diff --git "a/exp/log/log-train-2023-03-08-13-46-28-0" "b/exp/log/log-train-2023-03-08-13-46-28-0" new file mode 100644--- /dev/null +++ "b/exp/log/log-train-2023-03-08-13-46-28-0" @@ -0,0 +1,15109 @@ +2023-03-08 13:46:28,357 INFO [train.py:970] (0/4) Training started +2023-03-08 13:46:28,361 INFO [train.py:980] (0/4) Device: cuda:0 +2023-03-08 13:46:28,367 INFO [train.py:989] (0/4) {'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': '1.22', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '96c9a2aece2a3a7633da07740e24fa3d96f5498c', 'k2-git-date': 'Thu Nov 10 08:14:02 2022', 'lhotse-version': '1.13.0.dev+git.527d964.clean', 'torch-version': '1.12.1', 'torch-cuda-available': True, 'torch-cuda-version': '11.6', 'python-version': '3.8', 'icefall-git-branch': 'random_padding', 'icefall-git-sha1': '4cf2472-dirty', 'icefall-git-date': 'Wed Mar 1 23:53:23 2023', 'icefall-path': '/ceph-data4/yangxiaoyu/softwares/icefall_development/icefall_random_padding', 'k2-path': '/ceph-data4/yangxiaoyu/softwares/anaconda3/envs/k2_latest/lib/python3.8/site-packages/k2/__init__.py', 'lhotse-path': '/ceph-data4/yangxiaoyu/softwares/lhotse_development/lhotse_random_padding_left/lhotse/__init__.py', 'hostname': 'de-74279-k2-train-9-0208143539-7dcb6bfd79-b6fdq', 'IP address': '10.177.13.150'}, 'world_size': 4, 'master_port': 18180, 'tensorboard': True, 'num_epochs': 30, 'start_epoch': 1, 'start_batch': 0, 'exp_dir': PosixPath('pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4'), 'bpe_model': 'data/lang_bpe_500/bpe.model', 'base_lr': 0.05, 'lr_batches': 5000, 'lr_epochs': 3.5, '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, 'full_libri': True, 'manifest_dir': PosixPath('data/fbank'), 'max_duration': 750, 'bucketing_sampler': True, 'num_buckets': 30, 'concatenate_cuts': False, 'random_left_padding': False, 'num_left_padding': 8, '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} +2023-03-08 13:46:28,367 INFO [train.py:991] (0/4) About to create model +2023-03-08 13:46:29,235 INFO [zipformer.py:178] (0/4) 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. +2023-03-08 13:46:29,259 INFO [train.py:995] (0/4) Number of model parameters: 70369391 +2023-03-08 13:46:32,477 INFO [train.py:1010] (0/4) Using DDP +2023-03-08 13:46:32,852 INFO [asr_datamodule.py:439] (0/4) About to get the shuffled train-clean-100, train-clean-360 and train-other-500 cuts +2023-03-08 13:46:32,864 INFO [asr_datamodule.py:244] (0/4) Enable MUSAN +2023-03-08 13:46:32,864 INFO [asr_datamodule.py:245] (0/4) About to get Musan cuts +2023-03-08 13:46:35,101 INFO [asr_datamodule.py:269] (0/4) Enable SpecAugment +2023-03-08 13:46:35,101 INFO [asr_datamodule.py:270] (0/4) Time warp factor: 80 +2023-03-08 13:46:35,101 INFO [asr_datamodule.py:280] (0/4) Num frame mask: 10 +2023-03-08 13:46:35,101 INFO [asr_datamodule.py:293] (0/4) About to create train dataset +2023-03-08 13:46:35,102 INFO [asr_datamodule.py:320] (0/4) Using DynamicBucketingSampler. +2023-03-08 13:46:42,057 INFO [asr_datamodule.py:335] (0/4) About to create train dataloader +2023-03-08 13:46:42,061 INFO [asr_datamodule.py:449] (0/4) About to get dev-clean cuts +2023-03-08 13:46:42,065 INFO [asr_datamodule.py:456] (0/4) About to get dev-other cuts +2023-03-08 13:46:42,066 INFO [asr_datamodule.py:366] (0/4) About to create dev dataset +2023-03-08 13:46:42,415 INFO [asr_datamodule.py:383] (0/4) About to create dev dataloader +2023-03-08 13:47:06,862 INFO [train.py:898] (0/4) Epoch 1, batch 0, loss[loss=7.449, simple_loss=6.745, pruned_loss=7.021, over 18395.00 frames. ], tot_loss[loss=7.449, simple_loss=6.745, pruned_loss=7.021, over 18395.00 frames. ], batch size: 43, lr: 2.50e-02, grad_scale: 2.0 +2023-03-08 13:47:06,865 INFO [train.py:923] (0/4) Computing validation loss +2023-03-08 13:47:18,765 INFO [train.py:932] (0/4) Epoch 1, validation: loss=6.911, simple_loss=6.237, pruned_loss=6.721, over 944034.00 frames. +2023-03-08 13:47:18,766 INFO [train.py:933] (0/4) Maximum memory allocated so far is 15385MB +2023-03-08 13:47:22,967 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5.0, num_to_drop=2, layers_to_drop={1, 3} +2023-03-08 13:47:42,154 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=23.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 13:47:48,454 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.4537, 4.4526, 4.4209, 4.4105, 4.4502, 4.4481, 4.4490, 4.4483], + device='cuda:0'), covar=tensor([0.0011, 0.0007, 0.0009, 0.0008, 0.0012, 0.0010, 0.0012, 0.0007], + device='cuda:0'), in_proj_covar=tensor([0.0009, 0.0009, 0.0009, 0.0009, 0.0009, 0.0009, 0.0009, 0.0009], + device='cuda:0'), out_proj_covar=tensor([8.9100e-06, 8.8415e-06, 9.1061e-06, 8.8321e-06, 9.0736e-06, 8.9275e-06, + 9.0014e-06, 8.9923e-06], device='cuda:0') +2023-03-08 13:48:01,467 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.9515, 5.9454, 5.9511, 5.9525, 5.9538, 5.9556, 5.8364, 5.9584], + device='cuda:0'), covar=tensor([0.0006, 0.0005, 0.0009, 0.0008, 0.0007, 0.0009, 0.0009, 0.0007], + device='cuda:0'), in_proj_covar=tensor([0.0009, 0.0009, 0.0009, 0.0009, 0.0009, 0.0009, 0.0009, 0.0009], + device='cuda:0'), out_proj_covar=tensor([8.9180e-06, 9.0272e-06, 8.8533e-06, 9.0669e-06, 8.8809e-06, 8.9910e-06, + 8.9796e-06, 9.0382e-06], device='cuda:0') +2023-03-08 13:48:05,134 INFO [train.py:898] (0/4) Epoch 1, batch 50, loss[loss=1.353, simple_loss=1.199, pruned_loss=1.373, over 16048.00 frames. ], tot_loss[loss=2.138, simple_loss=1.933, pruned_loss=1.961, over 814820.04 frames. ], batch size: 94, lr: 2.75e-02, grad_scale: 1.0 +2023-03-08 13:48:08,631 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=49.90 vs. limit=5.0 +2023-03-08 13:48:18,220 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=4.86 vs. limit=2.0 +2023-03-08 13:48:26,229 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=4.04 vs. limit=2.0 +2023-03-08 13:48:33,942 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 13:48:51,691 WARNING [train.py:888] (0/4) Grad scale is small: 0.0009765625 +2023-03-08 13:48:51,692 INFO [train.py:898] (0/4) Epoch 1, batch 100, loss[loss=1.039, simple_loss=0.888, pruned_loss=1.19, over 18417.00 frames. ], tot_loss[loss=1.613, simple_loss=1.435, pruned_loss=1.599, over 1430034.33 frames. ], batch size: 43, lr: 3.00e-02, grad_scale: 0.001953125 +2023-03-08 13:48:55,120 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=71.98 vs. limit=5.0 +2023-03-08 13:49:00,254 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.8888, 5.8889, 5.8884, 5.8892, 5.8885, 5.5788, 5.8852, 5.8888], + device='cuda:0'), covar=tensor([0.0152, 0.0060, 0.0146, 0.0112, 0.0202, 0.0527, 0.0101, 0.0061], + device='cuda:0'), in_proj_covar=tensor([0.0009, 0.0009, 0.0009, 0.0009, 0.0009, 0.0009, 0.0009, 0.0009], + device='cuda:0'), out_proj_covar=tensor([9.2525e-06, 9.0149e-06, 8.9938e-06, 8.8840e-06, 9.0142e-06, 8.9836e-06, + 8.8558e-06, 9.0327e-06], device='cuda:0') +2023-03-08 13:49:00,732 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 6.526e+01 1.420e+02 2.842e+02 1.227e+03 3.323e+06, threshold=5.685e+02, percent-clipped=0.0 +2023-03-08 13:49:18,538 WARNING [optim.py:389] (0/4) Scaling gradients by 0.03670352324843407, model_norm_threshold=568.4981689453125 +2023-03-08 13:49:18,718 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.51, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=1.227e+08, grad_sumsq = 3.226e+09, orig_rms_sq=3.802e-02 +2023-03-08 13:49:27,501 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5337, 4.5135, 4.4280, 4.4938, 4.5349, 4.5325, 4.4959, 4.4817], + device='cuda:0'), covar=tensor([0.0087, 0.0043, 0.0045, 0.0041, 0.0055, 0.0050, 0.0073, 0.0026], + device='cuda:0'), in_proj_covar=tensor([0.0009, 0.0009, 0.0009, 0.0009, 0.0009, 0.0009, 0.0009, 0.0009], + device='cuda:0'), out_proj_covar=tensor([8.9161e-06, 8.8938e-06, 9.1402e-06, 8.9403e-06, 9.1647e-06, 8.8896e-06, + 9.0440e-06, 9.0103e-06], device='cuda:0') +2023-03-08 13:49:29,040 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=144.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-08 13:49:34,258 INFO [train.py:898] (0/4) Epoch 1, batch 150, loss[loss=0.991, simple_loss=0.8331, pruned_loss=1.127, over 18495.00 frames. ], tot_loss[loss=1.397, simple_loss=1.222, pruned_loss=1.461, over 1909470.77 frames. ], batch size: 51, lr: 3.25e-02, grad_scale: 0.001953125 +2023-03-08 13:49:55,941 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=3.71 vs. limit=2.0 +2023-03-08 13:49:57,988 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=3.07 vs. limit=2.0 +2023-03-08 13:50:02,809 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=3.81 vs. limit=2.0 +2023-03-08 13:50:16,356 WARNING [train.py:888] (0/4) Grad scale is small: 0.001953125 +2023-03-08 13:50:16,357 INFO [train.py:898] (0/4) Epoch 1, batch 200, loss[loss=1.016, simple_loss=0.8508, pruned_loss=1.092, over 18235.00 frames. ], tot_loss[loss=1.254, simple_loss=1.085, pruned_loss=1.334, over 2291854.29 frames. ], batch size: 60, lr: 3.50e-02, grad_scale: 0.00390625 +2023-03-08 13:50:32,159 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 7.044e+01 1.307e+02 2.204e+02 4.914e+02 1.549e+04, threshold=4.408e+02, percent-clipped=23.0 +2023-03-08 13:51:05,113 INFO [train.py:898] (0/4) Epoch 1, batch 250, loss[loss=0.8436, simple_loss=0.7042, pruned_loss=0.8623, over 17681.00 frames. ], tot_loss[loss=1.169, simple_loss=1.002, pruned_loss=1.238, over 2594332.85 frames. ], batch size: 39, lr: 3.75e-02, grad_scale: 0.00390625 +2023-03-08 13:51:21,648 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=31.50 vs. limit=5.0 +2023-03-08 13:51:37,154 WARNING [optim.py:389] (0/4) Scaling gradients by 0.0006386953755281866, model_norm_threshold=440.7669677734375 +2023-03-08 13:51:37,331 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.skip_modules.4.weight1 with proportion 0.43, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=2.042e+11, grad_sumsq = 2.042e+11, orig_rms_sq=1.000e+00 +2023-03-08 13:51:43,265 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=296.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 13:51:46,140 WARNING [optim.py:389] (0/4) Scaling gradients by 0.04052559658885002, model_norm_threshold=440.7669677734375 +2023-03-08 13:51:46,314 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.77, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=9.126e+07, grad_sumsq = 1.809e+09, orig_rms_sq=5.045e-02 +2023-03-08 13:51:46,530 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=300.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-08 13:51:47,136 WARNING [train.py:888] (0/4) Grad scale is small: 0.00390625 +2023-03-08 13:51:47,136 INFO [train.py:898] (0/4) Epoch 1, batch 300, loss[loss=0.9821, simple_loss=0.8134, pruned_loss=0.9766, over 18568.00 frames. ], tot_loss[loss=1.108, simple_loss=0.9432, pruned_loss=1.159, over 2815123.64 frames. ], batch size: 54, lr: 4.00e-02, grad_scale: 0.0078125 +2023-03-08 13:51:55,810 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 8.726e+01 1.552e+02 2.190e+02 3.865e+02 6.901e+05, threshold=4.380e+02, percent-clipped=20.0 +2023-03-08 13:52:01,031 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=3.65 vs. limit=2.0 +2023-03-08 13:52:20,493 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=3.11 vs. limit=2.0 +2023-03-08 13:52:21,625 WARNING [optim.py:389] (0/4) Scaling gradients by 0.00015154901484493166, model_norm_threshold=438.01873779296875 +2023-03-08 13:52:21,807 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.70, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=5.875e+12, grad_sumsq = 1.517e+14, orig_rms_sq=3.874e-02 +2023-03-08 13:52:23,420 WARNING [optim.py:389] (0/4) Scaling gradients by 0.01597026363015175, model_norm_threshold=438.01873779296875 +2023-03-08 13:52:23,581 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.80, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=6.028e+08, grad_sumsq = 1.556e+10, orig_rms_sq=3.874e-02 +2023-03-08 13:52:24,510 WARNING [optim.py:389] (0/4) Scaling gradients by 0.022203104570508003, model_norm_threshold=438.01873779296875 +2023-03-08 13:52:24,679 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.86, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=3.351e+08, grad_sumsq = 8.650e+09, orig_rms_sq=3.874e-02 +2023-03-08 13:52:26,179 WARNING [optim.py:389] (0/4) Scaling gradients by 0.008352968841791153, model_norm_threshold=438.01873779296875 +2023-03-08 13:52:26,350 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.77, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=2.124e+09, grad_sumsq = 5.347e+10, orig_rms_sq=3.973e-02 +2023-03-08 13:52:27,664 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=11.68 vs. limit=2.0 +2023-03-08 13:52:30,348 INFO [train.py:898] (0/4) Epoch 1, batch 350, loss[loss=0.8523, simple_loss=0.6939, pruned_loss=0.8503, over 18264.00 frames. ], tot_loss[loss=1.066, simple_loss=0.9001, pruned_loss=1.101, over 2985011.70 frames. ], batch size: 45, lr: 4.25e-02, grad_scale: 0.00390625 +2023-03-08 13:52:31,223 WARNING [optim.py:389] (0/4) Scaling gradients by 0.00011210949014639482, model_norm_threshold=438.01873779296875 +2023-03-08 13:52:31,391 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.85, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=1.296e+13, grad_sumsq = 3.160e+14, orig_rms_sq=4.100e-02 +2023-03-08 13:52:34,662 WARNING [optim.py:389] (0/4) Scaling gradients by 0.06386774033308029, model_norm_threshold=438.01873779296875 +2023-03-08 13:52:34,836 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.58, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=2.733e+07, grad_sumsq = 6.508e+08, orig_rms_sq=4.200e-02 +2023-03-08 13:52:35,612 WARNING [optim.py:389] (0/4) Scaling gradients by 0.0002155240799766034, model_norm_threshold=438.01873779296875 +2023-03-08 13:52:35,770 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.85, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=3.507e+12, grad_sumsq = 8.280e+13, orig_rms_sq=4.236e-02 +2023-03-08 13:52:36,060 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=357.0, num_to_drop=2, layers_to_drop={0, 3} +2023-03-08 13:52:43,645 WARNING [optim.py:389] (0/4) Scaling gradients by 0.08033499121665955, model_norm_threshold=438.01873779296875 +2023-03-08 13:52:43,801 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.60, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=1.796e+07, grad_sumsq = 4.255e+08, orig_rms_sq=4.221e-02 +2023-03-08 13:52:58,804 WARNING [optim.py:389] (0/4) Scaling gradients by 0.00024505704641342163, model_norm_threshold=438.01873779296875 +2023-03-08 13:52:58,966 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.67, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=2.127e+12, grad_sumsq = 5.153e+13, orig_rms_sq=4.128e-02 +2023-03-08 13:53:00,764 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=387.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 13:53:01,253 WARNING [optim.py:389] (0/4) Scaling gradients by 0.0035240945871919394, model_norm_threshold=438.01873779296875 +2023-03-08 13:53:01,415 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.51, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=7.954e+09, grad_sumsq = 1.927e+11, orig_rms_sq=4.128e-02 +2023-03-08 13:53:02,949 WARNING [optim.py:389] (0/4) Scaling gradients by 0.00012842776777688414, model_norm_threshold=438.01873779296875 +2023-03-08 13:53:03,116 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.65, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=7.588e+12, grad_sumsq = 1.863e+14, orig_rms_sq=4.072e-02 +2023-03-08 13:53:08,947 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=6.28 vs. limit=2.0 +2023-03-08 13:53:10,038 WARNING [optim.py:389] (0/4) Scaling gradients by 0.007196913007646799, model_norm_threshold=438.01873779296875 +2023-03-08 13:53:10,204 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoders.4.encoder.layers.0.norm_final.eps with proportion 0.38, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=1.420e+09, grad_sumsq = 1.420e+09, orig_rms_sq=1.000e+00 +2023-03-08 13:53:13,987 WARNING [train.py:888] (0/4) Grad scale is small: 0.00390625 +2023-03-08 13:53:13,987 INFO [train.py:898] (0/4) Epoch 1, batch 400, loss[loss=0.9905, simple_loss=0.8023, pruned_loss=0.958, over 18199.00 frames. ], tot_loss[loss=1.036, simple_loss=0.8669, pruned_loss=1.056, over 3124216.67 frames. ], batch size: 60, lr: 4.50e-02, grad_scale: 0.0078125 +2023-03-08 13:53:23,502 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.343e+02 2.188e+02 3.037e+02 6.402e+02 3.907e+06, threshold=6.074e+02, percent-clipped=33.0 +2023-03-08 13:53:35,681 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=47.83 vs. limit=5.0 +2023-03-08 13:53:43,973 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=439.