:55,184 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166503.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:59:58,011 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166507.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:59:59,266 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166509.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 11:59:59,344 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-27 12:00:20,212 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166531.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:00:35,350 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166542.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:00:45,992 INFO [finetune.py:976] (4/7) Epoch 30, batch 450, loss[loss=0.1841, simple_loss=0.2499, pruned_loss=0.05914, over 4927.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2413, pruned_loss=0.04718, over 856521.01 frames. ], batch size: 33, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:00:57,399 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166570.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:01:00,793 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.283e+01 1.415e+02 1.694e+02 2.083e+02 4.695e+02, threshold=3.388e+02, percent-clipped=2.0 +2023-03-27 12:01:08,424 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1708, 2.0425, 1.6624, 2.1700, 2.1046, 1.8574, 2.5403, 2.1543], + device='cuda:4'), covar=tensor([0.1353, 0.2140, 0.3161, 0.2434, 0.2696, 0.1845, 0.2519, 0.1876], + device='cuda:4'), in_proj_covar=tensor([0.0191, 0.0191, 0.0238, 0.0254, 0.0251, 0.0210, 0.0216, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:01:22,340 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166592.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:01:32,869 INFO [finetune.py:976] (4/7) Epoch 30, batch 500, loss[loss=0.1896, simple_loss=0.2602, pruned_loss=0.05952, over 4888.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.24, pruned_loss=0.0468, over 878536.46 frames. ], batch size: 32, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:01:46,187 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([3.3097, 2.9219, 3.0561, 3.2571, 3.0891, 2.8948, 3.3584, 1.0516], + device='cuda:4'), covar=tensor([0.1165, 0.1117, 0.1223, 0.1060, 0.1795, 0.2029, 0.1216, 0.6030], + device='cuda:4'), in_proj_covar=tensor([0.0355, 0.0249, 0.0288, 0.0298, 0.0339, 0.0288, 0.0306, 0.0304], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:02:06,079 INFO [finetune.py:976] (4/7) Epoch 30, batch 550, loss[loss=0.2154, simple_loss=0.266, pruned_loss=0.0824, over 4911.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2382, pruned_loss=0.04674, over 894675.20 frames. ], batch size: 32, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:02:12,199 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166663.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:02:20,414 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 7.871e+01 1.519e+02 1.786e+02 2.255e+02 3.834e+02, threshold=3.573e+02, percent-clipped=3.0 +2023-03-27 12:02:32,829 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5166, 1.4602, 1.9573, 1.7092, 1.5804, 3.4387, 1.4080, 1.5500], + device='cuda:4'), covar=tensor([0.1043, 0.1854, 0.1055, 0.0969, 0.1641, 0.0240, 0.1530, 0.1829], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0082, 0.0073, 0.0076, 0.0091, 0.0081, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 12:02:34,063 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1035, 1.9741, 1.7984, 2.1294, 2.5996, 2.1955, 2.0984, 1.7941], + device='cuda:4'), covar=tensor([0.1647, 0.1666, 0.1536, 0.1298, 0.1398, 0.0983, 0.1905, 0.1562], + device='cuda:4'), in_proj_covar=tensor([0.0250, 0.0215, 0.0219, 0.0202, 0.0250, 0.0194, 0.0221, 0.0210], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:02:34,084 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9666, 1.6762, 2.1572, 1.9422, 1.7938, 1.7268, 1.9046, 2.0681], + device='cuda:4'), covar=tensor([0.3299, 0.3305, 0.2511, 0.3372, 0.4205, 0.3786, 0.3982, 0.2401], + device='cuda:4'), in_proj_covar=tensor([0.0270, 0.0250, 0.0269, 0.0300, 0.0300, 0.0278, 0.0305, 0.0255], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:02:39,362 INFO [finetune.py:976] (4/7) Epoch 30, batch 600, loss[loss=0.1748, simple_loss=0.2412, pruned_loss=0.05418, over 4862.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2379, pruned_loss=0.04646, over 907678.16 frames. ], batch size: 31, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:02:43,719 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166710.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 12:03:00,948 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166735.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:03:00,972 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0344, 1.9712, 1.5604, 1.9375, 1.9476, 1.7334, 2.2457, 2.0309], + device='cuda:4'), covar=tensor([0.1364, 0.1909, 0.2881, 0.2250, 0.2669, 0.1728, 0.3066, 0.1711], + device='cuda:4'), in_proj_covar=tensor([0.0191, 0.0192, 0.0238, 0.0254, 0.0251, 0.0210, 0.0216, 0.0204], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:03:06,930 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1144, 2.1638, 1.8701, 2.1735, 2.0854, 2.1228, 2.1137, 2.7633], + device='cuda:4'), covar=tensor([0.3565, 0.4513, 0.3111, 0.4343, 0.4721, 0.2266, 0.3883, 0.1582], + device='cuda:4'), in_proj_covar=tensor([0.0292, 0.0267, 0.0241, 0.0277, 0.0263, 0.0234, 0.0262, 0.0241], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:03:12,742 INFO [finetune.py:976] (4/7) Epoch 30, batch 650, loss[loss=0.165, simple_loss=0.246, pruned_loss=0.04205, over 4828.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2408, pruned_loss=0.04716, over 918309.45 frames. ], batch size: 39, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:03:15,792 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166758.0, num_to_drop=1, layers_to_drop={1} +2023-03-27 12:03:19,477 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166764.