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-08 13:53:53,232 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=448.0, num_to_drop=2, layers_to_drop={0, 3} +2023-03-08 13:53:55,309 INFO [train.py:898] (0/4) Epoch 1, batch 450, loss[loss=1.001, simple_loss=0.8012, pruned_loss=0.9572, over 18461.00 frames. ], tot_loss[loss=1.013, simple_loss=0.8401, pruned_loss=1.02, over 3230920.84 frames. ], batch size: 59, lr: 4.75e-02, grad_scale: 0.0078125 +2023-03-08 13:53:55,801 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=3.58 vs. limit=2.0 +2023-03-08 13:53:59,213 WARNING [optim.py:389] (0/4) Scaling gradients by 0.001993334386497736, model_norm_threshold=607.3988037109375 +2023-03-08 13:53:59,379 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoders.4.encoder.layers.1.norm_final.eps with proportion 0.46, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=4.249e+10, grad_sumsq = 4.249e+10, orig_rms_sq=1.000e+00 +2023-03-08 13:54:00,197 WARNING [optim.py:389] (0/4) Scaling gradients by 0.009787621907889843, model_norm_threshold=607.3988037109375 +2023-03-08 13:54:00,385 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoders.4.encoder.layers.0.norm_final.eps with proportion 0.37, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=1.407e+09, grad_sumsq = 1.407e+09, orig_rms_sq=1.000e+00 +2023-03-08 13:54:05,676 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=14.54 vs. limit=2.0 +2023-03-08 13:54:09,174 WARNING [optim.py:389] (0/4) Scaling gradients by 0.07029950618743896, model_norm_threshold=607.3988037109375 +2023-03-08 13:54:09,345 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.83, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=6.195e+07, grad_sumsq = 1.377e+09, orig_rms_sq=4.498e-02 +2023-03-08 13:54:20,902 WARNING [optim.py:389] (0/4) Scaling gradients by 0.008813662454485893, model_norm_threshold=607.3988037109375 +2023-03-08 13:54:21,088 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.85, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=4.028e+09, grad_sumsq = 8.824e+10, orig_rms_sq=4.564e-02 +2023-03-08 13:54:27,982 WARNING [optim.py:389] (0/4) Scaling gradients by 0.024284733459353447, model_norm_threshold=607.3988037109375 +2023-03-08 13:54:28,142 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.83, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=5.221e+08, grad_sumsq = 1.167e+10, orig_rms_sq=4.473e-02 +2023-03-08 13:54:38,281 WARNING [optim.py:389] (0/4) Scaling gradients by 0.0006707996362820268, model_norm_threshold=607.3988037109375 +2023-03-08 13:54:38,448 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoders.4.encoder.layers.1.norm_final.eps with proportion 0.69, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=5.647e+11, grad_sumsq = 5.647e+11, orig_rms_sq=1.000e+00 +2023-03-08 13:54:38,477 WARNING [train.py:888] (0/4) Grad scale is small: 0.0078125 +2023-03-08 13:54:38,478 INFO [train.py:898] (0/4) Epoch 1, batch 500, loss[loss=1.05, simple_loss=0.8238, pruned_loss=1.012, over 18311.00 frames. ], tot_loss[loss=1.003, simple_loss=0.8225, pruned_loss=1.001, over 3312202.83 frames. ], batch size: 54, lr: 4.99e-02, grad_scale: 0.015625 +2023-03-08 13:54:41,508 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=14.54 vs. limit=2.0 +2023-03-08 13:54:42,456 WARNING [optim.py:389] (0/4) Scaling gradients by 0.006503281649202108, model_norm_threshold=607.3988037109375 +2023-03-08 13:54:42,623 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoders.1.out_combiner.weight1 with proportion 0.48, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=4.173e+09, grad_sumsq = 4.173e+09, orig_rms_sq=1.000e+00 +2023-03-08 13:54:48,219 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.357e+02 2.649e+02 4.541e+02 8.074e+02 9.055e+05, threshold=9.081e+02, percent-clipped=35.0 +2023-03-08 13:54:48,219 WARNING [optim.py:389] (0/4) Scaling gradients by 0.07500762492418289, model_norm_threshold=908.1141357421875 +2023-03-08 13:54:48,387 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoders.1.out_combiner.weight1 with proportion 0.80, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=1.178e+08, grad_sumsq = 1.178e+08, orig_rms_sq=1.000e+00 +2023-03-08 13:54:53,814 WARNING [optim.py:389] (0/4) Scaling gradients by 0.00848373118788004, model_norm_threshold=908.1141357421875 +2023-03-08 13:54:53,980 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoders.2.out_combiner.weight1 with proportion 0.43, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=4.921e+09, grad_sumsq = 4.921e+09, orig_rms_sq=1.000e+00 +2023-03-08 13:54:54,844 WARNING [optim.py:389] (0/4) Scaling gradients by 0.0037236642092466354, model_norm_threshold=908.1141357421875 +2023-03-08 13:54:55,010 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.skip_modules.4.weight1 with proportion 0.69, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=4.127e+10, grad_sumsq = 4.127e+10, orig_rms_sq=1.000e+00 +2023-03-08 13:55:00,558 WARNING [optim.py:389] (0/4) Scaling gradients by 0.0036443807184696198, model_norm_threshold=908.1141357421875 +2023-03-08 13:55:00,722 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.88, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=5.460e+10, grad_sumsq = 1.061e+12, orig_rms_sq=5.145e-02 +2023-03-08 13:55:01,568 WARNING [optim.py:389] (0/4) Scaling gradients by 0.004900030791759491, model_norm_threshold=908.1141357421875 +2023-03-08 13:55:01,733 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.35, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=1.186e+10, grad_sumsq = 2.280e+11, orig_rms_sq=5.204e-02 +2023-03-08 13:55:02,552 WARNING [optim.py:389] (0/4) Scaling gradients by 0.07598941773176193, model_norm_threshold=908.1141357421875 +2023-03-08 13:55:02,721 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.93, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=1.329e+08, grad_sumsq = 2.553e+09, orig_rms_sq=5.204e-02 +2023-03-08 13:55:05,077 WARNING [optim.py:389] (0/4) Scaling gradients by 0.028503550216555595, model_norm_threshold=908.1141357421875 +2023-03-08 13:55:05,262 INFO [optim.py:451] (0/4) Parameter Dominanting tot_sumsq module.encoder.encoder_embed.conv.0.weight with proportion 0.78, where dominant_sumsq=(grad_sumsq*orig_rms_sq)=7.871e+08, grad_sumsq = 1.620e+10, orig_rms_sq=4.859e-02 +2023-03-08 13:55:21,069 INFO [train.py:898] (0/4) Epoch 1, batch 550, loss[loss=1.037, simple_loss=0.8105, pruned_loss=0.9707, over 18483.00 frames. ], tot_loss[loss=0.9989, simple_loss=0.8114, pruned_loss=0.9834, over 3364128.92 frames. ], batch size: 59, lr: 4.98e-02, grad_scale: 0.015625 +2023-03-08 13:55:29,886 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=562.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 13:55:34,762 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=568.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 13:55:51,754 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=590.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 13:55:52,006 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=81.76 vs. limit=5.0 +2023-03-08 13:56:00,912 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=600.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-08 13:56:01,487 INFO [train.py:898] (0/4) Epoch 1, batch 600, loss[loss=1.036, simple_loss=0.7994, pruned_loss=0.9626, over 18503.00 frames. ], tot_loss[loss=0.9974, simple_loss=0.8014, pruned_loss=0.971, over 3413745.92 frames. ], batch size: 51, lr: 4.98e-02, grad_scale: 0.03125 +2023-03-08 13:56:11,698 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.187e+02 4.468e+02 7.848e+02 1.243e+03 2.492e+05, threshold=1.570e+03, percent-clipped=35.0 +2023-03-08 13:56:18,062 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([6.0629, 6.0671, 6.0614, 6.0654, 6.0272, 6.0698, 6.0155, 6.0745], + device='cuda:0'), covar=tensor([0.0070, 0.0119, 0.0144, 0.0084, 0.0199, 0.0117, 0.0237, 0.0085], + device='cuda:0'), in_proj_covar=tensor([0.0011, 0.0012, 0.0011, 0.0011, 0.0012, 0.0011, 0.0011, 0.0012], + device='cuda:0'), out_proj_covar=tensor([1.1527e-05, 1.2271e-05, 1.1646e-05, 1.1613e-05, 1.1416e-05, 1.1521e-05, + 1.1517e-05, 1.1718e-05], device='cuda:0') +2023-03-08 13:56:19,740 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=623.0, num_to_drop=2, layers_to_drop={0, 3} +2023-03-08 13:56:24,214 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=629.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-08 13:56:31,404 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=8.05 vs. limit=2.0 +2023-03-08 13:56:31,487 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=6.39 vs. limit=2.0 +2023-03-08 13:56:35,408 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=49.26 vs. limit=5.0 +2023-03-08 13:56:39,061 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=648.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 13:56:41,331 INFO [train.py:898] (0/4) Epoch 1, batch 650, loss[loss=1.001, simple_loss=0.7701, pruned_loss=0.9066, over 12887.00 frames. ], tot_loss[loss=0.9959, simple_loss=0.7918, pruned_loss=0.9585, over 3451620.01 frames. ], batch size: 129, lr: 4.98e-02, grad_scale: 0.03125 +2023-03-08 13:56:41,583 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=651.0, num_to_drop=2, layers_to_drop={1, 3} +2023-03-08 13:56:42,896 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=652.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-08 13:57:22,328 INFO [train.py:898] (0/4) Epoch 1, batch 700, loss[loss=0.9286, simple_loss=0.7018, pruned_loss=0.842, over 18583.00 frames. ], tot_loss[loss=0.9911, simple_loss=0.7801, pruned_loss=0.9423, over 3486994.07 frames. ], batch size: 45, lr: 4.98e-02, grad_scale: 0.0625 +2023-03-08 13:57:25,998 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=57.70 vs. limit=5.0 +2023-03-08 13:57:33,534 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.059e+02 3.958e+02 5.670e+02 9.058e+02 3.205e+03, threshold=1.134e+03, percent-clipped=9.0 +2023-03-08 13:57:54,347 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=739.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-08 13:57:57,087 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=15.23 vs. limit=2.0 +2023-03-08 13:57:57,484 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=743.0, num_to_drop=2, layers_to_drop={2, 3} +2023-03-08 13:58:03,595 INFO [train.py:898] (0/4) Epoch 1, batch 750, loss[loss=0.9221, simple_loss=0.6884, pruned_loss=0.8287, over 18375.00 frames. ], tot_loss[loss=0.9935, simple_loss=0.7749, pruned_loss=0.9315, over 3510623.34 frames. ], batch size: 42, lr: 4.97e-02, grad_scale: 0.0625 +2023-03-08 13:58:04,697 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8746, 3.8806, 3.7509, 3.8726, 3.8079, 3.8318, 3.9279, 3.8520], + device='cuda:0'), covar=tensor([0.0491, 0.0211, 0.0380, 0.0440, 0.0265, 0.0489, 0.0212, 0.0305], + device='cuda:0'), in_proj_covar=tensor([0.0012, 0.0012, 0.0012, 0.0012, 0.0012, 0.0012, 0.0012, 0.0012], + device='cuda:0'), out_proj_covar=tensor([1.2019e-05, 1.1792e-05, 1.2163e-05, 1.1678e-05, 1.2181e-05, 1.2326e-05, + 1.1657e-05, 1.1513e-05], device='cuda:0') +2023-03-08 13:58:34,246 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=787.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 13:58:34,600 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=4.60 vs. limit=2.0 +2023-03-08 13:58:45,696 INFO [train.py:898] (0/4) Epoch 1, batch 800, loss[loss=0.9472, simple_loss=0.7058, pruned_loss=0.8313, over 18245.00 frames. ], tot_loss[loss=0.9952, simple_loss=0.7688, pruned_loss=0.9209, over 3539422.99 frames. ], batch size: 45, lr: 4.97e-02, grad_scale: 0.125 +2023-03-08 13:58:55,624 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.968e+02 3.286e+02 5.870e+02 9.737e+02 2.138e+03, threshold=1.174e+03, percent-clipped=19.0 +2023-03-08 13:59:23,875 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=847.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 13:59:24,896 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=31.68 vs. limit=5.0 +2023-03-08 13:59:26,826 INFO [train.py:898] (0/4) Epoch 1, batch 850, loss[loss=1.051, simple_loss=0.782, pruned_loss=0.9023, over 17862.00 frames. ], tot_loss[loss=0.9993, simple_loss=0.7653, pruned_loss=0.912, over 3544562.68 frames. ], batch size: 70, lr: 4.96e-02, grad_scale: 0.125 +2023-03-08 14:00:09,695 INFO [train.py:898] (0/4) Epoch 1, batch 900, loss[loss=1.068, simple_loss=0.785, pruned_loss=0.9092, over 18488.00 frames. ], tot_loss[loss=1.003, simple_loss=0.762, pruned_loss=0.9017, over 3552656.27 frames. ], batch size: 51, lr: 4.96e-02, grad_scale: 0.25 +2023-03-08 14:00:15,515 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=908.0, num_to_drop=2, layers_to_drop={0, 3} +2023-03-08 14:00:19,232 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.803e+02 2.708e+02 4.690e+02 7.118e+02 2.600e+03, threshold=9.379e+02, percent-clipped=5.0 +2023-03-08 14:00:20,068 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=12.29 vs. limit=2.0 +2023-03-08 14:00:23,466 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=918.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-08 14:00:28,870 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=924.0, num_to_drop=2, layers_to_drop={0, 3} +2023-03-08 14:00:42,065 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=24.78 vs. limit=5.0 +2023-03-08 14:00:47,890 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=946.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-08 14:00:51,732 INFO [train.py:898] (0/4) Epoch 1, batch 950, loss[loss=1.071, simple_loss=0.7884, pruned_loss=0.8913, over 18352.00 frames. ], tot_loss[loss=1.008, simple_loss=0.7609, pruned_loss=0.8931, over 3566515.04 frames. ], batch size: 56, lr: 4.96e-02, grad_scale: 0.25 +2023-03-08 14:00:52,786 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=952.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-08 14:01:33,198 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=1000.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:01:33,770 INFO [train.py:898] (0/4) Epoch 1, batch 1000, loss[loss=1.063, simple_loss=0.7862, pruned_loss=0.8603, over 17098.00 frames. ], tot_loss[loss=1.008, simple_loss=0.7556, pruned_loss=0.8785, over 3584604.40 frames. ], batch size: 78, lr: 4.95e-02, grad_scale: 0.5 +2023-03-08 14:01:35,003 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=2.97 vs. limit=2.0 +2023-03-08 14:01:43,698 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.541e+02 3.491e+02 4.812e+02 7.825e+02 1.437e+03, threshold=9.623e+02, percent-clipped=15.0 +2023-03-08 14:02:10,222 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1043.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-08 14:02:17,220 INFO [train.py:898] (0/4) Epoch 1, batch 1050, loss[loss=1.065, simple_loss=0.7975, pruned_loss=0.8322, over 17890.00 frames. ], tot_loss[loss=1.009, simple_loss=0.755, pruned_loss=0.8636, over 3593819.69 frames. ], batch size: 70, lr: 4.95e-02, grad_scale: 0.5 +2023-03-08 14:02:26,108 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=2.84 vs. limit=2.0 +2023-03-08 14:02:32,653 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=2.95 vs. limit=2.0 +2023-03-08 14:02:52,430 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=1091.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:03:01,210 INFO [train.py:898] (0/4) Epoch 1, batch 1100, loss[loss=0.899, simple_loss=0.6906, pruned_loss=0.6693, over 18347.00 frames. ], tot_loss[loss=1.003, simple_loss=0.7511, pruned_loss=0.8392, over 3583325.37 frames. ], batch size: 46, lr: 4.94e-02, grad_scale: 1.0 +2023-03-08 14:03:05,171 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=4.41 vs. limit=2.0 +2023-03-08 14:03:10,797 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.010e+02 4.676e+02 7.290e+02 9.507e+02 1.781e+03, threshold=1.458e+03, percent-clipped=22.0 +2023-03-08 14:03:22,653 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4751, 5.5906, 5.4123, 5.1657, 5.7678, 5.6736, 5.3458, 5.5545], + device='cuda:0'), covar=tensor([0.1990, 0.1814, 0.2589, 0.3884, 0.0719, 0.1389, 0.4501, 0.2426], + device='cuda:0'), in_proj_covar=tensor([0.0060, 0.0061, 0.0061, 0.0061, 0.0059, 0.0066, 0.0070, 0.0066], + device='cuda:0'), out_proj_covar=tensor([6.2466e-05, 6.4484e-05, 6.0040e-05, 6.9100e-05, 6.1949e-05, 6.2909e-05, + 6.6386e-05, 6.5915e-05], device='cuda:0') +2023-03-08 14:03:44,978 INFO [train.py:898] (0/4) Epoch 1, batch 1150, loss[loss=0.9178, simple_loss=0.7207, pruned_loss=0.6546, over 18488.00 frames. ], tot_loss[loss=0.9843, simple_loss=0.7425, pruned_loss=0.8025, over 3571299.47 frames. ], batch size: 53, lr: 4.94e-02, grad_scale: 1.0 +2023-03-08 14:03:46,877 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1153.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:04:28,512 INFO [train.py:898] (0/4) Epoch 1, batch 1200, loss[loss=0.8498, simple_loss=0.6834, pruned_loss=0.5802, over 18259.00 frames. ], tot_loss[loss=0.961, simple_loss=0.7318, pruned_loss=0.7615, over 3578139.45 frames. ], batch size: 60, lr: 4.93e-02, grad_scale: 2.0 +2023-03-08 14:04:30,909 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1203.0, num_to_drop=2, layers_to_drop={1, 3} +2023-03-08 14:04:38,783 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.547e+02 6.963e+02 9.525e+02 1.418e+03 3.358e+03, threshold=1.905e+03, percent-clipped=24.0 +2023-03-08 14:04:39,878 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1214.0, num_to_drop=2, layers_to_drop={0, 3} +2023-03-08 14:04:43,089 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1218.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-08 14:04:44,956 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.93 vs. limit=2.0 +2023-03-08 14:04:47,858 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1224.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 14:05:06,360 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1246.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 14:05:11,318 INFO [train.py:898] (0/4) Epoch 1, batch 1250, loss[loss=0.8111, simple_loss=0.6655, pruned_loss=0.5336, over 16232.00 frames. ], tot_loss[loss=0.9292, simple_loss=0.7165, pruned_loss=0.7143, over 3578705.41 frames. ], batch size: 94, lr: 4.92e-02, grad_scale: 2.0 +2023-03-08 14:05:24,175 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=1266.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:05:28,876 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=1272.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:05:47,787 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=1294.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:05:53,953 INFO [train.py:898] (0/4) Epoch 1, batch 1300, loss[loss=0.7625, simple_loss=0.6356, pruned_loss=0.4867, over 18329.00 frames. ], tot_loss[loss=0.8953, simple_loss=0.6993, pruned_loss=0.6683, over 3573310.33 frames. ], batch size: 56, lr: 4.92e-02, grad_scale: 2.0 +2023-03-08 14:06:04,917 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 4.724e+02 7.521e+02 1.026e+03 1.273e+03 2.275e+03, threshold=2.053e+03, percent-clipped=5.0 +2023-03-08 14:06:08,313 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-08 14:06:38,443 INFO [train.py:898] (0/4) Epoch 1, batch 1350, loss[loss=0.776, simple_loss=0.639, pruned_loss=0.4984, over 12411.00 frames. ], tot_loss[loss=0.8565, simple_loss=0.6787, pruned_loss=0.6203, over 3576934.82 frames. ], batch size: 129, lr: 4.91e-02, grad_scale: 2.0 +2023-03-08 14:07:03,360 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=6.27 vs. limit=5.0 +2023-03-08 14:07:17,226 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=2.36 vs. limit=2.0 +2023-03-08 14:07:24,099 INFO [train.py:898] (0/4) Epoch 1, batch 1400, loss[loss=0.735, simple_loss=0.6296, pruned_loss=0.4461, over 18491.00 frames. ], tot_loss[loss=0.818, simple_loss=0.6587, pruned_loss=0.5743, over 3580743.62 frames. ], batch size: 53, lr: 4.91e-02, grad_scale: 2.0 +2023-03-08 14:07:35,946 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 4.189e+02 7.494e+02 9.042e+02 1.119e+03 2.290e+03, threshold=1.808e+03, percent-clipped=1.0 +2023-03-08 14:07:37,122 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3773, 4.2750, 4.2570, 4.3770, 4.2218, 4.1303, 4.3153, 4.2564], + device='cuda:0'), covar=tensor([0.6380, 0.5539, 0.5658, 0.4879, 0.5405, 0.5512, 0.6372, 0.6575], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0071, 0.0083, 0.0072, 0.0084, 0.0081, 0.0082, 0.0076], + device='cuda:0'), out_proj_covar=tensor([6.8954e-05, 6.3796e-05, 7.6647e-05, 6.2689e-05, 7.6487e-05, 7.4666e-05, + 7.1239e-05, 6.7563e-05], device='cuda:0') +2023-03-08 14:07:42,417 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=6.60 vs. limit=5.0 +2023-03-08 14:07:55,208 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1434.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:08:09,499 INFO [train.py:898] (0/4) Epoch 1, batch 1450, loss[loss=0.7211, simple_loss=0.6112, pruned_loss=0.4404, over 18503.00 frames. ], tot_loss[loss=0.783, simple_loss=0.6403, pruned_loss=0.5334, over 3582332.55 frames. ], batch size: 51, lr: 4.90e-02, grad_scale: 2.0 +2023-03-08 14:08:20,631 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.98 vs. limit=2.0 +2023-03-08 14:08:50,616 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1495.0, num_to_drop=2, layers_to_drop={2, 3} +2023-03-08 14:08:55,493 INFO [train.py:898] (0/4) Epoch 1, batch 1500, loss[loss=0.6466, simple_loss=0.581, pruned_loss=0.3657, over 18307.00 frames. ], tot_loss[loss=0.7479, simple_loss=0.6209, pruned_loss=0.4952, over 3584962.32 frames. ], batch size: 54, lr: 4.89e-02, grad_scale: 2.0 +2023-03-08 14:08:57,545 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1503.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:09:03,785 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1509.0, num_to_drop=2, layers_to_drop={2, 3} +2023-03-08 14:09:07,631 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.367e+02 6.843e+02 9.252e+02 1.183e+03 2.667e+03, threshold=1.850e+03, percent-clipped=7.0 +2023-03-08 14:09:08,092 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=5.74 vs. limit=5.0 +2023-03-08 14:09:25,609 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.09 vs. limit=2.0 +2023-03-08 14:09:43,238 INFO [train.py:898] (0/4) Epoch 1, batch 1550, loss[loss=0.5993, simple_loss=0.55, pruned_loss=0.3291, over 18374.00 frames. ], tot_loss[loss=0.7175, simple_loss=0.6043, pruned_loss=0.4625, over 3563671.42 frames. ], batch size: 56, lr: 4.89e-02, grad_scale: 2.0 +2023-03-08 14:09:43,425 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=1551.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:09:57,941 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1566.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 14:10:30,625 INFO [train.py:898] (0/4) Epoch 1, batch 1600, loss[loss=0.6496, simple_loss=0.5806, pruned_loss=0.3674, over 18466.00 frames. ], tot_loss[loss=0.6897, simple_loss=0.5892, pruned_loss=0.4332, over 3552385.83 frames. ], batch size: 59, lr: 4.88e-02, grad_scale: 4.0 +2023-03-08 14:10:39,894 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1611.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:10:41,383 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 4.398e+02 7.144e+02 8.708e+02 1.119e+03 2.244e+03, threshold=1.742e+03, percent-clipped=2.0 +2023-03-08 14:10:55,102 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1627.0, num_to_drop=2, layers_to_drop={0, 3} +2023-03-08 14:11:10,139 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=5.37 vs. limit=5.0 +2023-03-08 14:11:16,711 INFO [train.py:898] (0/4) Epoch 1, batch 1650, loss[loss=0.5041, simple_loss=0.4778, pruned_loss=0.2652, over 18164.00 frames. ], tot_loss[loss=0.6621, simple_loss=0.5742, pruned_loss=0.4053, over 3564934.08 frames. ], batch size: 44, lr: 4.87e-02, grad_scale: 4.0 +2023-03-08 14:11:36,858 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1672.0, num_to_drop=2, layers_to_drop={1, 3} +2023-03-08 14:11:55,136 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3939, 5.8028, 6.0575, 5.8821, 5.7816, 5.7571, 6.0066, 5.6593], + device='cuda:0'), covar=tensor([0.1175, 0.1125, 0.0538, 0.0675, 0.0914, 0.1151, 0.0813, 0.1150], + device='cuda:0'), in_proj_covar=tensor([0.0086, 0.0072, 0.0064, 0.0073, 0.0077, 0.0076, 0.0078, 0.0076], + device='cuda:0'), out_proj_covar=tensor([8.4147e-05, 7.2105e-05, 6.0245e-05, 6.6172e-05, 7.4169e-05, 7.3199e-05, + 7.9261e-05, 7.4196e-05], device='cuda:0') +2023-03-08 14:12:04,568 INFO [train.py:898] (0/4) Epoch 1, batch 1700, loss[loss=0.697, simple_loss=0.5962, pruned_loss=0.4094, over 12697.00 frames. ], tot_loss[loss=0.6374, simple_loss=0.5611, pruned_loss=0.3808, over 3556365.40 frames. ], batch size: 129, lr: 4.86e-02, grad_scale: 4.0 +2023-03-08 14:12:15,865 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.590e+02 6.591e+02 8.765e+02 1.033e+03 1.987e+03, threshold=1.753e+03, percent-clipped=3.0 +2023-03-08 14:12:53,411 INFO [train.py:898] (0/4) Epoch 1, batch 1750, loss[loss=0.6106, simple_loss=0.5592, pruned_loss=0.3335, over 12079.00 frames. ], tot_loss[loss=0.6144, simple_loss=0.5488, pruned_loss=0.3585, over 3562069.54 frames. ], batch size: 129, lr: 4.86e-02, grad_scale: 4.0 +2023-03-08 14:13:11,950 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9638, 2.7929, 3.3627, 3.2143, 3.0054, 3.0730, 3.5311, 2.5252], + device='cuda:0'), covar=tensor([0.2133, 0.2977, 0.1514, 0.1865, 0.2564, 0.2481, 0.1215, 0.4919], + device='cuda:0'), in_proj_covar=tensor([0.0058, 0.0061, 0.0053, 0.0054, 0.0058, 0.0061, 0.0056, 0.0055], + device='cuda:0'), out_proj_covar=tensor([5.1367e-05, 5.7931e-05, 4.8471e-05, 4.7376e-05, 5.1755e-05, 5.5838e-05, + 4.9609e-05, 4.9014e-05], device='cuda:0') +2023-03-08 14:13:30,719 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1790.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 14:13:41,358 INFO [train.py:898] (0/4) Epoch 1, batch 1800, loss[loss=0.4584, simple_loss=0.4475, pruned_loss=0.2333, over 17631.00 frames. ], tot_loss[loss=0.5966, simple_loss=0.5391, pruned_loss=0.3416, over 3551829.91 frames. ], batch size: 39, lr: 4.85e-02, grad_scale: 4.0 +2023-03-08 14:13:48,982 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=1809.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:13:52,143 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 4.336e+02 7.100e+02 8.751e+02 1.042e+03 2.911e+03, threshold=1.750e+03, percent-clipped=4.0 +2023-03-08 14:13:54,344 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7082, 3.6720, 3.4638, 3.6311, 3.6765, 3.4913, 3.4251, 3.2415], + device='cuda:0'), covar=tensor([0.1344, 0.1044, 0.2919, 0.1157, 0.0879, 0.1806, 0.1894, 0.1872], + device='cuda:0'), in_proj_covar=tensor([0.0062, 0.0057, 0.0068, 0.0056, 0.0058, 0.0063, 0.0066, 0.0070], + device='cuda:0'), out_proj_covar=tensor([6.1289e-05, 5.5704e-05, 6.5950e-05, 5.6124e-05, 5.8807e-05, 6.0836e-05, + 6.2984e-05, 6.8702e-05], device='cuda:0') +2023-03-08 14:14:10,622 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.97 vs. limit=5.0 +2023-03-08 14:14:13,095 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.93 vs. limit=2.0 +2023-03-08 14:14:17,008 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1838.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:14:28,036 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([6.0178, 5.9818, 5.7063, 5.7059, 5.9786, 6.3227, 5.6866, 5.7713], + device='cuda:0'), covar=tensor([0.0716, 0.0783, 0.0959, 0.0804, 0.0867, 0.0458, 0.0932, 0.1234], + device='cuda:0'), in_proj_covar=tensor([0.0107, 0.0098, 0.0092, 0.0100, 0.0100, 0.0078, 0.0095, 0.0105], + device='cuda:0'), out_proj_covar=tensor([9.5710e-05, 9.6481e-05, 8.9141e-05, 9.1565e-05, 1.0009e-04, 7.4387e-05, + 8.4644e-05, 8.9079e-05], device='cuda:0') +2023-03-08 14:14:28,756 INFO [train.py:898] (0/4) Epoch 1, batch 1850, loss[loss=0.439, simple_loss=0.4362, pruned_loss=0.2193, over 18564.00 frames. ], tot_loss[loss=0.5766, simple_loss=0.5284, pruned_loss=0.3235, over 3567395.55 frames. ], batch size: 45, lr: 4.84e-02, grad_scale: 4.0 +2023-03-08 14:14:35,071 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=1857.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:14:35,183 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1857.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:14:39,342 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.66 vs. limit=5.0 +2023-03-08 14:14:46,072 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=5.24 vs. limit=5.0 +2023-03-08 14:14:47,323 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=1870.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:15:15,787 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1899.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-08 14:15:17,456 INFO [train.py:898] (0/4) Epoch 1, batch 1900, loss[loss=0.4601, simple_loss=0.4555, pruned_loss=0.2314, over 18367.00 frames. ], tot_loss[loss=0.5593, simple_loss=0.5188, pruned_loss=0.3084, over 3572138.84 frames. ], batch size: 46, lr: 4.83e-02, grad_scale: 4.0 +2023-03-08 14:15:29,153 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.595e+02 6.983e+02 8.615e+02 1.111e+03 2.145e+03, threshold=1.723e+03, percent-clipped=1.0 +2023-03-08 14:15:34,098 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1918.0, num_to_drop=2, layers_to_drop={1, 3} +2023-03-08 14:15:37,686 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1922.0, num_to_drop=1, layers_to_drop={3} +2023-03-08 14:15:46,027 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=1931.0, num_to_drop=2, layers_to_drop={2, 3} +2023-03-08 14:15:56,280 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6189, 4.6239, 4.6464, 4.6007, 4.0719, 4.5278, 4.6214, 5.2858], + device='cuda:0'), covar=tensor([0.0307, 0.0279, 0.0280, 0.0249, 0.0647, 0.0263, 0.0261, 0.0034], + device='cuda:0'), in_proj_covar=tensor([0.0019, 0.0019, 0.0018, 0.0019, 0.0020, 0.0019, 0.0018, 0.0014], + device='cuda:0'), out_proj_covar=tensor([2.0016e-05, 1.7912e-05, 1.5300e-05, 1.7453e-05, 2.1946e-05, 1.9197e-05, + 1.6318e-05, 1.2412e-05], device='cuda:0') +2023-03-08 14:16:06,107 INFO [train.py:898] (0/4) Epoch 1, batch 1950, loss[loss=0.5249, simple_loss=0.5147, pruned_loss=0.2671, over 17658.00 frames. ], tot_loss[loss=0.5413, simple_loss=0.5086, pruned_loss=0.2934, over 3577752.62 frames. ], batch size: 70, lr: 4.83e-02, grad_scale: 4.0 +2023-03-08 14:16:21,794 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=1967.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:16:54,758 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-2000.pt +2023-03-08 14:16:59,458 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0793, 4.8156, 4.4298, 4.9097, 4.3761, 4.9275, 4.9165, 5.5787], + device='cuda:0'), covar=tensor([0.0297, 0.0389, 0.0808, 0.0271, 0.0941, 0.0272, 0.0390, 0.0042], + device='cuda:0'), in_proj_covar=tensor([0.0019, 0.0019, 0.0018, 0.0019, 0.0021, 0.0019, 0.0018, 0.0014], + device='cuda:0'), out_proj_covar=tensor([1.9558e-05, 1.8031e-05, 1.6070e-05, 1.7259e-05, 2.2850e-05, 1.9109e-05, + 1.6795e-05, 1.2273e-05], device='cuda:0') +2023-03-08 14:17:00,021 INFO [train.py:898] (0/4) Epoch 1, batch 2000, loss[loss=0.448, simple_loss=0.4563, pruned_loss=0.2199, over 18404.00 frames. ], tot_loss[loss=0.5281, simple_loss=0.5019, pruned_loss=0.2821, over 3585793.40 frames. ], batch size: 48, lr: 4.82e-02, grad_scale: 8.0 +2023-03-08 14:17:12,313 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.598e+02 6.794e+02 8.524e+02 1.041e+03 1.894e+03, threshold=1.705e+03, percent-clipped=5.0 +2023-03-08 14:17:37,066 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.15 vs. limit=2.0 +2023-03-08 14:17:53,239 INFO [train.py:898] (0/4) Epoch 1, batch 2050, loss[loss=0.4756, simple_loss=0.4795, pruned_loss=0.2358, over 18018.00 frames. ], tot_loss[loss=0.5089, simple_loss=0.4909, pruned_loss=0.2673, over 3589165.94 frames. ], batch size: 65, lr: 4.81e-02, grad_scale: 8.0 +2023-03-08 14:18:34,049 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2090.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-08 14:18:45,443 INFO [train.py:898] (0/4) Epoch 1, batch 2100, loss[loss=0.4349, simple_loss=0.4657, pruned_loss=0.202, over 18623.00 frames. ], tot_loss[loss=0.492, simple_loss=0.4818, pruned_loss=0.254, over 3594154.33 frames. ], batch size: 52, lr: 4.80e-02, grad_scale: 8.0 +2023-03-08 14:18:58,620 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.519e+02 5.789e+02 7.175e+02 9.165e+02 1.305e+03, threshold=1.435e+03, percent-clipped=0.0 +2023-03-08 14:19:24,271 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=2138.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:19:26,864 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.43 vs. limit=5.0 +2023-03-08 14:19:38,123 INFO [train.py:898] (0/4) Epoch 1, batch 2150, loss[loss=0.4719, simple_loss=0.4759, pruned_loss=0.2339, over 18284.00 frames. ], tot_loss[loss=0.4811, simple_loss=0.476, pruned_loss=0.2454, over 3590091.79 frames. ], batch size: 57, lr: 4.79e-02, grad_scale: 8.0 +2023-03-08 14:19:42,299 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([6.0075, 5.7746, 5.7406, 5.5518, 5.8767, 6.2268, 5.8356, 5.4743], + device='cuda:0'), covar=tensor([0.0554, 0.0701, 0.0626, 0.0571, 0.0744, 0.0389, 0.0558, 0.1380], + device='cuda:0'), in_proj_covar=tensor([0.0117, 0.0115, 0.0096, 0.0100, 0.0111, 0.0088, 0.0099, 0.0124], + device='cuda:0'), out_proj_covar=tensor([1.0636e-04, 1.1509e-04, 9.7243e-05, 9.3232e-05, 1.1209e-04, 8.5339e-05, + 9.1582e-05, 1.1592e-04], device='cuda:0') +2023-03-08 14:20:23,777 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2194.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:20:31,428 INFO [train.py:898] (0/4) Epoch 1, batch 2200, loss[loss=0.478, simple_loss=0.4836, pruned_loss=0.2362, over 18495.00 frames. ], tot_loss[loss=0.4688, simple_loss=0.4693, pruned_loss=0.2359, over 3586881.23 frames. ], batch size: 59, lr: 4.78e-02, grad_scale: 8.0 +2023-03-08 14:20:44,693 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 4.109e+02 6.450e+02 8.005e+02 9.321e+02 1.802e+03, threshold=1.601e+03, percent-clipped=2.0 +2023-03-08 14:20:44,867 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2213.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:20:46,195 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4463, 3.2139, 3.2886, 3.7565, 3.7710, 3.4783, 3.7181, 2.8752], + device='cuda:0'), covar=tensor([0.0842, 0.0916, 0.2830, 0.0482, 0.0444, 0.0644, 0.0476, 0.4951], + device='cuda:0'), in_proj_covar=tensor([0.0063, 0.0062, 0.0062, 0.0056, 0.0059, 0.0064, 0.0058, 0.0058], + device='cuda:0'), out_proj_covar=tensor([5.3520e-05, 5.2512e-05, 5.8060e-05, 4.7764e-05, 4.9170e-05, 5.2434e-05, + 4.7499e-05, 5.7301e-05], device='cuda:0') +2023-03-08 14:20:54,178 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2222.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-08 14:20:58,188 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2226.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:21:23,713 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9409, 4.8461, 4.7958, 4.6019, 5.1771, 4.8759, 4.7979, 4.6668], + device='cuda:0'), covar=tensor([0.0294, 0.0264, 0.0314, 0.0282, 0.0142, 0.0304, 0.0237, 0.0469], + device='cuda:0'), in_proj_covar=tensor([0.0019, 0.0017, 0.0018, 0.0019, 0.0015, 0.0017, 0.0016, 0.0019], + device='cuda:0'), out_proj_covar=tensor([9.8038e-06, 9.4974e-06, 9.4463e-06, 1.1026e-05, 6.4745e-06, 8.8683e-06, + 7.7270e-06, 9.7576e-06], device='cuda:0') +2023-03-08 14:21:24,249 INFO [train.py:898] (0/4) Epoch 1, batch 2250, loss[loss=0.4777, simple_loss=0.4912, pruned_loss=0.2321, over 17938.00 frames. ], tot_loss[loss=0.4577, simple_loss=0.4629, pruned_loss=0.2277, over 3581948.47 frames. ], batch size: 70, lr: 4.77e-02, grad_scale: 8.0 +2023-03-08 14:21:42,680 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2267.