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:03:26,540 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.144e+02 1.552e+02 1.821e+02 2.218e+02 3.611e+02, threshold=3.642e+02, percent-clipped=1.0 +2023-03-27 12:03:40,973 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5524, 1.8453, 1.5084, 1.5134, 2.0794, 2.1109, 1.8463, 1.7996], + device='cuda:4'), covar=tensor([0.0644, 0.0335, 0.0610, 0.0373, 0.0302, 0.0611, 0.0372, 0.0369], + device='cuda:4'), in_proj_covar=tensor([0.0103, 0.0106, 0.0148, 0.0111, 0.0102, 0.0118, 0.0104, 0.0115], + device='cuda:4'), out_proj_covar=tensor([7.9484e-05, 8.1227e-05, 1.1473e-04, 8.4079e-05, 7.9029e-05, 8.6654e-05, + 7.6676e-05, 8.7460e-05], device='cuda:4') +2023-03-27 12:03:41,542 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166796.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:03:45,813 INFO [finetune.py:976] (4/7) Epoch 30, batch 700, loss[loss=0.2394, simple_loss=0.2923, pruned_loss=0.09331, over 4816.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2422, pruned_loss=0.04732, over 926774.55 frames. ], batch size: 33, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:03:45,907 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166803.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:03:49,508 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-27 12:04:00,327 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166825.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:04:08,593 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166837.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:04:18,071 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=166851.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:04:18,720 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7511, 1.2088, 0.9080, 1.6238, 2.1349, 1.3783, 1.5226, 1.6299], + device='cuda:4'), covar=tensor([0.1444, 0.2132, 0.1797, 0.1200, 0.1887, 0.1848, 0.1454, 0.1929], + device='cuda:4'), in_proj_covar=tensor([0.0089, 0.0094, 0.0109, 0.0093, 0.0120, 0.0092, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 12:04:19,223 INFO [finetune.py:976] (4/7) Epoch 30, batch 750, loss[loss=0.1967, simple_loss=0.2624, pruned_loss=0.06545, over 4868.00 frames. ], tot_loss[loss=0.169, simple_loss=0.243, pruned_loss=0.04751, over 933651.44 frames. ], batch size: 34, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:04:27,158 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166865.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:04:33,236 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.605e+01 1.409e+02 1.752e+02 2.054e+02 3.389e+02, threshold=3.504e+02, percent-clipped=0.0 +2023-03-27 12:04:40,767 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.47 vs. limit=5.0 +2023-03-27 12:04:41,698 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=166887.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:04:46,248 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9014, 1.8011, 1.6389, 2.0670, 2.3175, 2.0411, 1.7368, 1.5652], + device='cuda:4'), covar=tensor([0.2124, 0.1916, 0.1931, 0.1540, 0.1465, 0.1092, 0.2193, 0.1927], + device='cuda:4'), in_proj_covar=tensor([0.0249, 0.0214, 0.0218, 0.0201, 0.0248, 0.0193, 0.0219, 0.0208], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:04:50,771 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-27 12:04:52,670 INFO [finetune.py:976] (4/7) Epoch 30, batch 800, loss[loss=0.1761, simple_loss=0.2551, pruned_loss=0.04858, over 4923.00 frames. ], tot_loss[loss=0.1691, simple_loss=0.2433, pruned_loss=0.04747, over 938799.93 frames. ], batch size: 42, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:05:09,485 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=166929.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:05:14,253 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4907, 1.3827, 1.3679, 1.3627, 0.8898, 2.4207, 0.8239, 1.2635], + device='cuda:4'), covar=tensor([0.4169, 0.3389, 0.2574, 0.3158, 0.2026, 0.0468, 0.2762, 0.1362], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0117, 0.0121, 0.0124, 0.0113, 0.0095, 0.0094, 0.0095], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 12:05:32,105 INFO [finetune.py:976] (4/7) Epoch 30, batch 850, loss[loss=0.178, simple_loss=0.2444, pruned_loss=0.05581, over 4827.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2402, pruned_loss=0.0462, over 943660.06 frames. ], batch size: 38, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:05:42,365 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=166963.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:05:44,836 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0230, 1.5903, 2.2552, 1.4908, 2.0144, 2.1087, 1.4003, 2.1422], + device='cuda:4'), covar=tensor([0.1302, 0.2396, 0.1383, 0.2105, 0.1043, 0.1652, 0.3177, 0.1211], + device='cuda:4'), in_proj_covar=tensor([0.0192, 0.0207, 0.0194, 0.0190, 0.0175, 0.0213, 0.0219, 0.0200], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:05:53,846 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.050e+02 1.507e+02 1.769e+02 2.105e+02 4.355e+02, threshold=3.539e+02, percent-clipped=2.0 +2023-03-27 12:06:07,535 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=166990.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:06:16,006 INFO [finetune.py:976] (4/7) Epoch 30, batch 900, loss[loss=0.1523, simple_loss=0.2216, pruned_loss=0.04154, over 4813.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.2373, pruned_loss=0.04559, over 945616.66 frames. ], batch size: 41, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:06:23,180 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167011.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:06:59,311 INFO [finetune.py:976] (4/7) Epoch 30, batch 950, loss[loss=0.1672, simple_loss=0.2346, pruned_loss=0.04993, over 4901.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.237, pruned_loss=0.04637, over 948551.65 frames. ], batch size: 32, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:07:13,661 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.122e+02 1.522e+02 1.848e+02 2.259e+02 3.601e+02, threshold=3.696e+02, percent-clipped=1.0 +2023-03-27 12:07:24,439 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167091.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:07:31,403 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167100.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:07:31,575 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.27 vs. limit=5.0 +2023-03-27 12:07:33,103 INFO [finetune.py:976] (4/7) Epoch 30, batch 1000, loss[loss=0.1855, simple_loss=0.258, pruned_loss=0.05654, over 4812.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2394, pruned_loss=0.04734, over 948010.77 frames. ], batch size: 38, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:07:33,853 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167104.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:07:44,463 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167120.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:07:51,842 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.1611, 1.2425, 1.4133, 1.2962, 1.3805, 2.4575, 1.1773, 1.3694], + device='cuda:4'), covar=tensor([0.0988, 0.1992, 0.1085, 0.1011, 0.1765, 0.0353, 0.1661, 0.1980], + device='cuda:4'), in_proj_covar=tensor([0.0075, 0.0083, 0.0073, 0.0076, 0.0092, 0.0081, 0.0086, 0.0081], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 12:07:55,373 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167137.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:08:06,884 INFO [finetune.py:976] (4/7) Epoch 30, batch 1050, loss[loss=0.2087, simple_loss=0.2741, pruned_loss=0.07171, over 4896.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.243, pruned_loss=0.04819, over 950299.50 frames. ], batch size: 43, lr: 2.81e-03, grad_scale: 64.0 +2023-03-27 12:08:11,880 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167161.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:08:14,778 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167165.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:08:14,824 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167165.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:08:21,246 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.094e+02 1.490e+02 1.818e+02 2.230e+02 3.401e+02, threshold=3.636e+02, percent-clipped=0.0 +2023-03-27 12:08:27,353 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167185.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:08:28,605 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167187.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:08:38,853 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.86 vs. limit=5.0 +2023-03-27 12:08:39,897 INFO [finetune.py:976] (4/7) Epoch 30, batch 1100, loss[loss=0.1851, simple_loss=0.2679, pruned_loss=0.05116, over 4867.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2433, pruned_loss=0.0479, over 952543.01 frames. ], batch size: 34, lr: 2.81e-03, grad_scale: 64.0 +2023-03-27 12:08:46,972 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167213.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:08:54,313 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5842, 0.6284, 1.6249, 1.5402, 1.4424, 1.4128, 1.4578, 1.6749], + device='cuda:4'), covar=tensor([0.4360, 0.4130, 0.3358, 0.3494, 0.4828, 0.4089, 0.4143, 0.3258], + device='cuda:4'), in_proj_covar=tensor([0.0272, 0.0252, 0.0271, 0.0302, 0.0302, 0.0280, 0.0307, 0.0256], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:09:01,420 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167235.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:09:13,765 INFO [finetune.py:976] (4/7) Epoch 30, batch 1150, loss[loss=0.1649, simple_loss=0.2496, pruned_loss=0.04011, over 4877.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2438, pruned_loss=0.04763, over 953876.07 frames. ], batch size: 32, lr: 2.81e-03, grad_scale: 64.0 +2023-03-27 12:09:28,393 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.072e+02 1.511e+02 1.772e+02 2.131e+02 4.281e+02, threshold=3.544e+02, percent-clipped=3.0 +2023-03-27 12:09:34,986 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167285.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:09:47,291 INFO [finetune.py:976] (4/7) Epoch 30, batch 1200, loss[loss=0.1751, simple_loss=0.252, pruned_loss=0.04913, over 4841.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2426, pruned_loss=0.04735, over 951873.12 frames. ], batch size: 44, lr: 2.81e-03, grad_scale: 64.0 +2023-03-27 12:09:56,420 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-27 12:10:20,453 INFO [finetune.py:976] (4/7) Epoch 30, batch 1250, loss[loss=0.1738, simple_loss=0.2315, pruned_loss=0.0581, over 4902.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2408, pruned_loss=0.04719, over 952731.95 frames. ], batch size: 43, lr: 2.81e-03, grad_scale: 64.0 +2023-03-27 12:10:36,986 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.068e+02 1.462e+02 1.711e+02 2.091e+02 4.240e+02, threshold=3.422e+02, percent-clipped=1.0 +2023-03-27 12:10:56,488 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167391.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:11:06,194 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7953, 1.3256, 0.9692, 1.6649, 2.2393, 1.5233, 1.6806, 1.8677], + device='cuda:4'), covar=tensor([0.1375, 0.1915, 0.1778, 0.1159, 0.1765, 0.1752, 0.1310, 0.1677], + device='cuda:4'), in_proj_covar=tensor([0.0090, 0.0094, 0.0109, 0.0093, 0.0120, 0.0092, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 12:11:09,177 INFO [finetune.py:976] (4/7) Epoch 30, batch 1300, loss[loss=0.1478, simple_loss=0.2169, pruned_loss=0.0394, over 4816.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.238, pruned_loss=0.04622, over 954230.37 frames. ], batch size: 25, lr: 2.81e-03, grad_scale: 64.0 +2023-03-27 12:11:24,564 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167420.