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 14:21:45,786 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=2270.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:21:46,741 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2388, 4.9788, 4.8607, 5.2960, 5.0901, 4.8343, 4.9700, 4.6263], + device='cuda:0'), covar=tensor([0.0439, 0.0342, 0.0897, 0.0391, 0.0320, 0.0468, 0.0464, 0.0527], + device='cuda:0'), in_proj_covar=tensor([0.0082, 0.0096, 0.0103, 0.0089, 0.0093, 0.0113, 0.0115, 0.0113], + device='cuda:0'), out_proj_covar=tensor([8.2766e-05, 1.0097e-04, 1.1103e-04, 9.0540e-05, 9.0480e-05, 1.2319e-04, + 1.2329e-04, 1.1548e-04], device='cuda:0') +2023-03-08 14:21:54,305 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.93 vs. limit=2.0 +2023-03-08 14:22:17,952 INFO [train.py:898] (0/4) Epoch 1, batch 2300, loss[loss=0.4506, simple_loss=0.4676, pruned_loss=0.2169, over 16285.00 frames. ], tot_loss[loss=0.4491, simple_loss=0.4581, pruned_loss=0.2211, over 3579558.99 frames. ], batch size: 94, lr: 4.77e-02, grad_scale: 8.0 +2023-03-08 14:22:22,373 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6081, 3.1465, 3.2372, 3.7192, 3.5217, 3.6445, 3.9325, 3.3520], + device='cuda:0'), covar=tensor([0.0873, 0.1048, 0.4460, 0.0712, 0.0799, 0.0791, 0.0493, 0.4178], + device='cuda:0'), in_proj_covar=tensor([0.0066, 0.0063, 0.0064, 0.0058, 0.0062, 0.0066, 0.0061, 0.0057], + device='cuda:0'), out_proj_covar=tensor([5.4736e-05, 5.3410e-05, 6.1449e-05, 4.8306e-05, 5.0565e-05, 5.3862e-05, + 4.8353e-05, 5.8902e-05], device='cuda:0') +2023-03-08 14:22:31,078 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.655e+02 6.513e+02 7.644e+02 9.367e+02 1.783e+03, threshold=1.529e+03, percent-clipped=1.0 +2023-03-08 14:22:33,316 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=2315.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:23:00,868 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2340.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:23:11,897 INFO [train.py:898] (0/4) Epoch 1, batch 2350, loss[loss=0.3571, simple_loss=0.3877, pruned_loss=0.1632, over 18499.00 frames. ], tot_loss[loss=0.4399, simple_loss=0.4526, pruned_loss=0.2144, over 3584859.53 frames. ], batch size: 47, lr: 4.76e-02, grad_scale: 16.0 +2023-03-08 14:24:05,195 INFO [train.py:898] (0/4) Epoch 1, batch 2400, loss[loss=0.4703, simple_loss=0.476, pruned_loss=0.2323, over 18279.00 frames. ], tot_loss[loss=0.4309, simple_loss=0.4471, pruned_loss=0.208, over 3590246.66 frames. ], batch size: 60, lr: 4.75e-02, grad_scale: 16.0 +2023-03-08 14:24:05,539 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2401.0, num_to_drop=2, layers_to_drop={1, 2} +2023-03-08 14:24:09,514 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9996, 3.4041, 3.7882, 3.9681, 3.8171, 4.0570, 4.1632, 3.8175], + device='cuda:0'), covar=tensor([0.0326, 0.0877, 0.1697, 0.0261, 0.0416, 0.0236, 0.0215, 0.1721], + device='cuda:0'), in_proj_covar=tensor([0.0064, 0.0064, 0.0063, 0.0056, 0.0061, 0.0064, 0.0060, 0.0055], + device='cuda:0'), out_proj_covar=tensor([5.3412e-05, 5.4651e-05, 6.1543e-05, 4.7014e-05, 5.1655e-05, 5.0988e-05, + 4.7085e-05, 5.7834e-05], device='cuda:0') +2023-03-08 14:24:17,834 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.759e+02 6.046e+02 7.526e+02 8.987e+02 1.408e+03, threshold=1.505e+03, percent-clipped=0.0 +2023-03-08 14:24:50,587 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5685, 4.4710, 4.2313, 4.0566, 4.5534, 4.8648, 4.1952, 4.0725], + device='cuda:0'), covar=tensor([0.0802, 0.0442, 0.0336, 0.0792, 0.0294, 0.0292, 0.0751, 0.0462], + device='cuda:0'), in_proj_covar=tensor([0.0030, 0.0023, 0.0026, 0.0025, 0.0027, 0.0024, 0.0027, 0.0028], + device='cuda:0'), out_proj_covar=tensor([3.1985e-05, 2.3807e-05, 2.5189e-05, 2.4616e-05, 2.6099e-05, 2.2618e-05, + 2.7895e-05, 2.9019e-05], device='cuda:0') +2023-03-08 14:24:58,101 INFO [train.py:898] (0/4) Epoch 1, batch 2450, loss[loss=0.4139, simple_loss=0.4446, pruned_loss=0.1916, over 18493.00 frames. ], tot_loss[loss=0.4215, simple_loss=0.4411, pruned_loss=0.2015, over 3606551.29 frames. ], batch size: 51, lr: 4.74e-02, grad_scale: 16.0 +2023-03-08 14:25:40,431 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1804, 5.2674, 5.2340, 4.7915, 4.6714, 4.8821, 4.6406, 4.9367], + device='cuda:0'), covar=tensor([0.0654, 0.0242, 0.0239, 0.0326, 0.0439, 0.0274, 0.0502, 0.0367], + device='cuda:0'), in_proj_covar=tensor([0.0060, 0.0068, 0.0055, 0.0058, 0.0069, 0.0066, 0.0071, 0.0063], + device='cuda:0'), out_proj_covar=tensor([6.0628e-05, 6.6626e-05, 5.3811e-05, 6.2590e-05, 7.3017e-05, 6.6923e-05, + 7.8224e-05, 6.1541e-05], device='cuda:0') +2023-03-08 14:25:45,438 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2494.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 14:25:53,139 INFO [train.py:898] (0/4) Epoch 1, batch 2500, loss[loss=0.3868, simple_loss=0.4331, pruned_loss=0.1703, over 18026.00 frames. ], tot_loss[loss=0.4149, simple_loss=0.437, pruned_loss=0.1968, over 3596859.97 frames. ], batch size: 62, lr: 4.73e-02, grad_scale: 16.0 +2023-03-08 14:25:54,857 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.97 vs. limit=2.0 +2023-03-08 14:26:05,162 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.211e+02 5.891e+02 7.187e+02 8.781e+02 1.767e+03, threshold=1.437e+03, percent-clipped=2.0 +2023-03-08 14:26:05,445 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2513.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:26:19,807 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2526.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:26:37,666 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=2542.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:26:47,805 INFO [train.py:898] (0/4) Epoch 1, batch 2550, loss[loss=0.3277, simple_loss=0.382, pruned_loss=0.1367, over 18267.00 frames. ], tot_loss[loss=0.4089, simple_loss=0.4331, pruned_loss=0.1927, over 3600838.11 frames. ], batch size: 49, lr: 4.72e-02, grad_scale: 16.0 +2023-03-08 14:26:58,313 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=2561.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:27:12,127 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=2574.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:27:42,176 INFO [train.py:898] (0/4) Epoch 1, batch 2600, loss[loss=0.3702, simple_loss=0.4079, pruned_loss=0.1662, over 18362.00 frames. ], tot_loss[loss=0.4036, simple_loss=0.4302, pruned_loss=0.1887, over 3596885.91 frames. ], batch size: 46, lr: 4.71e-02, grad_scale: 16.0 +2023-03-08 14:27:48,448 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5780, 6.0207, 5.9245, 5.8472, 5.5070, 5.8835, 6.0980, 5.9062], + device='cuda:0'), covar=tensor([0.0458, 0.0548, 0.0248, 0.0293, 0.0480, 0.0375, 0.0336, 0.0371], + device='cuda:0'), in_proj_covar=tensor([0.0140, 0.0136, 0.0099, 0.0114, 0.0129, 0.0104, 0.0108, 0.0112], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-08 14:27:55,523 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.922e+02 6.237e+02 7.424e+02 9.705e+02 2.435e+03, threshold=1.485e+03, percent-clipped=1.0 +2023-03-08 14:28:20,589 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9656, 5.0147, 4.4515, 4.7656, 4.9034, 4.9113, 4.4877, 4.1734], + device='cuda:0'), covar=tensor([0.0391, 0.0200, 0.0172, 0.0185, 0.0138, 0.0216, 0.0387, 0.0318], + device='cuda:0'), in_proj_covar=tensor([0.0030, 0.0022, 0.0025, 0.0023, 0.0026, 0.0024, 0.0027, 0.0028], + device='cuda:0'), out_proj_covar=tensor([3.5635e-05, 2.4020e-05, 2.5537e-05, 2.5487e-05, 2.6134e-05, 2.4129e-05, + 2.9325e-05, 3.1399e-05], device='cuda:0') +2023-03-08 14:28:26,240 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.05 vs. limit=2.0 +2023-03-08 14:28:36,992 INFO [train.py:898] (0/4) Epoch 1, batch 2650, loss[loss=0.4335, simple_loss=0.4576, pruned_loss=0.2047, over 18264.00 frames. ], tot_loss[loss=0.3988, simple_loss=0.4275, pruned_loss=0.1852, over 3590402.87 frames. ], batch size: 57, lr: 4.70e-02, grad_scale: 16.0 +2023-03-08 14:28:44,397 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5001, 3.6851, 3.5425, 3.8683, 3.5920, 4.0062, 3.5707, 3.5183], + device='cuda:0'), covar=tensor([0.0274, 0.0322, 0.0220, 0.0179, 0.0264, 0.0150, 0.0182, 0.0304], + device='cuda:0'), in_proj_covar=tensor([0.0022, 0.0020, 0.0023, 0.0023, 0.0022, 0.0021, 0.0019, 0.0022], + device='cuda:0'), out_proj_covar=tensor([1.5996e-05, 1.3256e-05, 1.5638e-05, 1.6936e-05, 1.4908e-05, 1.3939e-05, + 1.2769e-05, 1.5634e-05], device='cuda:0') +2023-03-08 14:28:53,331 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.88 vs. limit=5.0 +2023-03-08 14:29:14,977 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2685.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:29:27,194 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=2696.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 14:29:32,054 INFO [train.py:898] (0/4) Epoch 1, batch 2700, loss[loss=0.3956, simple_loss=0.4201, pruned_loss=0.1856, over 18385.00 frames. ], tot_loss[loss=0.3945, simple_loss=0.4246, pruned_loss=0.1824, over 3587590.90 frames. ], batch size: 48, lr: 4.69e-02, grad_scale: 16.0 +2023-03-08 14:29:45,132 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.900e+02 5.817e+02 6.597e+02 8.414e+02 1.857e+03, threshold=1.319e+03, percent-clipped=1.0 +2023-03-08 14:30:16,191 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.63 vs. limit=5.0 +2023-03-08 14:30:21,408 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7245, 4.3055, 4.3817, 4.5340, 4.3936, 4.3587, 4.9274, 4.6617], + device='cuda:0'), covar=tensor([0.0183, 0.0375, 0.0377, 0.0293, 0.0337, 0.0317, 0.0248, 0.0253], + device='cuda:0'), in_proj_covar=tensor([0.0053, 0.0052, 0.0053, 0.0052, 0.0056, 0.0061, 0.0053, 0.0050], + device='cuda:0'), out_proj_covar=tensor([5.4205e-05, 4.9496e-05, 5.3795e-05, 5.0271e-05, 5.9588e-05, 7.1349e-05, + 5.4809e-05, 4.7272e-05], device='cuda:0') +2023-03-08 14:30:22,439 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2746.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-08 14:30:27,318 INFO [train.py:898] (0/4) Epoch 1, batch 2750, loss[loss=0.4064, simple_loss=0.4473, pruned_loss=0.1827, over 18393.00 frames. ], tot_loss[loss=0.3912, simple_loss=0.4233, pruned_loss=0.1797, over 3591895.14 frames. ], batch size: 52, lr: 4.68e-02, grad_scale: 16.0 +2023-03-08 14:30:51,313 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4261, 3.0574, 2.8140, 3.1411, 3.1900, 3.4349, 2.8799, 3.2744], + device='cuda:0'), covar=tensor([0.0096, 0.0237, 0.0341, 0.0171, 0.0136, 0.0106, 0.0340, 0.0093], + device='cuda:0'), in_proj_covar=tensor([0.0019, 0.0023, 0.0023, 0.0019, 0.0016, 0.0019, 0.0022, 0.0019], + device='cuda:0'), out_proj_covar=tensor([1.5326e-05, 1.9269e-05, 2.0496e-05, 1.5626e-05, 1.2327e-05, 1.4888e-05, + 1.8542e-05, 1.5228e-05], device='cuda:0') +2023-03-08 14:30:52,553 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.25 vs. limit=5.0 +2023-03-08 14:31:19,077 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2567, 5.7686, 5.6311, 5.4467, 5.1522, 5.5853, 5.7579, 5.6008], + device='cuda:0'), covar=tensor([0.0645, 0.0612, 0.0369, 0.0401, 0.0661, 0.0435, 0.0439, 0.0549], + device='cuda:0'), in_proj_covar=tensor([0.0156, 0.0151, 0.0115, 0.0130, 0.0146, 0.0118, 0.0121, 0.0129], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-08 14:31:22,077 INFO [train.py:898] (0/4) Epoch 1, batch 2800, loss[loss=0.32, simple_loss=0.3599, pruned_loss=0.1401, over 18164.00 frames. ], tot_loss[loss=0.3889, simple_loss=0.4214, pruned_loss=0.1783, over 3595748.78 frames. ], batch size: 44, lr: 4.67e-02, grad_scale: 16.0 +2023-03-08 14:31:27,952 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.16 vs. limit=5.0 +2023-03-08 14:31:35,363 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.828e+02 5.568e+02 7.020e+02 9.459e+02 2.422e+03, threshold=1.404e+03, percent-clipped=9.0 +2023-03-08 14:31:45,266 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3985, 5.8808, 5.8792, 5.6646, 5.3044, 5.8627, 5.9303, 5.7480], + device='cuda:0'), covar=tensor([0.0504, 0.0664, 0.0283, 0.0333, 0.0555, 0.0277, 0.0407, 0.0399], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0148, 0.0112, 0.0127, 0.0143, 0.0115, 0.0118, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-08 14:31:47,504 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2824.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:32:16,541 INFO [train.py:898] (0/4) Epoch 1, batch 2850, loss[loss=0.4125, simple_loss=0.4452, pruned_loss=0.1899, over 18566.00 frames. ], tot_loss[loss=0.3863, simple_loss=0.4198, pruned_loss=0.1764, over 3584192.14 frames. ], batch size: 54, lr: 4.66e-02, grad_scale: 16.0 +2023-03-08 14:32:53,592 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2885.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-08 14:33:10,882 INFO [train.py:898] (0/4) Epoch 1, batch 2900, loss[loss=0.3989, simple_loss=0.4379, pruned_loss=0.1799, over 17860.00 frames. ], tot_loss[loss=0.3839, simple_loss=0.4186, pruned_loss=0.1747, over 3578836.48 frames. ], batch size: 70, lr: 4.65e-02, grad_scale: 16.0 +2023-03-08 14:33:21,902 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0406, 3.0171, 2.7636, 3.5954, 3.2908, 3.4138, 2.7363, 3.3815], + device='cuda:0'), covar=tensor([0.0132, 0.0600, 0.0667, 0.0282, 0.0303, 0.0337, 0.0609, 0.0223], + device='cuda:0'), in_proj_covar=tensor([0.0019, 0.0026, 0.0025, 0.0020, 0.0019, 0.0020, 0.0024, 0.0020], + device='cuda:0'), out_proj_covar=tensor([1.5409e-05, 2.2776e-05, 2.2579e-05, 1.7486e-05, 1.4578e-05, 1.5987e-05, + 2.1117e-05, 1.7011e-05], device='cuda:0') +2023-03-08 14:33:23,642 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.949e+02 5.658e+02 6.946e+02 8.980e+02 2.248e+03, threshold=1.389e+03, percent-clipped=2.0 +2023-03-08 14:33:25,125 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0181, 3.0816, 3.1435, 3.3354, 3.1211, 3.3218, 3.1648, 3.5738], + device='cuda:0'), covar=tensor([0.2091, 0.1586, 0.0416, 0.0329, 0.0677, 0.0703, 0.0461, 0.0209], + device='cuda:0'), in_proj_covar=tensor([0.0046, 0.0046, 0.0052, 0.0048, 0.0048, 0.0046, 0.0041, 0.0047], + device='cuda:0'), out_proj_covar=tensor([5.2061e-05, 5.1196e-05, 4.6421e-05, 3.9992e-05, 4.3795e-05, 4.6661e-05, + 3.7192e-05, 3.7822e-05], device='cuda:0') +2023-03-08 14:33:50,318 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=2937.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 14:34:05,709 INFO [train.py:898] (0/4) Epoch 1, batch 2950, loss[loss=0.3985, simple_loss=0.4394, pruned_loss=0.1788, over 18506.00 frames. ], tot_loss[loss=0.3792, simple_loss=0.416, pruned_loss=0.1713, over 3594812.72 frames. ], batch size: 59, lr: 4.64e-02, grad_scale: 16.0 +2023-03-08 14:34:54,933 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=2996.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:34:57,619 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=2998.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-08 14:35:00,656 INFO [train.py:898] (0/4) Epoch 1, batch 3000, loss[loss=0.383, simple_loss=0.4198, pruned_loss=0.1731, over 18342.00 frames. ], tot_loss[loss=0.3745, simple_loss=0.4122, pruned_loss=0.1685, over 3599533.77 frames. ], batch size: 56, lr: 4.63e-02, grad_scale: 8.0 +2023-03-08 14:35:00,659 INFO [train.py:923] (0/4) Computing validation loss +2023-03-08 14:35:12,699 INFO [train.py:932] (0/4) Epoch 1, validation: loss=0.2954, simple_loss=0.387, pruned_loss=0.102, over 944034.00 frames. +2023-03-08 14:35:12,700 INFO [train.py:933] (0/4) Maximum memory allocated so far is 17514MB +2023-03-08 14:35:26,916 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.611e+02 5.968e+02 7.272e+02 9.035e+02 2.166e+03, threshold=1.454e+03, percent-clipped=4.0 +2023-03-08 14:35:32,550 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3019.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:35:57,653 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3041.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:35:58,025 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.00 vs. limit=2.0 +2023-03-08 14:36:00,947 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=3044.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:36:08,339 INFO [train.py:898] (0/4) Epoch 1, batch 3050, loss[loss=0.3898, simple_loss=0.4314, pruned_loss=0.1741, over 16033.00 frames. ], tot_loss[loss=0.3727, simple_loss=0.411, pruned_loss=0.1673, over 3597292.31 frames. ], batch size: 94, lr: 4.62e-02, grad_scale: 8.0 +2023-03-08 14:36:24,750 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.99 vs. limit=2.0 +2023-03-08 14:36:41,067 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3080.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-08 14:36:49,154 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7114, 5.0586, 5.3884, 5.2625, 5.2298, 4.9452, 5.5157, 5.1351], + device='cuda:0'), covar=tensor([0.0123, 0.0078, 0.0027, 0.0028, 0.0026, 0.0069, 0.0022, 0.0062], + device='cuda:0'), in_proj_covar=tensor([0.0022, 0.0020, 0.0018, 0.0019, 0.0013, 0.0019, 0.0015, 0.0018], + device='cuda:0'), out_proj_covar=tensor([1.2978e-05, 1.2092e-05, 9.1798e-06, 1.0968e-05, 6.4565e-06, 1.1001e-05, + 8.3216e-06, 1.0518e-05], device='cuda:0') +2023-03-08 14:37:04,048 INFO [train.py:898] (0/4) Epoch 1, batch 3100, loss[loss=0.3275, simple_loss=0.3564, pruned_loss=0.1493, over 18466.00 frames. ], tot_loss[loss=0.3693, simple_loss=0.4085, pruned_loss=0.165, over 3590078.53 frames. ], batch size: 43, lr: 4.61e-02, grad_scale: 8.0 +2023-03-08 14:37:18,603 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.360e+02 5.789e+02 7.194e+02 8.721e+02 2.161e+03, threshold=1.439e+03, percent-clipped=3.0 +2023-03-08 14:37:30,616 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8207, 4.6680, 4.8109, 4.4641, 3.8716, 4.9835, 4.7831, 4.8633], + device='cuda:0'), covar=tensor([0.0187, 0.0238, 0.0163, 0.0398, 0.1866, 0.0130, 0.0174, 0.0098], + device='cuda:0'), in_proj_covar=tensor([0.0035, 0.0036, 0.0029, 0.0037, 0.0057, 0.0029, 0.0033, 0.0032], + device='cuda:0'), out_proj_covar=tensor([1.8270e-05, 2.1628e-05, 1.3560e-05, 2.2250e-05, 4.0635e-05, 1.3294e-05, + 1.5873e-05, 1.6054e-05], device='cuda:0') +2023-03-08 14:37:59,563 INFO [train.py:898] (0/4) Epoch 1, batch 3150, loss[loss=0.3806, simple_loss=0.4203, pruned_loss=0.1705, over 17763.00 frames. ], tot_loss[loss=0.3659, simple_loss=0.4063, pruned_loss=0.1628, over 3598709.05 frames. ], batch size: 70, lr: 4.60e-02, grad_scale: 8.0 +2023-03-08 14:38:30,706 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3180.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:38:54,779 INFO [train.py:898] (0/4) Epoch 1, batch 3200, loss[loss=0.3376, simple_loss=0.3738, pruned_loss=0.1507, over 18256.00 frames. ], tot_loss[loss=0.3629, simple_loss=0.4043, pruned_loss=0.1607, over 3601954.18 frames. ], batch size: 45, lr: 4.59e-02, grad_scale: 8.0 +2023-03-08 14:39:08,065 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.309e+02 6.227e+02 7.762e+02 9.563e+02 2.131e+03, threshold=1.552e+03, percent-clipped=3.0 +2023-03-08 14:39:49,492 INFO [train.py:898] (0/4) Epoch 1, batch 3250, loss[loss=0.3397, simple_loss=0.3889, pruned_loss=0.1453, over 18565.00 frames. ], tot_loss[loss=0.3605, simple_loss=0.4025, pruned_loss=0.1592, over 3607212.53 frames. ], batch size: 49, lr: 4.58e-02, grad_scale: 8.0 +2023-03-08 14:39:52,195 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=2.07 vs. limit=2.0 +2023-03-08 14:39:54,432 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.71 vs. limit=2.0 +2023-03-08 14:40:00,596 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.38 vs. limit=2.0 +2023-03-08 14:40:00,725 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.84 vs. limit=2.0 +2023-03-08 14:40:30,553 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3288.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:40:35,657 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3293.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:40:45,135 INFO [train.py:898] (0/4) Epoch 1, batch 3300, loss[loss=0.3381, simple_loss=0.3828, pruned_loss=0.1467, over 18287.00 frames. ], tot_loss[loss=0.3601, simple_loss=0.4023, pruned_loss=0.159, over 3590057.00 frames. ], batch size: 49, lr: 4.57e-02, grad_scale: 8.0 +2023-03-08 14:40:48,028 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=5.08 vs. limit=5.0 +2023-03-08 14:40:58,793 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 4.218e+02 5.885e+02 6.635e+02 8.607e+02 2.408e+03, threshold=1.327e+03, percent-clipped=3.0 +2023-03-08 14:41:29,216 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3341.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:41:38,100 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3349.0, num_to_drop=2, layers_to_drop={0, 2} +2023-03-08 14:41:40,517 INFO [train.py:898] (0/4) Epoch 1, batch 3350, loss[loss=0.342, simple_loss=0.3865, pruned_loss=0.1488, over 18500.00 frames. ], tot_loss[loss=0.3573, simple_loss=0.4004, pruned_loss=0.1571, over 3597844.63 frames. ], batch size: 47, lr: 4.56e-02, grad_scale: 8.0 +2023-03-08 14:41:54,361 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.92 vs. limit=2.0 +2023-03-08 14:42:06,535 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3375.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:42:22,237 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=3389.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:42:36,022 INFO [train.py:898] (0/4) Epoch 1, batch 3400, loss[loss=0.3914, simple_loss=0.4291, pruned_loss=0.1768, over 16389.00 frames. ], tot_loss[loss=0.3546, simple_loss=0.3985, pruned_loss=0.1553, over 3600368.21 frames. ], batch size: 94, lr: 4.55e-02, grad_scale: 8.0 +2023-03-08 14:42:50,979 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.931e+02 5.703e+02 6.610e+02 8.336e+02 1.336e+03, threshold=1.322e+03, percent-clipped=1.0 +2023-03-08 14:43:31,445 INFO [train.py:898] (0/4) Epoch 1, batch 3450, loss[loss=0.3352, simple_loss=0.3942, pruned_loss=0.1381, over 18280.00 frames. ], tot_loss[loss=0.3522, simple_loss=0.3962, pruned_loss=0.1541, over 3593189.71 frames. ], batch size: 57, lr: 4.54e-02, grad_scale: 8.0 +2023-03-08 14:43:50,108 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8808, 2.8969, 1.2858, 3.3493, 2.6945, 3.8307, 2.1621, 3.4106], + device='cuda:0'), covar=tensor([0.0132, 0.1021, 0.2055, 0.0483, 0.0612, 0.0179, 0.1115, 0.0333], + device='cuda:0'), in_proj_covar=tensor([0.0027, 0.0045, 0.0045, 0.0027, 0.0032, 0.0029, 0.0041, 0.0032], + device='cuda:0'), out_proj_covar=tensor([2.2628e-05, 4.5715e-05, 4.6483e-05, 2.7172e-05, 2.6216e-05, 2.4672e-05, + 3.8518e-05, 2.7343e-05], device='cuda:0') +2023-03-08 14:44:04,074 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3480.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:44:27,331 INFO [train.py:898] (0/4) Epoch 1, batch 3500, loss[loss=0.3854, simple_loss=0.4279, pruned_loss=0.1714, over 16225.00 frames. ], tot_loss[loss=0.3496, simple_loss=0.3946, pruned_loss=0.1523, over 3599676.75 frames. ], batch size: 96, lr: 4.53e-02, grad_scale: 8.0 +2023-03-08 14:44:42,687 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 4.394e+02 6.336e+02 7.785e+02 9.331e+02 2.014e+03, threshold=1.557e+03, percent-clipped=6.0 +2023-03-08 14:44:58,031 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=3528.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:45:16,115 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3546.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:45:20,682 INFO [train.py:898] (0/4) Epoch 1, batch 3550, loss[loss=0.3871, simple_loss=0.4294, pruned_loss=0.1724, over 17911.00 frames. ], tot_loss[loss=0.3505, simple_loss=0.3948, pruned_loss=0.1531, over 3599385.17 frames. ], batch size: 65, lr: 4.51e-02, grad_scale: 8.0 +2023-03-08 14:46:04,622 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3593.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:46:12,801 INFO [train.py:898] (0/4) Epoch 1, batch 3600, loss[loss=0.367, simple_loss=0.4119, pruned_loss=0.161, over 18619.00 frames. ], tot_loss[loss=0.3488, simple_loss=0.3933, pruned_loss=0.1521, over 3590475.63 frames. ], batch size: 52, lr: 4.50e-02, grad_scale: 8.0 +2023-03-08 14:46:19,835 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3607.0, num_to_drop=2, layers_to_drop={0, 1} +2023-03-08 14:46:26,697 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.777e+02 7.280e+02 9.972e+02 1.228e+03 2.916e+03, threshold=1.994e+03, percent-clipped=11.0 +2023-03-08 14:46:48,018 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-1.pt +2023-03-08 14:47:17,209 INFO [train.py:898] (0/4) Epoch 2, batch 0, loss[loss=0.3739, simple_loss=0.4219, pruned_loss=0.1629, over 18019.00 frames. ], tot_loss[loss=0.3739, simple_loss=0.4219, pruned_loss=0.1629, over 18019.00 frames. ], batch size: 65, lr: 4.41e-02, grad_scale: 8.0 +2023-03-08 14:47:17,211 INFO [train.py:923] (0/4) Computing validation loss +2023-03-08 14:47:23,864 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1267, 4.1971, 3.4790, 4.1341, 4.0605, 3.7208, 4.1081, 3.5286], + device='cuda:0'), covar=tensor([0.0352, 0.0302, 0.1689, 0.0390, 0.0352, 0.0578, 0.0391, 0.0615], + device='cuda:0'), in_proj_covar=tensor([0.0115, 0.0122, 0.0169, 0.0114, 0.0111, 0.0139, 0.0134, 0.0139], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 14:47:29,038 INFO [train.py:932] (0/4) Epoch 2, validation: loss=0.2643, simple_loss=0.3556, pruned_loss=0.08646, over 944034.00 frames. +2023-03-08 14:47:29,038 INFO [train.py:933] (0/4) Maximum memory allocated so far is 18384MB +2023-03-08 14:47:35,830 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=3641.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:47:39,164 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3644.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:48:16,082 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3675.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:48:17,163 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3676.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:48:26,904 INFO [train.py:898] (0/4) Epoch 2, batch 50, loss[loss=0.3314, simple_loss=0.3806, pruned_loss=0.141, over 18408.00 frames. ], tot_loss[loss=0.3318, simple_loss=0.381, pruned_loss=0.1413, over 821905.97 frames. ], batch size: 48, lr: 4.40e-02, grad_scale: 8.0 +2023-03-08 14:48:54,761 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1471, 5.1076, 4.0619, 5.1566, 5.0871, 4.7904, 5.1391, 4.3694], + device='cuda:0'), covar=tensor([0.0518, 0.0354, 0.1898, 0.0676, 0.0384, 0.0412, 0.0452, 0.0653], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0132, 0.0189, 0.0122, 0.0120, 0.0148, 0.0142, 0.0151], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 14:49:00,566 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 4.344e+02 6.253e+02 7.943e+02 9.806e+02 1.695e+03, threshold=1.589e+03, percent-clipped=0.0 +2023-03-08 14:49:11,365 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=3723.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:49:25,277 INFO [train.py:898] (0/4) Epoch 2, batch 100, loss[loss=0.3132, simple_loss=0.3674, pruned_loss=0.1295, over 18361.00 frames. ], tot_loss[loss=0.329, simple_loss=0.3791, pruned_loss=0.1394, over 1430160.37 frames. ], batch size: 46, lr: 4.39e-02, grad_scale: 8.0 +2023-03-08 14:49:27,925 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3737.0, num_to_drop=2, layers_to_drop={2, 3} +2023-03-08 14:50:22,838 INFO [train.py:898] (0/4) Epoch 2, batch 150, loss[loss=0.3452, simple_loss=0.3908, pruned_loss=0.1498, over 18502.00 frames. ], tot_loss[loss=0.3328, simple_loss=0.3825, pruned_loss=0.1415, over 1913811.39 frames. ], batch size: 51, lr: 4.38e-02, grad_scale: 8.0 +2023-03-08 14:50:28,583 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8596, 4.3105, 4.2711, 3.9716, 3.5378, 3.2382, 3.7052, 3.1906], + device='cuda:0'), covar=tensor([0.0212, 0.0301, 0.0217, 0.0140, 0.0237, 0.0403, 0.0221, 0.0307], + device='cuda:0'), in_proj_covar=tensor([0.0028, 0.0029, 0.0027, 0.0026, 0.0035, 0.0027, 0.0030, 0.0032], + device='cuda:0'), out_proj_covar=tensor([5.1732e-05, 6.1996e-05, 4.5461e-05, 4.7415e-05, 5.7854e-05, 4.7605e-05, + 4.6087e-05, 5.6105e-05], device='cuda:0') +2023-03-08 14:50:55,228 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.556e+02 6.308e+02 7.748e+02 1.009e+03 1.866e+03, threshold=1.550e+03, percent-clipped=3.0 +2023-03-08 14:51:21,237 INFO [train.py:898] (0/4) Epoch 2, batch 200, loss[loss=0.3593, simple_loss=0.4033, pruned_loss=0.1577, over 18131.00 frames. ], tot_loss[loss=0.3318, simple_loss=0.3818, pruned_loss=0.1409, over 2298314.12 frames. ], batch size: 62, lr: 4.37e-02, grad_scale: 8.0 +2023-03-08 14:52:20,102 INFO [train.py:898] (0/4) Epoch 2, batch 250, loss[loss=0.334, simple_loss=0.3897, pruned_loss=0.1391, over 17177.00 frames. ], tot_loss[loss=0.3334, simple_loss=0.3825, pruned_loss=0.1421, over 2585649.18 frames. ], batch size: 78, lr: 4.36e-02, grad_scale: 8.0 +2023-03-08 14:52:39,890 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=3902.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:52:53,090 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.269e+02 6.436e+02 8.265e+02 1.082e+03 2.310e+03, threshold=1.653e+03, percent-clipped=6.0 +2023-03-08 14:53:08,313 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=3926.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:53:09,586 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1724, 3.9171, 4.0427, 3.9536, 2.8305, 3.1650, 3.7819, 3.7052], + device='cuda:0'), covar=tensor([0.0412, 0.0617, 0.0057, 0.0148, 0.0650, 0.0756, 0.0151, 0.0055], + device='cuda:0'), in_proj_covar=tensor([0.0032, 0.0022, 0.0017, 0.0025, 0.0037, 0.0040, 0.0025, 0.0017], + device='cuda:0'), out_proj_covar=tensor([3.7465e-05, 2.5614e-05, 1.7193e-05, 2.7304e-05, 4.2230e-05, 4.5974e-05, + 2.8156e-05, 1.6798e-05], device='cuda:0') +2023-03-08 14:53:17,083 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2084, 3.9633, 3.7437, 3.6475, 3.8935, 3.7423, 3.7720, 3.4635], + device='cuda:0'), covar=tensor([0.0697, 0.0237, 0.0188, 0.0308, 0.0115, 0.0328, 0.0153, 0.0464], + device='cuda:0'), in_proj_covar=tensor([0.0028, 0.0022, 0.0020, 0.0022, 0.0014, 0.0025, 0.0017, 0.0026], + device='cuda:0'), out_proj_covar=tensor([1.7596e-05, 1.3512e-05, 1.1386e-05, 1.3415e-05, 7.5386e-06, 1.6288e-05, + 1.0145e-05, 1.6361e-05], device='cuda:0') +2023-03-08 14:53:18,696 INFO [train.py:898] (0/4) Epoch 2, batch 300, loss[loss=0.3016, simple_loss=0.355, pruned_loss=0.1241, over 18483.00 frames. ], tot_loss[loss=0.3312, simple_loss=0.3813, pruned_loss=0.1406, over 2806482.95 frames. ], batch size: 47, lr: 4.35e-02, grad_scale: 8.0 +2023-03-08 14:53:22,635 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.95 vs. limit=2.0 +2023-03-08 14:53:29,565 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=3944.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 14:54:17,745 INFO [train.py:898] (0/4) Epoch 2, batch 350, loss[loss=0.2989, simple_loss=0.3554, pruned_loss=0.1212, over 18502.00 frames. ], tot_loss[loss=0.3276, simple_loss=0.3785, pruned_loss=0.1383, over 2991456.50 frames. ], batch size: 47, lr: 4.34e-02, grad_scale: 8.0 +2023-03-08 14:54:18,661 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.79 vs. limit=5.0 +2023-03-08 14:54:20,170 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.6210, 5.3800, 5.5890, 5.3552, 5.4343, 6.1037, 5.6045, 5.6091], + device='cuda:0'), covar=tensor([0.0420, 0.0493, 0.0498, 0.0472, 0.0807, 0.0424, 0.0478, 0.1104], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0146, 0.0132, 0.0119, 0.0168, 0.0163, 0.0122, 0.0175], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002], + device='cuda:0') +2023-03-08 14:54:20,395 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=3987.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:54:26,778 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=3992.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 14:54:27,914 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1268, 2.5563, 1.6453, 2.9838, 2.5199, 4.3060, 1.8853, 3.4638], + device='cuda:0'), covar=tensor([0.0121, 0.1165, 0.1998, 0.1280, 0.0690, 0.0094, 0.1447, 0.0363], + device='cuda:0'), in_proj_covar=tensor([0.0039, 0.0064, 0.0067, 0.0034, 0.0051, 0.0038, 0.0063, 0.0050], + device='cuda:0'), out_proj_covar=tensor([3.0611e-05, 6.7906e-05, 7.4566e-05, 3.6332e-05, 4.5578e-05, 3.0549e-05, + 6.2787e-05, 4.4591e-05], device='cuda:0') +2023-03-08 14:54:35,970 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-4000.pt +2023-03-08 14:54:55,409 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.6101, 6.1138, 5.9494, 5.7149, 5.3998, 5.8300, 6.1686, 6.0526], + device='cuda:0'), covar=tensor([0.0634, 0.0450, 0.0271, 0.0465, 0.0975, 0.0405, 0.0433, 0.0404], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0192, 0.0149, 0.0182, 0.0221, 0.0162, 0.0170, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 14:54:56,245 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.948e+02 5.340e+02 6.954e+02 8.950e+02 2.037e+03, threshold=1.391e+03, percent-clipped=3.0 +2023-03-08 14:55:03,890 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6445, 4.7584, 3.4178, 4.3558, 4.8462, 3.2202, 4.4136, 3.9971], + device='cuda:0'), covar=tensor([0.0066, 0.0067, 0.1460, 0.0096, 0.0050, 0.0738, 0.0268, 0.0418], + device='cuda:0'), in_proj_covar=tensor([0.0046, 0.0041, 0.0115, 0.0047, 0.0040, 0.0080, 0.0072, 0.0067], + device='cuda:0'), out_proj_covar=tensor([3.5375e-05, 3.4151e-05, 1.0401e-04, 3.6272e-05, 3.1002e-05, 7.2401e-05, + 6.6161e-05, 6.2928e-05], device='cuda:0') +2023-03-08 14:55:18,947 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4032.0, num_to_drop=1, layers_to_drop={3} +2023-03-08 14:55:22,113 INFO [train.py:898] (0/4) Epoch 2, batch 400, loss[loss=0.3236, simple_loss=0.3848, pruned_loss=0.1312, over 18389.00 frames. ], tot_loss[loss=0.3273, simple_loss=0.3781, pruned_loss=0.1382, over 3123631.08 frames. ], batch size: 52, lr: 4.33e-02, grad_scale: 8.0 +2023-03-08 14:56:19,527 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3663, 3.9278, 4.0261, 3.4710, 3.3097, 2.7450, 3.4718, 2.4998], + device='cuda:0'), covar=tensor([0.0359, 0.0218, 0.0213, 0.0224, 0.0273, 0.0490, 0.0251, 0.0514], + device='cuda:0'), in_proj_covar=tensor([0.0026, 0.0026, 0.0027, 0.0025, 0.0032, 0.0025, 0.0028, 0.0031], + device='cuda:0'), out_proj_covar=tensor([5.2893e-05, 6.0491e-05, 4.7626e-05, 4.9556e-05, 6.1728e-05, 4.9633e-05, + 4.7836e-05, 5.9114e-05], device='cuda:0') +2023-03-08 14:56:22,111 INFO [train.py:898] (0/4) Epoch 2, batch 450, loss[loss=0.3913, simple_loss=0.4303, pruned_loss=0.1762, over 18072.00 frames. ], tot_loss[loss=0.3311, simple_loss=0.3814, pruned_loss=0.1405, over 3217960.09 frames. ], batch size: 62, lr: 4.31e-02, grad_scale: 8.0 +2023-03-08 14:56:44,543 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=2.17 vs. limit=2.0 +2023-03-08 14:56:56,419 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.730e+02 6.143e+02 7.834e+02 1.006e+03 1.697e+03, threshold=1.567e+03, percent-clipped=3.0 +2023-03-08 14:56:57,896 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4693, 3.4211, 2.6335, 3.2430, 3.0667, 3.3883, 2.4575, 2.9440], + device='cuda:0'), covar=tensor([0.0622, 0.0475, 0.0583, 0.0345, 0.0981, 0.0527, 0.0441, 0.0296], + device='cuda:0'), in_proj_covar=tensor([0.0047, 0.0040, 0.0067, 0.0055, 0.0049, 0.0039, 0.0044, 0.0055], + device='cuda:0'), out_proj_covar=tensor([6.0625e-05, 5.4458e-05, 7.5490e-05, 5.6330e-05, 6.2205e-05, 4.6963e-05, + 4.7492e-05, 5.5559e-05], device='cuda:0') +2023-03-08 14:57:21,302 INFO [train.py:898] (0/4) Epoch 2, batch 500, loss[loss=0.2989, simple_loss=0.3488, pruned_loss=0.1245, over 17685.00 frames. ], tot_loss[loss=0.3283, simple_loss=0.3797, pruned_loss=0.1385, over 3306311.25 frames. ], batch size: 39, lr: 4.30e-02, grad_scale: 8.0 +2023-03-08 14:57:28,812 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=4141.