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:11:32,403 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167432.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:11:38,998 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167439.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:11:47,734 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5636, 1.5731, 1.2716, 1.5481, 1.9049, 1.8410, 1.5621, 1.4045], + device='cuda:4'), covar=tensor([0.0361, 0.0355, 0.0722, 0.0333, 0.0233, 0.0483, 0.0367, 0.0442], + device='cuda:4'), in_proj_covar=tensor([0.0104, 0.0107, 0.0149, 0.0112, 0.0103, 0.0118, 0.0104, 0.0116], + device='cuda:4'), out_proj_covar=tensor([7.9990e-05, 8.1833e-05, 1.1603e-04, 8.4729e-05, 7.9501e-05, 8.7181e-05, + 7.7443e-05, 8.7955e-05], device='cuda:4') +2023-03-27 12:11:55,872 INFO [finetune.py:976] (4/7) Epoch 30, batch 1350, loss[loss=0.1709, simple_loss=0.2463, pruned_loss=0.04778, over 4927.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2388, pruned_loss=0.04668, over 953773.56 frames. ], batch size: 33, lr: 2.81e-03, grad_scale: 64.0 +2023-03-27 12:11:58,248 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167456.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:12:00,698 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167460.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:12:07,069 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167468.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:12:11,227 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.213e+01 1.429e+02 1.665e+02 1.960e+02 3.889e+02, threshold=3.329e+02, percent-clipped=2.0 +2023-03-27 12:12:23,210 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167493.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:12:29,688 INFO [finetune.py:976] (4/7) Epoch 30, batch 1400, loss[loss=0.1408, simple_loss=0.223, pruned_loss=0.02935, over 4828.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2409, pruned_loss=0.04746, over 953080.81 frames. ], batch size: 33, lr: 2.81e-03, grad_scale: 64.0 +2023-03-27 12:13:02,946 INFO [finetune.py:976] (4/7) Epoch 30, batch 1450, loss[loss=0.1647, simple_loss=0.2471, pruned_loss=0.04114, over 4750.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2431, pruned_loss=0.0479, over 954161.46 frames. ], batch size: 54, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:13:12,962 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6620, 1.4652, 1.0690, 0.2940, 1.2143, 1.4719, 1.4410, 1.4329], + device='cuda:4'), covar=tensor([0.1013, 0.0840, 0.1489, 0.2047, 0.1380, 0.2626, 0.2439, 0.0901], + device='cuda:4'), in_proj_covar=tensor([0.0172, 0.0190, 0.0203, 0.0182, 0.0210, 0.0212, 0.0225, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:13:19,228 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.032e+02 1.535e+02 1.844e+02 2.174e+02 4.375e+02, threshold=3.689e+02, percent-clipped=4.0 +2023-03-27 12:13:25,327 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167585.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:13:36,787 INFO [finetune.py:976] (4/7) Epoch 30, batch 1500, loss[loss=0.1645, simple_loss=0.2487, pruned_loss=0.04011, over 4862.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2437, pruned_loss=0.04797, over 954987.99 frames. ], batch size: 34, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:13:52,817 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-03-27 12:13:57,381 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167633.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:14:10,430 INFO [finetune.py:976] (4/7) Epoch 30, batch 1550, loss[loss=0.2295, simple_loss=0.2902, pruned_loss=0.08442, over 4705.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2432, pruned_loss=0.04777, over 954980.55 frames. ], batch size: 59, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:14:26,676 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.931e+01 1.416e+02 1.751e+02 2.105e+02 4.024e+02, threshold=3.503e+02, percent-clipped=1.0 +2023-03-27 12:14:44,010 INFO [finetune.py:976] (4/7) Epoch 30, batch 1600, loss[loss=0.1716, simple_loss=0.25, pruned_loss=0.04658, over 4767.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2422, pruned_loss=0.04783, over 954850.39 frames. ], batch size: 27, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:14:54,740 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167719.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:15:17,589 INFO [finetune.py:976] (4/7) Epoch 30, batch 1650, loss[loss=0.147, simple_loss=0.2191, pruned_loss=0.03749, over 4769.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2381, pruned_loss=0.04627, over 955373.36 frames. ], batch size: 26, lr: 2.81e-03, grad_scale: 32.0 +2023-03-27 12:15:19,517 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167756.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:15:21,893 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=167760.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:15:26,698 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=167767.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:15:32,529 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.722e+01 1.428e+02 1.633e+02 1.926e+02 4.440e+02, threshold=3.266e+02, percent-clipped=1.0 +2023-03-27 12:15:36,064 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167780.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:15:39,493 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4956, 2.4915, 2.4643, 1.9228, 2.4034, 2.7926, 2.6501, 2.1791], + device='cuda:4'), covar=tensor([0.0593, 0.0573, 0.0701, 0.0769, 0.0930, 0.0563, 0.0615, 0.1042], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0138, 0.0141, 0.0119, 0.0129, 0.0139, 0.0139, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:15:41,288 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=167788.