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:58:19,878 INFO [train.py:898] (0/4) Epoch 2, batch 550, loss[loss=0.3855, simple_loss=0.4262, pruned_loss=0.1724, over 18284.00 frames. ], tot_loss[loss=0.3274, simple_loss=0.3791, pruned_loss=0.1378, over 3376097.39 frames. ], batch size: 57, lr: 4.29e-02, grad_scale: 8.0 +2023-03-08 14:58:30,533 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3540, 5.1355, 5.2851, 4.7596, 4.8151, 5.0552, 4.6433, 4.8598], + device='cuda:0'), covar=tensor([0.0382, 0.0403, 0.0191, 0.0286, 0.0483, 0.0249, 0.0571, 0.0321], + device='cuda:0'), in_proj_covar=tensor([0.0080, 0.0097, 0.0083, 0.0079, 0.0091, 0.0095, 0.0107, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-08 14:58:42,082 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4202.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 14:58:42,182 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=4202.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 14:58:53,157 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6615, 4.5944, 3.2437, 4.2785, 4.6992, 2.6985, 4.1374, 4.0015], + device='cuda:0'), covar=tensor([0.0093, 0.0160, 0.1567, 0.0087, 0.0059, 0.1043, 0.0357, 0.0496], + device='cuda:0'), in_proj_covar=tensor([0.0049, 0.0043, 0.0118, 0.0049, 0.0042, 0.0084, 0.0075, 0.0071], + device='cuda:0'), out_proj_covar=tensor([3.8799e-05, 3.6694e-05, 1.0599e-04, 3.9042e-05, 3.4116e-05, 7.6720e-05, + 7.0309e-05, 6.8096e-05], device='cuda:0') +2023-03-08 14:58:56,104 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.937e+02 6.401e+02 7.607e+02 9.685e+02 1.732e+03, threshold=1.521e+03, percent-clipped=4.0 +2023-03-08 14:59:20,552 INFO [train.py:898] (0/4) Epoch 2, batch 600, loss[loss=0.3158, simple_loss=0.3638, pruned_loss=0.134, over 18246.00 frames. ], tot_loss[loss=0.326, simple_loss=0.3781, pruned_loss=0.1369, over 3428799.29 frames. ], batch size: 45, lr: 4.28e-02, grad_scale: 8.0 +2023-03-08 14:59:39,658 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=4250.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 14:59:57,914 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5865, 4.3645, 4.7097, 4.5267, 4.3971, 4.3834, 4.8794, 4.7494], + device='cuda:0'), covar=tensor([0.0121, 0.0239, 0.0158, 0.0149, 0.0190, 0.0142, 0.0148, 0.0148], + device='cuda:0'), in_proj_covar=tensor([0.0058, 0.0050, 0.0049, 0.0053, 0.0054, 0.0060, 0.0052, 0.0049], + device='cuda:0'), out_proj_covar=tensor([1.0549e-04, 7.7935e-05, 7.6159e-05, 7.6658e-05, 9.2329e-05, 1.0306e-04, + 8.1486e-05, 7.3238e-05], device='cuda:0') +2023-03-08 15:00:17,454 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4282.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:00:20,601 INFO [train.py:898] (0/4) Epoch 2, batch 650, loss[loss=0.3157, simple_loss=0.3725, pruned_loss=0.1294, over 18377.00 frames. ], tot_loss[loss=0.3255, simple_loss=0.3774, pruned_loss=0.1368, over 3465125.95 frames. ], batch size: 50, lr: 4.27e-02, grad_scale: 8.0 +2023-03-08 15:00:24,404 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=4288.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:00:44,554 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5862, 5.3838, 5.4420, 5.3507, 5.4093, 5.9460, 5.4389, 5.3689], + device='cuda:0'), covar=tensor([0.0444, 0.0497, 0.0538, 0.0425, 0.0919, 0.0430, 0.0533, 0.1261], + device='cuda:0'), in_proj_covar=tensor([0.0157, 0.0136, 0.0126, 0.0117, 0.0169, 0.0162, 0.0115, 0.0176], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002], + device='cuda:0') +2023-03-08 15:00:55,826 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.352e+02 6.066e+02 7.398e+02 9.036e+02 1.375e+03, threshold=1.480e+03, percent-clipped=0.0 +2023-03-08 15:00:59,227 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.92 vs. limit=5.0 +2023-03-08 15:01:02,581 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7413, 3.4807, 3.7234, 2.7628, 3.5280, 3.5033, 3.8253, 3.9463], + device='cuda:0'), covar=tensor([0.0512, 0.0508, 0.0293, 0.1062, 0.1470, 0.0152, 0.0317, 0.0316], + device='cuda:0'), in_proj_covar=tensor([0.0041, 0.0042, 0.0032, 0.0041, 0.0069, 0.0033, 0.0036, 0.0036], + device='cuda:0'), out_proj_covar=tensor([2.2098e-05, 2.5309e-05, 1.7659e-05, 2.5371e-05, 4.5493e-05, 1.6745e-05, + 1.8512e-05, 1.9495e-05], device='cuda:0') +2023-03-08 15:01:16,747 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.06 vs. limit=2.0 +2023-03-08 15:01:17,493 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4332.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 15:01:20,635 INFO [train.py:898] (0/4) Epoch 2, batch 700, loss[loss=0.3373, simple_loss=0.3901, pruned_loss=0.1423, over 18614.00 frames. ], tot_loss[loss=0.3226, simple_loss=0.375, pruned_loss=0.1351, over 3504855.42 frames. ], batch size: 52, lr: 4.26e-02, grad_scale: 8.0 +2023-03-08 15:01:37,694 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=4349.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:02:14,426 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=4380.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 15:02:19,779 INFO [train.py:898] (0/4) Epoch 2, batch 750, loss[loss=0.3119, simple_loss=0.3671, pruned_loss=0.1284, over 18598.00 frames. ], tot_loss[loss=0.3227, simple_loss=0.3749, pruned_loss=0.1352, over 3526046.41 frames. ], batch size: 54, lr: 4.25e-02, grad_scale: 8.0 +2023-03-08 15:02:36,968 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-08 15:02:42,813 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9637, 4.0239, 4.4574, 4.1317, 3.7040, 4.5254, 4.6859, 4.5464], + device='cuda:0'), covar=tensor([0.0144, 0.0341, 0.0095, 0.0340, 0.1324, 0.0074, 0.0101, 0.0135], + device='cuda:0'), in_proj_covar=tensor([0.0044, 0.0045, 0.0035, 0.0045, 0.0077, 0.0035, 0.0039, 0.0040], + device='cuda:0'), out_proj_covar=tensor([2.4324e-05, 2.7704e-05, 1.9278e-05, 2.7839e-05, 5.0439e-05, 1.8264e-05, + 2.0461e-05, 2.1715e-05], device='cuda:0') +2023-03-08 15:02:54,752 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.989e+02 6.069e+02 8.253e+02 1.039e+03 2.142e+03, threshold=1.651e+03, percent-clipped=4.0 +2023-03-08 15:03:19,452 INFO [train.py:898] (0/4) Epoch 2, batch 800, loss[loss=0.3263, simple_loss=0.3825, pruned_loss=0.135, over 18098.00 frames. ], tot_loss[loss=0.3234, simple_loss=0.3753, pruned_loss=0.1358, over 3545543.80 frames. ], batch size: 62, lr: 4.24e-02, grad_scale: 8.0 +2023-03-08 15:03:53,490 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1596, 5.5318, 5.4152, 5.1860, 4.8099, 5.4600, 5.5831, 5.4249], + device='cuda:0'), covar=tensor([0.0821, 0.0647, 0.0325, 0.0581, 0.1620, 0.0447, 0.0510, 0.0527], + device='cuda:0'), in_proj_covar=tensor([0.0200, 0.0199, 0.0148, 0.0194, 0.0244, 0.0177, 0.0183, 0.0168], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 15:03:59,490 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7392, 3.2108, 1.1572, 3.8637, 2.7947, 4.1504, 1.5339, 3.4241], + device='cuda:0'), covar=tensor([0.0137, 0.0848, 0.2167, 0.0310, 0.0652, 0.0084, 0.1784, 0.0335], + device='cuda:0'), in_proj_covar=tensor([0.0044, 0.0074, 0.0076, 0.0037, 0.0062, 0.0041, 0.0073, 0.0059], + device='cuda:0'), out_proj_covar=tensor([3.7185e-05, 7.9438e-05, 8.5004e-05, 4.3067e-05, 5.9145e-05, 3.4850e-05, + 7.3923e-05, 5.5666e-05], device='cuda:0') +2023-03-08 15:04:08,603 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.81 vs. limit=2.0 +2023-03-08 15:04:19,436 INFO [train.py:898] (0/4) Epoch 2, batch 850, loss[loss=0.3005, simple_loss=0.3742, pruned_loss=0.1134, over 18622.00 frames. ], tot_loss[loss=0.3206, simple_loss=0.3734, pruned_loss=0.1339, over 3557332.23 frames. ], batch size: 52, lr: 4.23e-02, grad_scale: 8.0 +2023-03-08 15:04:33,165 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4497.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:04:53,484 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.476e+02 6.096e+02 7.477e+02 9.236e+02 1.546e+03, threshold=1.495e+03, percent-clipped=0.0 +2023-03-08 15:05:18,309 INFO [train.py:898] (0/4) Epoch 2, batch 900, loss[loss=0.3236, simple_loss=0.3813, pruned_loss=0.133, over 18350.00 frames. ], tot_loss[loss=0.3198, simple_loss=0.3731, pruned_loss=0.1333, over 3564854.30 frames. ], batch size: 56, lr: 4.22e-02, grad_scale: 8.0 +2023-03-08 15:06:14,749 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4582.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:06:17,754 INFO [train.py:898] (0/4) Epoch 2, batch 950, loss[loss=0.3169, simple_loss=0.3711, pruned_loss=0.1314, over 18623.00 frames. ], tot_loss[loss=0.3205, simple_loss=0.3737, pruned_loss=0.1337, over 3572160.45 frames. ], batch size: 52, lr: 4.21e-02, grad_scale: 8.0 +2023-03-08 15:06:52,447 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.369e+02 6.106e+02 7.676e+02 9.311e+02 1.838e+03, threshold=1.535e+03, percent-clipped=6.0 +2023-03-08 15:07:12,164 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=4630.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:07:18,131 INFO [train.py:898] (0/4) Epoch 2, batch 1000, loss[loss=0.289, simple_loss=0.3511, pruned_loss=0.1134, over 18281.00 frames. ], tot_loss[loss=0.3187, simple_loss=0.3722, pruned_loss=0.1326, over 3573086.69 frames. ], batch size: 49, lr: 4.20e-02, grad_scale: 8.0 +2023-03-08 15:07:28,735 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=4644.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:07:36,078 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7398, 3.7815, 4.2098, 3.2515, 3.8827, 4.1104, 4.1391, 3.7644], + device='cuda:0'), covar=tensor([0.0566, 0.0444, 0.0187, 0.0853, 0.1482, 0.0094, 0.0258, 0.0311], + device='cuda:0'), in_proj_covar=tensor([0.0049, 0.0049, 0.0038, 0.0051, 0.0088, 0.0038, 0.0042, 0.0044], + device='cuda:0'), out_proj_covar=tensor([2.8454e-05, 3.0610e-05, 2.2047e-05, 3.2092e-05, 5.8293e-05, 2.0752e-05, + 2.3259e-05, 2.5599e-05], device='cuda:0') +2023-03-08 15:07:39,389 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3732, 2.7664, 1.9324, 3.6284, 2.9002, 3.0694, 2.8057, 3.1431], + device='cuda:0'), covar=tensor([0.0332, 0.0344, 0.1827, 0.0101, 0.1083, 0.0323, 0.0322, 0.0351], + device='cuda:0'), in_proj_covar=tensor([0.0048, 0.0043, 0.0042, 0.0040, 0.0039, 0.0053, 0.0053, 0.0037], + device='cuda:0'), out_proj_covar=tensor([5.0197e-05, 4.7523e-05, 5.6703e-05, 3.9369e-05, 4.6052e-05, 5.5855e-05, + 5.5035e-05, 5.0859e-05], device='cuda:0') +2023-03-08 15:07:42,100 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.82 vs. limit=5.0 +2023-03-08 15:08:17,970 INFO [train.py:898] (0/4) Epoch 2, batch 1050, loss[loss=0.3025, simple_loss=0.3686, pruned_loss=0.1182, over 18615.00 frames. ], tot_loss[loss=0.3176, simple_loss=0.3716, pruned_loss=0.1318, over 3569672.51 frames. ], batch size: 52, lr: 4.19e-02, grad_scale: 8.0 +2023-03-08 15:08:37,951 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4687, 6.0118, 5.6615, 5.6073, 5.2375, 5.8583, 6.0230, 5.9763], + device='cuda:0'), covar=tensor([0.0634, 0.0441, 0.0256, 0.0474, 0.1386, 0.0372, 0.0385, 0.0359], + device='cuda:0'), in_proj_covar=tensor([0.0214, 0.0204, 0.0155, 0.0205, 0.0270, 0.0187, 0.0195, 0.0175], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 15:08:52,051 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.788e+02 5.479e+02 6.722e+02 8.127e+02 1.317e+03, threshold=1.344e+03, percent-clipped=0.0 +2023-03-08 15:08:57,406 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.75 vs. limit=5.0 +2023-03-08 15:09:17,312 INFO [train.py:898] (0/4) Epoch 2, batch 1100, loss[loss=0.3692, simple_loss=0.4054, pruned_loss=0.1665, over 13026.00 frames. ], tot_loss[loss=0.3162, simple_loss=0.3706, pruned_loss=0.131, over 3562265.75 frames. ], batch size: 132, lr: 4.18e-02, grad_scale: 4.0 +2023-03-08 15:10:17,510 INFO [train.py:898] (0/4) Epoch 2, batch 1150, loss[loss=0.3143, simple_loss=0.3592, pruned_loss=0.1347, over 18149.00 frames. ], tot_loss[loss=0.3159, simple_loss=0.3705, pruned_loss=0.1307, over 3572429.18 frames. ], batch size: 44, lr: 4.17e-02, grad_scale: 4.0 +2023-03-08 15:10:31,929 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4797.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:10:47,905 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.6256, 5.3491, 5.5100, 5.3488, 5.4637, 5.9880, 5.5575, 5.5624], + device='cuda:0'), covar=tensor([0.0443, 0.0447, 0.0525, 0.0455, 0.0783, 0.0406, 0.0456, 0.0828], + device='cuda:0'), in_proj_covar=tensor([0.0159, 0.0143, 0.0134, 0.0123, 0.0173, 0.0169, 0.0124, 0.0176], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002], + device='cuda:0') +2023-03-08 15:10:52,310 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.612e+02 6.045e+02 7.965e+02 9.549e+02 1.896e+03, threshold=1.593e+03, percent-clipped=3.0 +2023-03-08 15:11:11,465 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.66 vs. limit=2.0 +2023-03-08 15:11:13,664 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5176, 4.6024, 4.5805, 4.6355, 4.5838, 4.2987, 5.0596, 4.8305], + device='cuda:0'), covar=tensor([0.0130, 0.0153, 0.0200, 0.0121, 0.0146, 0.0151, 0.0108, 0.0168], + device='cuda:0'), in_proj_covar=tensor([0.0060, 0.0051, 0.0049, 0.0053, 0.0055, 0.0061, 0.0051, 0.0050], + device='cuda:0'), out_proj_covar=tensor([1.2201e-04, 9.0005e-05, 8.7867e-05, 8.7781e-05, 1.0531e-04, 1.2301e-04, + 9.1554e-05, 8.7371e-05], device='cuda:0') +2023-03-08 15:11:16,899 INFO [train.py:898] (0/4) Epoch 2, batch 1200, loss[loss=0.2933, simple_loss=0.3483, pruned_loss=0.1191, over 18275.00 frames. ], tot_loss[loss=0.3172, simple_loss=0.3714, pruned_loss=0.1315, over 3580710.37 frames. ], batch size: 47, lr: 4.16e-02, grad_scale: 8.0 +2023-03-08 15:11:29,238 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=4845.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:12:16,084 INFO [train.py:898] (0/4) Epoch 2, batch 1250, loss[loss=0.2956, simple_loss=0.3688, pruned_loss=0.1112, over 18353.00 frames. ], tot_loss[loss=0.3142, simple_loss=0.3695, pruned_loss=0.1294, over 3595191.27 frames. ], batch size: 55, lr: 4.15e-02, grad_scale: 8.0 +2023-03-08 15:12:51,967 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.909e+02 5.792e+02 6.888e+02 8.674e+02 1.521e+03, threshold=1.378e+03, percent-clipped=0.0 +2023-03-08 15:13:15,903 INFO [train.py:898] (0/4) Epoch 2, batch 1300, loss[loss=0.332, simple_loss=0.3958, pruned_loss=0.1341, over 18022.00 frames. ], tot_loss[loss=0.3134, simple_loss=0.3687, pruned_loss=0.1291, over 3598027.68 frames. ], batch size: 65, lr: 4.14e-02, grad_scale: 8.0 +2023-03-08 15:13:26,166 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2619, 3.5906, 4.0415, 3.3055, 3.7707, 3.8071, 3.7936, 2.7937], + device='cuda:0'), covar=tensor([0.0474, 0.0233, 0.0084, 0.0315, 0.0130, 0.0341, 0.0193, 0.0597], + device='cuda:0'), in_proj_covar=tensor([0.0045, 0.0039, 0.0029, 0.0039, 0.0027, 0.0050, 0.0026, 0.0045], + device='cuda:0'), out_proj_covar=tensor([3.3109e-05, 2.8780e-05, 2.0061e-05, 2.8512e-05, 2.0874e-05, 3.7645e-05, + 2.0735e-05, 3.2512e-05], device='cuda:0') +2023-03-08 15:13:27,100 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=4944.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:13:57,543 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.83 vs. limit=5.0 +2023-03-08 15:14:15,427 INFO [train.py:898] (0/4) Epoch 2, batch 1350, loss[loss=0.2885, simple_loss=0.337, pruned_loss=0.12, over 18373.00 frames. ], tot_loss[loss=0.3131, simple_loss=0.3682, pruned_loss=0.129, over 3595924.17 frames. ], batch size: 42, lr: 4.13e-02, grad_scale: 8.0 +2023-03-08 15:14:24,037 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=4992.