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:15:55,586 INFO [finetune.py:976] (4/7) Epoch 30, batch 1700, loss[loss=0.1705, simple_loss=0.2445, pruned_loss=0.04827, over 4898.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.2362, pruned_loss=0.04562, over 956444.87 frames. ], batch size: 43, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:15:56,726 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167804.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:16:03,659 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=167808.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:16:25,157 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=167828.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:16:42,119 INFO [finetune.py:976] (4/7) Epoch 30, batch 1750, loss[loss=0.1481, simple_loss=0.2329, pruned_loss=0.03166, over 4824.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2403, pruned_loss=0.04764, over 957388.75 frames. ], batch size: 33, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:16:47,665 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.0815, 1.9560, 2.0000, 1.6319, 1.8855, 2.1209, 2.1290, 1.6457], + device='cuda:4'), covar=tensor([0.0524, 0.0547, 0.0605, 0.0690, 0.0956, 0.0517, 0.0460, 0.1033], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0137, 0.0141, 0.0118, 0.0128, 0.0138, 0.0139, 0.0162], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:16:54,733 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.4955, 1.3449, 1.6284, 2.3462, 1.6430, 2.1470, 1.0203, 2.0530], + device='cuda:4'), covar=tensor([0.1461, 0.1242, 0.1040, 0.0750, 0.0785, 0.1120, 0.1416, 0.0560], + device='cuda:4'), in_proj_covar=tensor([0.0100, 0.0116, 0.0133, 0.0164, 0.0100, 0.0136, 0.0125, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 12:16:57,034 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.117e+02 1.486e+02 1.829e+02 2.140e+02 4.770e+02, threshold=3.658e+02, percent-clipped=2.0 +2023-03-27 12:17:22,142 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.83 vs. limit=5.0 +2023-03-27 12:17:25,521 INFO [finetune.py:976] (4/7) Epoch 30, batch 1800, loss[loss=0.2136, simple_loss=0.2874, pruned_loss=0.06986, over 4154.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2434, pruned_loss=0.04879, over 955143.78 frames. ], batch size: 65, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:17:49,531 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-27 12:17:58,718 INFO [finetune.py:976] (4/7) Epoch 30, batch 1850, loss[loss=0.1629, simple_loss=0.2411, pruned_loss=0.04233, over 4812.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2452, pruned_loss=0.04951, over 954497.73 frames. ], batch size: 33, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:18:13,650 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.002e+02 1.458e+02 1.787e+02 2.144e+02 3.700e+02, threshold=3.573e+02, percent-clipped=1.0 +2023-03-27 12:18:23,837 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.88 vs. limit=5.0 +2023-03-27 12:18:31,855 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6428, 1.3874, 1.9030, 3.0075, 2.0061, 2.2278, 1.0384, 2.6307], + device='cuda:4'), covar=tensor([0.1755, 0.1485, 0.1354, 0.0617, 0.0855, 0.1399, 0.1788, 0.0493], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0116, 0.0134, 0.0165, 0.0100, 0.0136, 0.0125, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 12:18:33,563 INFO [finetune.py:976] (4/7) Epoch 30, batch 1900, loss[loss=0.1508, simple_loss=0.2317, pruned_loss=0.03499, over 4749.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2442, pruned_loss=0.04844, over 952795.85 frames. ], batch size: 54, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:18:55,004 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=168035.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:18:58,555 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4423, 2.2994, 2.3470, 1.6971, 2.3246, 2.4613, 2.5361, 1.9811], + device='cuda:4'), covar=tensor([0.0540, 0.0661, 0.0703, 0.0893, 0.0722, 0.0676, 0.0591, 0.1174], + device='cuda:4'), in_proj_covar=tensor([0.0132, 0.0138, 0.0142, 0.0119, 0.0129, 0.0139, 0.0140, 0.0163], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:18:59,745 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5315, 1.3730, 1.8257, 2.8644, 1.9133, 2.1938, 0.9868, 2.4540], + device='cuda:4'), covar=tensor([0.1762, 0.1526, 0.1401, 0.0842, 0.0921, 0.1694, 0.1843, 0.0594], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0116, 0.0134, 0.0165, 0.0101, 0.0136, 0.0126, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 12:19:01,151 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-27 12:19:07,333 INFO [finetune.py:976] (4/7) Epoch 30, batch 1950, loss[loss=0.1214, simple_loss=0.1939, pruned_loss=0.0245, over 4775.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2416, pruned_loss=0.04689, over 953315.17 frames. ], batch size: 28, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:19:21,551 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=168075.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:19:22,073 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.065e+02 1.458e+02 1.758e+02 2.118e+02 3.555e+02, threshold=3.516e+02, percent-clipped=0.0 +2023-03-27 12:19:25,868 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3885, 2.0943, 2.4634, 2.4009, 2.1477, 2.1405, 2.3691, 2.2568], + device='cuda:4'), covar=tensor([0.4144, 0.4060, 0.3289, 0.3913, 0.4908, 0.3838, 0.4636, 0.3142], + device='cuda:4'), in_proj_covar=tensor([0.0270, 0.0250, 0.0269, 0.0300, 0.0301, 0.0278, 0.0306, 0.0255], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:19:29,933 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.1465, 1.8650, 2.2068, 2.1440, 1.8992, 1.9267, 2.1712, 2.1194], + device='cuda:4'), covar=tensor([0.4029, 0.3823, 0.2902, 0.3929, 0.4877, 0.3775, 0.4255, 0.2599], + device='cuda:4'), in_proj_covar=tensor([0.0270, 0.0250, 0.0269, 0.0300, 0.0301, 0.0278, 0.0306, 0.0255], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:19:30,462 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168088.