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:14:36,463 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6280, 5.0345, 3.4730, 4.5811, 4.6925, 4.9882, 4.8939, 2.9872], + device='cuda:0'), covar=tensor([0.0240, 0.0046, 0.0258, 0.0084, 0.0062, 0.0084, 0.0095, 0.0583], + device='cuda:0'), in_proj_covar=tensor([0.0037, 0.0022, 0.0032, 0.0025, 0.0028, 0.0025, 0.0028, 0.0042], + device='cuda:0'), out_proj_covar=tensor([8.8728e-05, 5.2329e-05, 7.9389e-05, 6.6324e-05, 5.9799e-05, 5.7790e-05, + 6.4990e-05, 9.2966e-05], device='cuda:0') +2023-03-08 15:14:51,637 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.370e+02 5.891e+02 6.846e+02 8.415e+02 1.418e+03, threshold=1.369e+03, percent-clipped=1.0 +2023-03-08 15:14:57,793 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6799, 3.2252, 3.3245, 2.8992, 2.7433, 2.4384, 2.8519, 2.5976], + device='cuda:0'), covar=tensor([0.0246, 0.0163, 0.0093, 0.0158, 0.0305, 0.0326, 0.0196, 0.0224], + device='cuda:0'), in_proj_covar=tensor([0.0026, 0.0025, 0.0026, 0.0024, 0.0034, 0.0024, 0.0028, 0.0029], + device='cuda:0'), out_proj_covar=tensor([6.9583e-05, 7.8909e-05, 6.0328e-05, 6.3130e-05, 9.1946e-05, 6.3380e-05, + 6.5976e-05, 7.3436e-05], device='cuda:0') +2023-03-08 15:15:10,680 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.84 vs. limit=2.0 +2023-03-08 15:15:15,343 INFO [train.py:898] (0/4) Epoch 2, batch 1400, loss[loss=0.3136, simple_loss=0.3683, pruned_loss=0.1295, over 18374.00 frames. ], tot_loss[loss=0.3122, simple_loss=0.3676, pruned_loss=0.1284, over 3590023.80 frames. ], batch size: 50, lr: 4.12e-02, grad_scale: 8.0 +2023-03-08 15:16:14,508 INFO [train.py:898] (0/4) Epoch 2, batch 1450, loss[loss=0.33, simple_loss=0.384, pruned_loss=0.138, over 17948.00 frames. ], tot_loss[loss=0.3119, simple_loss=0.3675, pruned_loss=0.1282, over 3595683.77 frames. ], batch size: 65, lr: 4.11e-02, grad_scale: 8.0 +2023-03-08 15:16:50,484 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.583e+02 6.040e+02 7.064e+02 8.729e+02 1.298e+03, threshold=1.413e+03, percent-clipped=0.0 +2023-03-08 15:17:13,678 INFO [train.py:898] (0/4) Epoch 2, batch 1500, loss[loss=0.2944, simple_loss=0.3505, pruned_loss=0.1191, over 18540.00 frames. ], tot_loss[loss=0.312, simple_loss=0.3675, pruned_loss=0.1283, over 3585124.29 frames. ], batch size: 49, lr: 4.10e-02, grad_scale: 8.0 +2023-03-08 15:17:17,245 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0009, 4.9100, 4.2349, 5.0841, 4.9717, 4.5901, 4.8612, 4.2522], + device='cuda:0'), covar=tensor([0.0294, 0.0292, 0.1591, 0.0367, 0.0294, 0.0334, 0.0311, 0.0568], + device='cuda:0'), in_proj_covar=tensor([0.0163, 0.0174, 0.0278, 0.0140, 0.0150, 0.0179, 0.0176, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0002, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 15:17:32,636 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.4597, 2.4980, 2.1500, 2.3649, 2.4070, 2.0969, 2.1156, 2.5060], + device='cuda:0'), covar=tensor([0.0224, 0.0144, 0.0462, 0.0177, 0.0347, 0.0310, 0.0335, 0.0276], + device='cuda:0'), in_proj_covar=tensor([0.0045, 0.0041, 0.0045, 0.0042, 0.0040, 0.0053, 0.0059, 0.0039], + device='cuda:0'), out_proj_covar=tensor([5.2233e-05, 4.7102e-05, 6.2729e-05, 4.6238e-05, 5.1783e-05, 6.0912e-05, + 6.5620e-05, 5.5045e-05], device='cuda:0') +2023-03-08 15:18:11,879 INFO [train.py:898] (0/4) Epoch 2, batch 1550, loss[loss=0.3187, simple_loss=0.3673, pruned_loss=0.135, over 18397.00 frames. ], tot_loss[loss=0.3104, simple_loss=0.3661, pruned_loss=0.1274, over 3591449.46 frames. ], batch size: 52, lr: 4.08e-02, grad_scale: 8.0 +2023-03-08 15:18:48,959 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.611e+02 6.188e+02 7.756e+02 9.428e+02 1.707e+03, threshold=1.551e+03, percent-clipped=5.0 +2023-03-08 15:18:49,224 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4227, 5.9900, 5.6925, 5.7695, 5.2948, 5.9259, 6.0919, 5.9227], + device='cuda:0'), covar=tensor([0.1034, 0.0637, 0.0292, 0.0582, 0.1695, 0.0424, 0.0451, 0.0632], + device='cuda:0'), in_proj_covar=tensor([0.0239, 0.0211, 0.0165, 0.0215, 0.0297, 0.0204, 0.0203, 0.0189], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0004, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 15:19:04,264 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4846, 5.9714, 5.6435, 5.7210, 5.3586, 5.8326, 6.0276, 5.9268], + device='cuda:0'), covar=tensor([0.0841, 0.0495, 0.0270, 0.0456, 0.1432, 0.0381, 0.0384, 0.0464], + device='cuda:0'), in_proj_covar=tensor([0.0236, 0.0207, 0.0162, 0.0211, 0.0290, 0.0204, 0.0200, 0.0186], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0004, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 15:19:11,984 INFO [train.py:898] (0/4) Epoch 2, batch 1600, loss[loss=0.2814, simple_loss=0.3368, pruned_loss=0.113, over 17591.00 frames. ], tot_loss[loss=0.3091, simple_loss=0.3652, pruned_loss=0.1265, over 3600756.99 frames. ], batch size: 39, lr: 4.07e-02, grad_scale: 8.0 +2023-03-08 15:20:11,418 INFO [train.py:898] (0/4) Epoch 2, batch 1650, loss[loss=0.2594, simple_loss=0.3359, pruned_loss=0.09143, over 18281.00 frames. ], tot_loss[loss=0.3072, simple_loss=0.3637, pruned_loss=0.1253, over 3608561.89 frames. ], batch size: 49, lr: 4.06e-02, grad_scale: 8.0 +2023-03-08 15:20:23,770 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([1.8097, 4.1796, 3.9705, 3.6668, 2.5759, 2.1964, 3.7534, 4.0803], + device='cuda:0'), covar=tensor([0.0925, 0.0206, 0.0078, 0.0163, 0.0961, 0.1240, 0.0202, 0.0042], + device='cuda:0'), in_proj_covar=tensor([0.0071, 0.0036, 0.0032, 0.0051, 0.0080, 0.0087, 0.0052, 0.0030], + device='cuda:0'), out_proj_covar=tensor([9.8730e-05, 5.8930e-05, 4.1875e-05, 6.7203e-05, 1.0215e-04, 1.1192e-04, + 6.9638e-05, 3.9375e-05], device='cuda:0') +2023-03-08 15:20:35,554 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3503, 3.9466, 4.1721, 3.6796, 3.0823, 2.5000, 3.0270, 2.3795], + device='cuda:0'), covar=tensor([0.0247, 0.0249, 0.0128, 0.0138, 0.0368, 0.0483, 0.0258, 0.0426], + device='cuda:0'), in_proj_covar=tensor([0.0025, 0.0025, 0.0025, 0.0023, 0.0034, 0.0023, 0.0028, 0.0030], + device='cuda:0'), out_proj_covar=tensor([7.3655e-05, 8.4978e-05, 6.2672e-05, 6.7036e-05, 1.0027e-04, 6.7505e-05, + 7.1082e-05, 8.0773e-05], device='cuda:0') +2023-03-08 15:20:47,169 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.457e+02 5.814e+02 6.706e+02 8.624e+02 1.400e+03, threshold=1.341e+03, percent-clipped=0.0 +2023-03-08 15:21:10,521 INFO [train.py:898] (0/4) Epoch 2, batch 1700, loss[loss=0.364, simple_loss=0.3945, pruned_loss=0.1667, over 11924.00 frames. ], tot_loss[loss=0.3074, simple_loss=0.3639, pruned_loss=0.1254, over 3597174.68 frames. ], batch size: 129, lr: 4.05e-02, grad_scale: 8.0 +2023-03-08 15:21:36,434 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=2.06 vs. limit=2.0 +2023-03-08 15:22:10,169 INFO [train.py:898] (0/4) Epoch 2, batch 1750, loss[loss=0.2704, simple_loss=0.3376, pruned_loss=0.1016, over 18521.00 frames. ], tot_loss[loss=0.3056, simple_loss=0.3628, pruned_loss=0.1242, over 3606539.60 frames. ], batch size: 47, lr: 4.04e-02, grad_scale: 8.0 +2023-03-08 15:22:34,933 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7436, 3.0133, 1.5456, 3.7779, 2.9449, 4.3864, 2.0601, 3.7387], + device='cuda:0'), covar=tensor([0.0244, 0.1072, 0.1984, 0.0349, 0.0836, 0.0071, 0.1597, 0.0334], + device='cuda:0'), in_proj_covar=tensor([0.0063, 0.0100, 0.0100, 0.0052, 0.0088, 0.0050, 0.0097, 0.0085], + device='cuda:0'), out_proj_covar=tensor([6.4787e-05, 1.0896e-04, 1.0981e-04, 6.9580e-05, 9.4218e-05, 4.7487e-05, + 1.0115e-04, 8.5256e-05], device='cuda:0') +2023-03-08 15:22:46,216 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.642e+02 5.505e+02 6.727e+02 8.573e+02 1.772e+03, threshold=1.345e+03, percent-clipped=4.0 +2023-03-08 15:22:54,199 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7644, 3.5680, 3.9562, 2.9652, 3.8144, 3.8952, 3.8492, 3.8609], + device='cuda:0'), covar=tensor([0.0419, 0.0425, 0.0122, 0.0713, 0.1291, 0.0081, 0.0292, 0.0279], + device='cuda:0'), in_proj_covar=tensor([0.0065, 0.0064, 0.0047, 0.0067, 0.0120, 0.0049, 0.0054, 0.0059], + device='cuda:0'), out_proj_covar=tensor([4.2686e-05, 4.2765e-05, 3.0299e-05, 4.5906e-05, 8.3759e-05, 2.9922e-05, + 3.3434e-05, 3.8454e-05], device='cuda:0') +2023-03-08 15:23:02,900 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.96 vs. limit=2.0 +2023-03-08 15:23:10,300 INFO [train.py:898] (0/4) Epoch 2, batch 1800, loss[loss=0.2841, simple_loss=0.351, pruned_loss=0.1086, over 18543.00 frames. ], tot_loss[loss=0.3062, simple_loss=0.3635, pruned_loss=0.1245, over 3604331.79 frames. ], batch size: 49, lr: 4.03e-02, grad_scale: 8.0 +2023-03-08 15:23:11,157 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.70 vs. limit=2.0 +2023-03-08 15:24:09,615 INFO [train.py:898] (0/4) Epoch 2, batch 1850, loss[loss=0.403, simple_loss=0.4203, pruned_loss=0.1929, over 12534.00 frames. ], tot_loss[loss=0.3075, simple_loss=0.3642, pruned_loss=0.1254, over 3587734.61 frames. ], batch size: 130, lr: 4.02e-02, grad_scale: 8.0 +2023-03-08 15:24:45,051 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.702e+02 5.666e+02 6.741e+02 8.273e+02 2.014e+03, threshold=1.348e+03, percent-clipped=2.0 +2023-03-08 15:25:08,313 INFO [train.py:898] (0/4) Epoch 2, batch 1900, loss[loss=0.2858, simple_loss=0.3431, pruned_loss=0.1142, over 18354.00 frames. ], tot_loss[loss=0.3071, simple_loss=0.3634, pruned_loss=0.1254, over 3574194.69 frames. ], batch size: 46, lr: 4.01e-02, grad_scale: 8.0 +2023-03-08 15:26:07,461 INFO [train.py:898] (0/4) Epoch 2, batch 1950, loss[loss=0.3014, simple_loss=0.3659, pruned_loss=0.1184, over 18107.00 frames. ], tot_loss[loss=0.306, simple_loss=0.363, pruned_loss=0.1245, over 3582214.43 frames. ], batch size: 62, lr: 4.00e-02, grad_scale: 8.0 +2023-03-08 15:26:42,648 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.871e+02 5.827e+02 7.380e+02 9.024e+02 1.685e+03, threshold=1.476e+03, percent-clipped=2.0 +2023-03-08 15:27:06,886 INFO [train.py:898] (0/4) Epoch 2, batch 2000, loss[loss=0.2892, simple_loss=0.3533, pruned_loss=0.1126, over 18407.00 frames. ], tot_loss[loss=0.3048, simple_loss=0.3616, pruned_loss=0.1239, over 3582986.80 frames. ], batch size: 48, lr: 3.99e-02, grad_scale: 8.0 +2023-03-08 15:27:24,002 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5650.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 15:28:05,536 INFO [train.py:898] (0/4) Epoch 2, batch 2050, loss[loss=0.3115, simple_loss=0.3685, pruned_loss=0.1272, over 18224.00 frames. ], tot_loss[loss=0.3049, simple_loss=0.3618, pruned_loss=0.124, over 3575653.62 frames. ], batch size: 60, lr: 3.98e-02, grad_scale: 8.0 +2023-03-08 15:28:35,678 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5711.0, num_to_drop=1, layers_to_drop={3} +2023-03-08 15:28:42,463 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.372e+02 5.884e+02 6.946e+02 9.338e+02 2.315e+03, threshold=1.389e+03, percent-clipped=7.0 +2023-03-08 15:29:03,506 INFO [train.py:898] (0/4) Epoch 2, batch 2100, loss[loss=0.3372, simple_loss=0.3891, pruned_loss=0.1427, over 17999.00 frames. ], tot_loss[loss=0.3045, simple_loss=0.3614, pruned_loss=0.1238, over 3585211.46 frames. ], batch size: 65, lr: 3.97e-02, grad_scale: 2.0 +2023-03-08 15:29:03,722 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4892, 5.2342, 5.4206, 5.2096, 5.2301, 6.0458, 5.5697, 5.5299], + device='cuda:0'), covar=tensor([0.0562, 0.0553, 0.0515, 0.0501, 0.0985, 0.0512, 0.0492, 0.0981], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0139, 0.0132, 0.0125, 0.0177, 0.0182, 0.0123, 0.0181], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002], + device='cuda:0') +2023-03-08 15:29:27,826 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5755.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:30:02,847 INFO [train.py:898] (0/4) Epoch 2, batch 2150, loss[loss=0.2706, simple_loss=0.3236, pruned_loss=0.1088, over 18493.00 frames. ], tot_loss[loss=0.3035, simple_loss=0.3607, pruned_loss=0.1232, over 3590924.84 frames. ], batch size: 44, lr: 3.96e-02, grad_scale: 2.0 +2023-03-08 15:30:40,445 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5816.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:30:41,159 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.344e+02 5.674e+02 6.514e+02 8.678e+02 1.455e+03, threshold=1.303e+03, percent-clipped=1.0 +2023-03-08 15:30:41,503 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0553, 2.9557, 3.1546, 2.9964, 2.2455, 3.2009, 3.1431, 2.4112], + device='cuda:0'), covar=tensor([0.0121, 0.0108, 0.0072, 0.0101, 0.0668, 0.0061, 0.0087, 0.0517], + device='cuda:0'), in_proj_covar=tensor([0.0053, 0.0049, 0.0047, 0.0044, 0.0089, 0.0044, 0.0038, 0.0085], + device='cuda:0'), out_proj_covar=tensor([5.2085e-05, 4.6240e-05, 4.6100e-05, 4.5312e-05, 8.4852e-05, 4.0193e-05, + 4.3150e-05, 8.3161e-05], device='cuda:0') +2023-03-08 15:31:01,994 INFO [train.py:898] (0/4) Epoch 2, batch 2200, loss[loss=0.2989, simple_loss=0.3628, pruned_loss=0.1175, over 18506.00 frames. ], tot_loss[loss=0.3039, simple_loss=0.3612, pruned_loss=0.1233, over 3587389.30 frames. ], batch size: 51, lr: 3.95e-02, grad_scale: 2.0 +2023-03-08 15:31:29,438 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5858.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:31:56,695 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.82 vs. limit=2.0 +2023-03-08 15:32:00,586 INFO [train.py:898] (0/4) Epoch 2, batch 2250, loss[loss=0.317, simple_loss=0.3751, pruned_loss=0.1294, over 18035.00 frames. ], tot_loss[loss=0.3033, simple_loss=0.3607, pruned_loss=0.123, over 3576829.63 frames. ], batch size: 65, lr: 3.95e-02, grad_scale: 2.0 +2023-03-08 15:32:09,118 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1098, 3.0475, 3.1712, 2.9916, 2.0524, 3.2966, 2.9155, 2.3183], + device='cuda:0'), covar=tensor([0.0098, 0.0115, 0.0085, 0.0084, 0.0765, 0.0073, 0.0145, 0.0593], + device='cuda:0'), in_proj_covar=tensor([0.0055, 0.0051, 0.0048, 0.0045, 0.0092, 0.0045, 0.0039, 0.0089], + device='cuda:0'), out_proj_covar=tensor([5.3849e-05, 4.7498e-05, 4.6752e-05, 4.5972e-05, 8.8334e-05, 4.0961e-05, + 4.3928e-05, 8.6826e-05], device='cuda:0') +2023-03-08 15:32:38,782 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.697e+02 5.715e+02 7.226e+02 9.109e+02 2.021e+03, threshold=1.445e+03, percent-clipped=5.0 +2023-03-08 15:32:41,288 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=5919.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:32:59,623 INFO [train.py:898] (0/4) Epoch 2, batch 2300, loss[loss=0.2611, simple_loss=0.3208, pruned_loss=0.1007, over 18255.00 frames. ], tot_loss[loss=0.305, simple_loss=0.3623, pruned_loss=0.1239, over 3571072.56 frames. ], batch size: 45, lr: 3.94e-02, grad_scale: 2.0 +2023-03-08 15:33:58,221 INFO [train.py:898] (0/4) Epoch 2, batch 2350, loss[loss=0.3272, simple_loss=0.3843, pruned_loss=0.135, over 18339.00 frames. ], tot_loss[loss=0.3033, simple_loss=0.3611, pruned_loss=0.1227, over 3579180.39 frames. ], batch size: 56, lr: 3.93e-02, grad_scale: 2.0 +2023-03-08 15:34:04,378 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=5990.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:34:05,471 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([1.5387, 4.1494, 4.2266, 3.4358, 2.4327, 2.1419, 3.7196, 4.2185], + device='cuda:0'), covar=tensor([0.1186, 0.0165, 0.0049, 0.0233, 0.0985, 0.1180, 0.0237, 0.0040], + device='cuda:0'), in_proj_covar=tensor([0.0083, 0.0039, 0.0035, 0.0058, 0.0090, 0.0099, 0.0059, 0.0032], + device='cuda:0'), out_proj_covar=tensor([1.1926e-04, 6.7776e-05, 5.0520e-05, 8.2127e-05, 1.2139e-04, 1.3365e-04, + 8.2822e-05, 4.5191e-05], device='cuda:0') +2023-03-08 15:34:15,181 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-6000.pt +2023-03-08 15:34:27,061 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6006.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 15:34:39,270 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.355e+02 5.403e+02 7.312e+02 8.931e+02 1.402e+03, threshold=1.462e+03, percent-clipped=0.0 +2023-03-08 15:35:00,406 INFO [train.py:898] (0/4) Epoch 2, batch 2400, loss[loss=0.3347, simple_loss=0.3845, pruned_loss=0.1424, over 17984.00 frames. ], tot_loss[loss=0.3043, simple_loss=0.3617, pruned_loss=0.1234, over 3569429.39 frames. ], batch size: 65, lr: 3.92e-02, grad_scale: 4.0 +2023-03-08 15:35:10,100 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6043.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:35:19,738 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6051.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:35:59,379 INFO [train.py:898] (0/4) Epoch 2, batch 2450, loss[loss=0.3207, simple_loss=0.3855, pruned_loss=0.128, over 18368.00 frames. ], tot_loss[loss=0.3033, simple_loss=0.3609, pruned_loss=0.1229, over 3571457.24 frames. ], batch size: 56, lr: 3.91e-02, grad_scale: 4.0 +2023-03-08 15:36:23,042 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6104.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:36:30,634 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6111.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:36:37,237 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.501e+02 5.639e+02 7.028e+02 8.696e+02 2.308e+03, threshold=1.406e+03, percent-clipped=2.0 +2023-03-08 15:36:39,783 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6053, 4.5667, 4.7421, 4.6773, 4.5166, 4.3637, 4.9105, 4.9111], + device='cuda:0'), covar=tensor([0.0118, 0.0132, 0.0149, 0.0112, 0.0136, 0.0155, 0.0144, 0.0124], + device='cuda:0'), in_proj_covar=tensor([0.0060, 0.0049, 0.0046, 0.0054, 0.0052, 0.0061, 0.0054, 0.0046], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0001, 0.0001, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:0') +2023-03-08 15:36:58,327 INFO [train.py:898] (0/4) Epoch 2, batch 2500, loss[loss=0.2778, simple_loss=0.3397, pruned_loss=0.1079, over 18541.00 frames. ], tot_loss[loss=0.3023, simple_loss=0.36, pruned_loss=0.1223, over 3571194.01 frames. ], batch size: 49, lr: 3.90e-02, grad_scale: 4.0 +2023-03-08 15:36:58,679 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6135.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:36:59,298 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-03-08 15:37:32,602 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=5.37 vs. limit=5.0 +2023-03-08 15:37:39,720 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6170.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:37:56,479 INFO [train.py:898] (0/4) Epoch 2, batch 2550, loss[loss=0.3311, simple_loss=0.3848, pruned_loss=0.1387, over 16243.00 frames. ], tot_loss[loss=0.3015, simple_loss=0.3596, pruned_loss=0.1217, over 3580453.45 frames. ], batch size: 94, lr: 3.89e-02, grad_scale: 4.0 +2023-03-08 15:38:10,468 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6196.