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:19:35,826 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=168096.0, num_to_drop=1, layers_to_drop={2} +2023-03-27 12:19:40,900 INFO [finetune.py:976] (4/7) Epoch 30, batch 2000, loss[loss=0.1486, simple_loss=0.2229, pruned_loss=0.03713, over 4894.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2393, pruned_loss=0.04653, over 953799.70 frames. ], batch size: 32, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:19:51,201 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7154, 1.7674, 1.5348, 1.9323, 1.8172, 1.8605, 1.4778, 1.4395], + device='cuda:4'), covar=tensor([0.2068, 0.1836, 0.1810, 0.1431, 0.1833, 0.1209, 0.2401, 0.1847], + device='cuda:4'), in_proj_covar=tensor([0.0251, 0.0215, 0.0220, 0.0202, 0.0250, 0.0195, 0.0222, 0.0209], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:19:54,029 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=168123.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:19:56,281 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-27 12:20:01,894 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168136.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:20:09,487 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9791, 2.1655, 1.7762, 1.8675, 2.5223, 2.6956, 2.1083, 2.0368], + device='cuda:4'), covar=tensor([0.0405, 0.0346, 0.0696, 0.0362, 0.0291, 0.0443, 0.0436, 0.0450], + device='cuda:4'), in_proj_covar=tensor([0.0104, 0.0107, 0.0149, 0.0112, 0.0103, 0.0118, 0.0105, 0.0116], + device='cuda:4'), out_proj_covar=tensor([7.9934e-05, 8.1725e-05, 1.1602e-04, 8.4931e-05, 7.9446e-05, 8.6816e-05, + 7.7724e-05, 8.7750e-05], device='cuda:4') +2023-03-27 12:20:14,572 INFO [finetune.py:976] (4/7) Epoch 30, batch 2050, loss[loss=0.1675, simple_loss=0.2511, pruned_loss=0.04191, over 4939.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2362, pruned_loss=0.04548, over 955328.51 frames. ], batch size: 33, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:20:29,509 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.016e+02 1.484e+02 1.729e+02 2.115e+02 4.273e+02, threshold=3.459e+02, percent-clipped=3.0 +2023-03-27 12:20:38,606 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.0483, 1.2604, 1.3888, 1.1339, 1.4248, 2.4617, 1.2167, 1.4533], + device='cuda:4'), covar=tensor([0.1130, 0.2035, 0.1087, 0.1034, 0.1833, 0.0395, 0.1626, 0.1873], + device='cuda:4'), in_proj_covar=tensor([0.0074, 0.0082, 0.0072, 0.0076, 0.0091, 0.0081, 0.0085, 0.0080], + device='cuda:4'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:4') +2023-03-27 12:20:47,507 INFO [finetune.py:976] (4/7) Epoch 30, batch 2100, loss[loss=0.1632, simple_loss=0.238, pruned_loss=0.04421, over 4814.00 frames. ], tot_loss[loss=0.1634, simple_loss=0.2358, pruned_loss=0.04552, over 955259.39 frames. ], batch size: 38, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:21:19,173 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=168233.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:21:41,933 INFO [finetune.py:976] (4/7) Epoch 30, batch 2150, loss[loss=0.1751, simple_loss=0.2457, pruned_loss=0.05229, over 4788.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2392, pruned_loss=0.04672, over 952833.69 frames. ], batch size: 29, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:21:53,968 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-27 12:22:00,997 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.149e+02 1.551e+02 1.892e+02 2.224e+02 4.404e+02, threshold=3.784e+02, percent-clipped=3.0 +2023-03-27 12:22:12,557 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=168294.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:22:18,788 INFO [finetune.py:976] (4/7) Epoch 30, batch 2200, loss[loss=0.1707, simple_loss=0.2528, pruned_loss=0.04436, over 4785.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2405, pruned_loss=0.04687, over 952040.06 frames. ], batch size: 29, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:22:50,781 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.7765, 1.2632, 1.8581, 1.8156, 1.6124, 1.5879, 1.8075, 1.8027], + device='cuda:4'), covar=tensor([0.3549, 0.3510, 0.2726, 0.3144, 0.4257, 0.3402, 0.3715, 0.2560], + device='cuda:4'), in_proj_covar=tensor([0.0270, 0.0250, 0.0269, 0.0300, 0.0301, 0.0278, 0.0306, 0.0255], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:22:51,943 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.4297, 2.2609, 1.8935, 0.9555, 2.0266, 1.8593, 1.7489, 2.1108], + device='cuda:4'), covar=tensor([0.0969, 0.0761, 0.1736, 0.2152, 0.1435, 0.2406, 0.2334, 0.1059], + device='cuda:4'), in_proj_covar=tensor([0.0170, 0.0188, 0.0201, 0.0181, 0.0208, 0.0209, 0.0223, 0.0195], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:23:02,562 INFO [finetune.py:976] (4/7) Epoch 30, batch 2250, loss[loss=0.1283, simple_loss=0.2158, pruned_loss=0.02041, over 4865.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2437, pruned_loss=0.04836, over 953012.53 frames. ], batch size: 31, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:23:17,417 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168375.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:23:17,907 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.029e+01 1.513e+02 1.826e+02 2.132e+02 3.584e+02, threshold=3.652e+02, percent-clipped=0.0 +2023-03-27 12:23:21,863 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.40 vs. limit=5.0 +2023-03-27 12:23:28,082 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=168391.0, num_to_drop=1, layers_to_drop={3} +2023-03-27 12:23:32,332 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6967, 1.5750, 2.2441, 3.5242, 2.3771, 2.5018, 1.0773, 3.0610], + device='cuda:4'), covar=tensor([0.1854, 0.1365, 0.1326, 0.0503, 0.0758, 0.1288, 0.1892, 0.0410], + device='cuda:4'), in_proj_covar=tensor([0.0101, 0.0116, 0.0133, 0.0164, 0.0100, 0.0136, 0.0125, 0.0102], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:4') +2023-03-27 12:23:36,284 INFO [finetune.py:976] (4/7) Epoch 30, batch 2300, loss[loss=0.1526, simple_loss=0.2214, pruned_loss=0.04194, over 4712.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2443, pruned_loss=0.04814, over 952534.32 frames. ], batch size: 23, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:23:49,962 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168423.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:23:49,985 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168423.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:23:57,184 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([4.5326, 3.9543, 4.1501, 4.3918, 4.2794, 3.9584, 4.6157, 1.4363], + device='cuda:4'), covar=tensor([0.0738, 0.0769, 0.0866, 0.0861, 0.1156, 0.1550, 0.0658, 0.5630], + device='cuda:4'), in_proj_covar=tensor([0.0355, 0.0248, 0.0289, 0.0299, 0.0339, 0.0288, 0.0308, 0.0304], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:24:09,555 INFO [finetune.py:976] (4/7) Epoch 30, batch 2350, loss[loss=0.1561, simple_loss=0.2227, pruned_loss=0.04469, over 4736.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2414, pruned_loss=0.04693, over 952827.35 frames. ], batch size: 59, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:24:21,868 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168471.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:24:24,774 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.258e+01 1.458e+02 1.699e+02 2.116e+02 4.301e+02, threshold=3.398e+02, percent-clipped=1.0 +2023-03-27 12:24:42,042 INFO [finetune.py:976] (4/7) Epoch 30, batch 2400, loss[loss=0.1473, simple_loss=0.2107, pruned_loss=0.04197, over 4839.00 frames. ], tot_loss[loss=0.165, simple_loss=0.238, pruned_loss=0.046, over 954300.59 frames. ], batch size: 40, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:24:45,657 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5437, 1.4225, 1.3115, 1.3141, 1.8170, 1.7743, 1.5091, 1.3115], + device='cuda:4'), covar=tensor([0.0357, 0.0367, 0.0721, 0.0368, 0.0240, 0.0432, 0.0423, 0.0521], + device='cuda:4'), in_proj_covar=tensor([0.0102, 0.0106, 0.0147, 0.0111, 0.0101, 0.0116, 0.0104, 0.0114], + device='cuda:4'), out_proj_covar=tensor([7.9130e-05, 8.0732e-05, 1.1452e-04, 8.4213e-05, 7.8375e-05, 8.5644e-05, + 7.6898e-05, 8.6716e-05], device='cuda:4') +2023-03-27 12:25:15,078 INFO [finetune.py:976] (4/7) Epoch 30, batch 2450, loss[loss=0.1591, simple_loss=0.2264, pruned_loss=0.04587, over 4894.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.236, pruned_loss=0.04572, over 954096.42 frames. ], batch size: 35, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:25:30,931 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.020e+02 1.487e+02 1.838e+02 2.245e+02 3.083e+02, threshold=3.676e+02, percent-clipped=0.0 +2023-03-27 12:25:39,369 INFO [zipformer.py:1188] (4/7) warmup_begin=1333.3, warmup_end=2000.0, batch_count=168589.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:25:48,869 INFO [finetune.py:976] (4/7) Epoch 30, batch 2500, loss[loss=0.1361, simple_loss=0.2121, pruned_loss=0.03008, over 4758.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2395, pruned_loss=0.04743, over 954362.91 frames. ], batch size: 28, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:26:05,757 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-27 12:26:27,846 INFO [finetune.py:976] (4/7) Epoch 30, batch 2550, loss[loss=0.186, simple_loss=0.2668, pruned_loss=0.05263, over 4903.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.243, pruned_loss=0.04839, over 955295.81 frames. ], batch size: 37, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:26:55,340 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.157e+02 1.538e+02 1.821e+02 2.163e+02 4.307e+02, threshold=3.641e+02, percent-clipped=2.0 +2023-03-27 12:27:09,672 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168691.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:27:17,863 INFO [finetune.py:976] (4/7) Epoch 30, batch 2600, loss[loss=0.1877, simple_loss=0.257, pruned_loss=0.05918, over 4863.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2446, pruned_loss=0.04887, over 955123.25 frames. ], batch size: 34, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:27:22,001 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-27 12:27:44,434 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168739.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:27:54,610 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([2.3606, 2.2313, 1.7614, 2.4013, 2.1697, 1.9866, 2.6167, 2.3620], + device='cuda:4'), covar=tensor([0.1323, 0.2136, 0.3047, 0.2407, 0.2568, 0.1775, 0.2738, 0.1651], + device='cuda:4'), in_proj_covar=tensor([0.0191, 0.0192, 0.0240, 0.0254, 0.0251, 0.0210, 0.0217, 0.0205], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:28:01,793 INFO [finetune.py:976] (4/7) Epoch 30, batch 2650, loss[loss=0.1801, simple_loss=0.2485, pruned_loss=0.05588, over 4774.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2451, pruned_loss=0.04887, over 955396.69 frames. ], batch size: 28, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:28:10,636 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-27 12:28:21,696 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 1.069e+02 1.510e+02 1.724e+02 1.979e+02 3.263e+02, threshold=3.448e+02, percent-clipped=0.0 +2023-03-27 12:28:43,722 INFO [finetune.py:976] (4/7) Epoch 30, batch 2700, loss[loss=0.1663, simple_loss=0.2283, pruned_loss=0.05211, over 4810.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.244, pruned_loss=0.04849, over 954052.49 frames. ], batch size: 38, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:29:09,177 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.6669, 1.5567, 1.5126, 1.5888, 1.1708, 3.0661, 1.3015, 1.5714], + device='cuda:4'), covar=tensor([0.3045, 0.2284, 0.1949, 0.2154, 0.1653, 0.0283, 0.2795, 0.1203], + device='cuda:4'), in_proj_covar=tensor([0.0131, 0.0116, 0.0120, 0.0123, 0.0113, 0.0094, 0.0093, 0.0094], + device='cuda:4'), out_proj_covar=tensor([0.0006, 0.0006, 0.0005, 0.0006, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:4') +2023-03-27 12:29:17,016 INFO [finetune.py:976] (4/7) Epoch 30, batch 2750, loss[loss=0.1796, simple_loss=0.2487, pruned_loss=0.05524, over 4759.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2415, pruned_loss=0.0482, over 954214.31 frames. ], batch size: 27, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:29:24,735 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([0.3378, 1.4485, 1.4023, 0.7204, 1.5519, 1.7029, 1.7253, 1.3618], + device='cuda:4'), covar=tensor([0.0922, 0.0566, 0.0530, 0.0550, 0.0515, 0.0572, 0.0319, 0.0701], + device='cuda:4'), in_proj_covar=tensor([0.0120, 0.0145, 0.0130, 0.0119, 0.0130, 0.0129, 0.0140, 0.0150], + device='cuda:4'), out_proj_covar=tensor([8.7248e-05, 1.0389e-04, 9.2107e-05, 8.3462e-05, 9.1257e-05, 9.1488e-05, + 9.9273e-05, 1.0690e-04], device='cuda:4') +2023-03-27 12:29:28,117 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=4.84 vs. limit=5.0 +2023-03-27 12:29:32,260 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.461e+01 1.436e+02 1.670e+02 1.989e+02 2.987e+02, threshold=3.340e+02, percent-clipped=0.0 +2023-03-27 12:29:41,551 INFO [zipformer.py:1188] (4/7) warmup_begin=2000.0, warmup_end=2666.7, batch_count=168889.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:29:43,592 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-27 12:29:49,731 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-27 12:29:50,504 INFO [finetune.py:976] (4/7) Epoch 30, batch 2800, loss[loss=0.1651, simple_loss=0.2366, pruned_loss=0.04682, over 4894.00 frames. ], tot_loss[loss=0.1662, simple_loss=0.2382, pruned_loss=0.0471, over 954611.46 frames. ], batch size: 35, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:30:13,016 INFO [zipformer.py:1188] (4/7) warmup_begin=666.7, warmup_end=1333.3, batch_count=168937.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:30:23,986 INFO [finetune.py:976] (4/7) Epoch 30, batch 2850, loss[loss=0.1562, simple_loss=0.2272, pruned_loss=0.04263, over 4144.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2372, pruned_loss=0.0469, over 954379.92 frames. ], batch size: 65, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:30:33,490 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=168967.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:30:34,713 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=168969.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 12:30:36,023 INFO [scaling.py:679] (4/7) Whitening: num_groups=1, num_channels=384, metric=3.63 vs. limit=5.0 +2023-03-27 12:30:38,409 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.3487, 1.2565, 1.1848, 1.1826, 1.5496, 1.4052, 1.2453, 1.1777], + device='cuda:4'), covar=tensor([0.0395, 0.0334, 0.0724, 0.0356, 0.0258, 0.0589, 0.0426, 0.0449], + device='cuda:4'), in_proj_covar=tensor([0.0104, 0.0106, 0.0149, 0.0111, 0.0102, 0.0118, 0.0105, 0.0115], + device='cuda:4'), out_proj_covar=tensor([8.0054e-05, 8.1077e-05, 1.1563e-04, 8.4684e-05, 7.9105e-05, 8.6455e-05, + 7.7896e-05, 8.7461e-05], device='cuda:4') +2023-03-27 12:30:38,882 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 8.606e+01 1.456e+02 1.728e+02 2.104e+02 5.266e+02, threshold=3.457e+02, percent-clipped=2.0 +2023-03-27 12:30:42,543 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.9794, 1.7405, 2.1629, 1.4774, 1.9865, 2.1922, 1.6210, 2.3567], + device='cuda:4'), covar=tensor([0.1347, 0.2061, 0.1427, 0.1884, 0.0967, 0.1390, 0.2665, 0.0818], + device='cuda:4'), in_proj_covar=tensor([0.0189, 0.0205, 0.0191, 0.0187, 0.0172, 0.0210, 0.0217, 0.0197], + device='cuda:4'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:4') +2023-03-27 12:30:57,848 INFO [finetune.py:976] (4/7) Epoch 30, batch 2900, loss[loss=0.1799, simple_loss=0.25, pruned_loss=0.05486, over 4840.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2405, pruned_loss=0.04794, over 955799.88 frames. ], batch size: 49, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:30:58,563 INFO [zipformer.py:2441] (4/7) attn_weights_entropy = tensor([1.5473, 1.1167, 0.7457, 1.3316, 1.9568, 0.8322, 1.2457, 1.3366], + device='cuda:4'), covar=tensor([0.1521, 0.2068, 0.1721, 0.1247, 0.1871, 0.1943, 0.1441, 0.2050], + device='cuda:4'), in_proj_covar=tensor([0.0091, 0.0094, 0.0110, 0.0093, 0.0121, 0.0093, 0.0098, 0.0089], + device='cuda:4'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:4') +2023-03-27 12:31:14,730 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169028.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:31:15,930 INFO [zipformer.py:1188] (4/7) warmup_begin=3333.3, warmup_end=4000.0, batch_count=169030.0, num_to_drop=1, layers_to_drop={0} +2023-03-27 12:31:31,773 INFO [finetune.py:976] (4/7) Epoch 30, batch 2950, loss[loss=0.1704, simple_loss=0.2519, pruned_loss=0.04451, over 4811.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2427, pruned_loss=0.04865, over 954478.11 frames. ], batch size: 39, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:31:34,934 INFO [scaling.py:679] (4/7) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-27 12:31:49,075 INFO [optim.py:369] (4/7) Clipping_scale=2.0, grad-norm quartiles 9.406e+01 1.615e+02 1.887e+02 2.255e+02 4.054e+02, threshold=3.773e+02, percent-clipped=1.0 +2023-03-27 12:31:53,289 INFO [zipformer.py:1188] (4/7) warmup_begin=2666.7, warmup_end=3333.3, batch_count=169082.0, num_to_drop=0, layers_to_drop=set() +2023-03-27 12:32:19,269 INFO [finetune.py:976] (4/7) Epoch 30, batch 3000, loss[loss=0.2433, simple_loss=0.2941, pruned_loss=0.09623, over 4123.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2433, pruned_loss=0.04814, over 955988.22 frames. ], batch size: 66, lr: 2.80e-03, grad_scale: 32.0 +2023-03-27 12:32:19,269 INFO [finetune.py:1001] (4/7) Computing validation loss