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:38:18,770 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6203.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:38:28,961 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6211.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:38:32,258 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6214.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:38:35,433 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.309e+02 5.777e+02 7.159e+02 8.638e+02 1.810e+03, threshold=1.432e+03, percent-clipped=8.0 +2023-03-08 15:38:36,364 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.66 vs. limit=2.0 +2023-03-08 15:38:37,546 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-08 15:38:51,721 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6231.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:38:55,993 INFO [train.py:898] (0/4) Epoch 2, batch 2600, loss[loss=0.2883, simple_loss=0.3595, pruned_loss=0.1086, over 18545.00 frames. ], tot_loss[loss=0.299, simple_loss=0.3575, pruned_loss=0.1202, over 3582746.80 frames. ], batch size: 54, lr: 3.88e-02, grad_scale: 4.0 +2023-03-08 15:39:31,881 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6264.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:39:34,038 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6266.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:39:40,773 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6272.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:39:55,317 INFO [train.py:898] (0/4) Epoch 2, batch 2650, loss[loss=0.2973, simple_loss=0.3562, pruned_loss=0.1192, over 18484.00 frames. ], tot_loss[loss=0.2988, simple_loss=0.3572, pruned_loss=0.1202, over 3589819.90 frames. ], batch size: 51, lr: 3.87e-02, grad_scale: 4.0 +2023-03-08 15:39:59,499 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-03-08 15:40:20,588 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6306.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 15:40:33,295 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.735e+02 5.474e+02 6.730e+02 8.945e+02 2.130e+03, threshold=1.346e+03, percent-clipped=7.0 +2023-03-08 15:40:45,715 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6327.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:40:48,163 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3267, 4.2187, 4.4094, 4.1500, 2.2925, 4.3344, 3.6855, 3.1286], + device='cuda:0'), covar=tensor([0.0089, 0.0088, 0.0079, 0.0098, 0.1056, 0.0059, 0.0099, 0.0586], + device='cuda:0'), in_proj_covar=tensor([0.0058, 0.0055, 0.0051, 0.0048, 0.0102, 0.0049, 0.0043, 0.0098], + device='cuda:0'), out_proj_covar=tensor([5.6753e-05, 5.2352e-05, 5.1239e-05, 5.0023e-05, 9.7563e-05, 4.5434e-05, + 4.8340e-05, 9.5542e-05], device='cuda:0') +2023-03-08 15:40:54,509 INFO [train.py:898] (0/4) Epoch 2, batch 2700, loss[loss=0.2944, simple_loss=0.3509, pruned_loss=0.1189, over 18258.00 frames. ], tot_loss[loss=0.3004, simple_loss=0.3586, pruned_loss=0.1211, over 3588689.39 frames. ], batch size: 47, lr: 3.86e-02, grad_scale: 4.0 +2023-03-08 15:41:07,625 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6346.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:41:17,125 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=6354.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 15:41:31,315 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9739, 3.0935, 4.4070, 2.6874, 3.4890, 4.1094, 4.3739, 4.1040], + device='cuda:0'), covar=tensor([0.0232, 0.0378, 0.0097, 0.0595, 0.1063, 0.0041, 0.0126, 0.0193], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0078, 0.0050, 0.0081, 0.0145, 0.0055, 0.0068, 0.0071], + device='cuda:0'), out_proj_covar=tensor([5.3566e-05, 5.7201e-05, 3.7788e-05, 6.0194e-05, 1.0709e-04, 3.6124e-05, + 4.7130e-05, 5.1948e-05], device='cuda:0') +2023-03-08 15:41:53,012 INFO [train.py:898] (0/4) Epoch 2, batch 2750, loss[loss=0.3344, simple_loss=0.397, pruned_loss=0.1359, over 18288.00 frames. ], tot_loss[loss=0.3006, simple_loss=0.3585, pruned_loss=0.1214, over 3584907.55 frames. ], batch size: 57, lr: 3.85e-02, grad_scale: 4.0 +2023-03-08 15:42:09,978 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6399.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:42:24,421 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6411.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:42:31,296 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.682e+02 5.887e+02 7.195e+02 9.085e+02 1.925e+03, threshold=1.439e+03, percent-clipped=3.0 +2023-03-08 15:42:52,440 INFO [train.py:898] (0/4) Epoch 2, batch 2800, loss[loss=0.3098, simple_loss=0.3767, pruned_loss=0.1215, over 18317.00 frames. ], tot_loss[loss=0.2995, simple_loss=0.3583, pruned_loss=0.1204, over 3586721.03 frames. ], batch size: 54, lr: 3.84e-02, grad_scale: 8.0 +2023-03-08 15:43:21,373 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=6459.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:43:25,109 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6462.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:43:30,876 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6467.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:43:51,662 INFO [train.py:898] (0/4) Epoch 2, batch 2850, loss[loss=0.2974, simple_loss=0.3571, pruned_loss=0.1189, over 18499.00 frames. ], tot_loss[loss=0.2992, simple_loss=0.358, pruned_loss=0.1202, over 3582010.61 frames. ], batch size: 47, lr: 3.83e-02, grad_scale: 8.0 +2023-03-08 15:43:58,770 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6491.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:44:26,094 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6514.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:44:29,244 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.462e+02 5.681e+02 6.833e+02 8.085e+02 2.080e+03, threshold=1.367e+03, percent-clipped=2.0 +2023-03-08 15:44:35,060 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=5.36 vs. limit=5.0 +2023-03-08 15:44:37,071 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6523.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 15:44:40,203 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6526.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:44:43,177 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6528.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:44:50,782 INFO [train.py:898] (0/4) Epoch 2, batch 2900, loss[loss=0.3175, simple_loss=0.3733, pruned_loss=0.1309, over 18247.00 frames. ], tot_loss[loss=0.2984, simple_loss=0.3574, pruned_loss=0.1197, over 3586013.19 frames. ], batch size: 60, lr: 3.82e-02, grad_scale: 8.0 +2023-03-08 15:44:51,336 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-08 15:44:56,289 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.00 vs. limit=2.0 +2023-03-08 15:45:09,564 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1481, 3.6713, 4.3874, 3.2642, 3.7035, 3.6322, 4.0336, 2.8376], + device='cuda:0'), covar=tensor([0.0470, 0.0293, 0.0084, 0.0330, 0.0205, 0.0693, 0.0400, 0.0748], + device='cuda:0'), in_proj_covar=tensor([0.0071, 0.0064, 0.0043, 0.0062, 0.0054, 0.0096, 0.0043, 0.0082], + device='cuda:0'), out_proj_covar=tensor([6.4120e-05, 5.7602e-05, 3.6361e-05, 5.7551e-05, 5.1672e-05, 8.4423e-05, + 4.4886e-05, 6.9883e-05], device='cuda:0') +2023-03-08 15:45:19,372 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6559.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:45:22,845 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=6562.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:45:28,528 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6567.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:45:49,893 INFO [train.py:898] (0/4) Epoch 2, batch 2950, loss[loss=0.2557, simple_loss=0.3223, pruned_loss=0.09458, over 18286.00 frames. ], tot_loss[loss=0.2965, simple_loss=0.3559, pruned_loss=0.1185, over 3593370.44 frames. ], batch size: 49, lr: 3.81e-02, grad_scale: 8.0 +2023-03-08 15:46:27,984 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.426e+02 5.603e+02 6.658e+02 8.574e+02 1.720e+03, threshold=1.332e+03, percent-clipped=3.0 +2023-03-08 15:46:33,935 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6622.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:46:42,588 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6629.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:46:45,896 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5922, 4.0575, 4.0513, 3.7381, 3.7359, 3.9494, 3.6507, 3.8877], + device='cuda:0'), covar=tensor([0.0335, 0.0311, 0.0250, 0.0298, 0.0526, 0.0237, 0.0691, 0.0319], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0117, 0.0105, 0.0091, 0.0112, 0.0112, 0.0145, 0.0106], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 15:46:49,586 INFO [train.py:898] (0/4) Epoch 2, batch 3000, loss[loss=0.2835, simple_loss=0.3547, pruned_loss=0.1062, over 18290.00 frames. ], tot_loss[loss=0.297, simple_loss=0.3568, pruned_loss=0.1186, over 3579373.02 frames. ], batch size: 57, lr: 3.80e-02, grad_scale: 8.0 +2023-03-08 15:46:49,588 INFO [train.py:923] (0/4) Computing validation loss +2023-03-08 15:46:57,527 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7412, 2.9646, 3.6010, 2.5993, 3.3420, 3.1387, 3.2325, 2.3522], + device='cuda:0'), covar=tensor([0.0509, 0.0382, 0.0118, 0.0353, 0.0240, 0.0687, 0.0346, 0.0785], + device='cuda:0'), in_proj_covar=tensor([0.0072, 0.0066, 0.0043, 0.0064, 0.0057, 0.0098, 0.0044, 0.0084], + device='cuda:0'), out_proj_covar=tensor([6.5517e-05, 5.9910e-05, 3.7127e-05, 5.8848e-05, 5.4287e-05, 8.6963e-05, + 4.6189e-05, 7.2130e-05], device='cuda:0') +2023-03-08 15:47:01,157 INFO [train.py:932] (0/4) Epoch 2, validation: loss=0.2202, simple_loss=0.3188, pruned_loss=0.06074, over 944034.00 frames. +2023-03-08 15:47:01,158 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19544MB +2023-03-08 15:47:14,425 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6646.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:47:37,792 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2853, 5.0481, 5.2672, 5.3527, 5.1285, 5.8523, 5.2588, 5.2909], + device='cuda:0'), covar=tensor([0.0620, 0.0698, 0.0631, 0.0512, 0.1382, 0.0627, 0.0531, 0.1417], + device='cuda:0'), in_proj_covar=tensor([0.0183, 0.0148, 0.0147, 0.0136, 0.0194, 0.0203, 0.0136, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 15:47:39,075 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8569, 5.3440, 3.5595, 4.6892, 4.8910, 5.3331, 4.9857, 2.8618], + device='cuda:0'), covar=tensor([0.0236, 0.0040, 0.0363, 0.0070, 0.0068, 0.0033, 0.0085, 0.0790], + device='cuda:0'), in_proj_covar=tensor([0.0044, 0.0027, 0.0050, 0.0031, 0.0035, 0.0029, 0.0036, 0.0063], + device='cuda:0'), out_proj_covar=tensor([1.4779e-04, 9.2237e-05, 1.5784e-04, 1.2114e-04, 1.0337e-04, 9.6279e-05, + 1.1401e-04, 1.7535e-04], device='cuda:0') +2023-03-08 15:47:59,747 INFO [train.py:898] (0/4) Epoch 2, batch 3050, loss[loss=0.266, simple_loss=0.337, pruned_loss=0.0975, over 18377.00 frames. ], tot_loss[loss=0.2962, simple_loss=0.3557, pruned_loss=0.1184, over 3591252.11 frames. ], batch size: 50, lr: 3.79e-02, grad_scale: 8.0 +2023-03-08 15:48:05,961 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6690.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:48:10,410 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=6694.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:48:11,600 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2632, 5.0067, 5.2887, 5.0059, 4.8107, 4.9506, 4.4683, 4.8486], + device='cuda:0'), covar=tensor([0.0434, 0.0475, 0.0239, 0.0232, 0.0491, 0.0361, 0.0818, 0.0328], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0118, 0.0103, 0.0090, 0.0110, 0.0114, 0.0146, 0.0106], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 15:48:16,920 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6699.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:48:37,954 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.591e+02 5.592e+02 6.340e+02 8.473e+02 1.907e+03, threshold=1.268e+03, percent-clipped=6.0 +2023-03-08 15:48:55,991 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3714, 3.0370, 1.3392, 4.1541, 2.6413, 4.4010, 2.0754, 3.6919], + device='cuda:0'), covar=tensor([0.0474, 0.1065, 0.2234, 0.0256, 0.1121, 0.0048, 0.1482, 0.0389], + device='cuda:0'), in_proj_covar=tensor([0.0081, 0.0121, 0.0120, 0.0069, 0.0110, 0.0053, 0.0116, 0.0103], + device='cuda:0'), out_proj_covar=tensor([1.0284e-04, 1.3751e-04, 1.3668e-04, 1.0129e-04, 1.3128e-04, 6.0233e-05, + 1.2799e-04, 1.1454e-04], device='cuda:0') +2023-03-08 15:48:58,811 INFO [train.py:898] (0/4) Epoch 2, batch 3100, loss[loss=0.2758, simple_loss=0.3317, pruned_loss=0.1099, over 17712.00 frames. ], tot_loss[loss=0.2961, simple_loss=0.3557, pruned_loss=0.1182, over 3581966.28 frames. ], batch size: 39, lr: 3.79e-02, grad_scale: 8.0 +2023-03-08 15:49:12,509 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=6747.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:49:24,947 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5126, 4.3842, 4.7066, 4.4589, 4.2700, 4.2991, 4.7119, 4.6886], + device='cuda:0'), covar=tensor([0.0091, 0.0137, 0.0075, 0.0099, 0.0131, 0.0129, 0.0091, 0.0092], + device='cuda:0'), in_proj_covar=tensor([0.0059, 0.0048, 0.0045, 0.0053, 0.0050, 0.0059, 0.0052, 0.0044], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0001, 0.0001, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:0') +2023-03-08 15:49:57,575 INFO [train.py:898] (0/4) Epoch 2, batch 3150, loss[loss=0.2968, simple_loss=0.3683, pruned_loss=0.1127, over 18375.00 frames. ], tot_loss[loss=0.2948, simple_loss=0.3547, pruned_loss=0.1174, over 3586806.24 frames. ], batch size: 50, lr: 3.78e-02, grad_scale: 8.0 +2023-03-08 15:50:04,876 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6791.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:50:08,331 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8388, 4.1570, 4.1282, 3.7191, 2.4652, 4.1427, 3.8788, 2.4473], + device='cuda:0'), covar=tensor([0.0159, 0.0109, 0.0072, 0.0151, 0.1070, 0.0072, 0.0093, 0.0877], + device='cuda:0'), in_proj_covar=tensor([0.0065, 0.0060, 0.0056, 0.0054, 0.0110, 0.0055, 0.0049, 0.0106], + device='cuda:0'), out_proj_covar=tensor([6.3216e-05, 5.6844e-05, 5.6296e-05, 5.5420e-05, 1.0623e-04, 5.2270e-05, + 5.5748e-05, 1.0459e-04], device='cuda:0') +2023-03-08 15:50:36,055 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.441e+02 5.682e+02 6.955e+02 9.535e+02 1.991e+03, threshold=1.391e+03, percent-clipped=9.0 +2023-03-08 15:50:37,877 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6818.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 15:50:43,374 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6823.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:50:46,862 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6826.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:50:50,300 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2464, 5.1583, 4.4388, 5.2378, 5.2207, 4.7800, 5.1262, 4.6762], + device='cuda:0'), covar=tensor([0.0286, 0.0312, 0.1891, 0.0400, 0.0246, 0.0335, 0.0308, 0.0430], + device='cuda:0'), in_proj_covar=tensor([0.0188, 0.0207, 0.0338, 0.0163, 0.0165, 0.0199, 0.0205, 0.0228], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 15:50:57,112 INFO [train.py:898] (0/4) Epoch 2, batch 3200, loss[loss=0.2484, simple_loss=0.311, pruned_loss=0.09292, over 18382.00 frames. ], tot_loss[loss=0.2936, simple_loss=0.3536, pruned_loss=0.1168, over 3586337.67 frames. ], batch size: 42, lr: 3.77e-02, grad_scale: 8.0 +2023-03-08 15:50:58,857 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.81 vs. limit=2.0 +2023-03-08 15:51:01,906 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=6839.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:51:25,990 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6859.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:51:35,025 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6867.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:51:43,493 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=6874.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:51:48,450 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6878.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 15:51:56,397 INFO [train.py:898] (0/4) Epoch 2, batch 3250, loss[loss=0.3233, simple_loss=0.3822, pruned_loss=0.1322, over 18045.00 frames. ], tot_loss[loss=0.2932, simple_loss=0.3535, pruned_loss=0.1165, over 3584933.61 frames. ], batch size: 65, lr: 3.76e-02, grad_scale: 8.0 +2023-03-08 15:52:22,315 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=6907.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:52:25,360 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=6909.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:52:29,081 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.76 vs. limit=2.0 +2023-03-08 15:52:31,932 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=6915.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:52:34,070 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.634e+02 5.392e+02 6.712e+02 8.838e+02 2.461e+03, threshold=1.342e+03, percent-clipped=3.0 +2023-03-08 15:52:39,931 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=6922.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:52:54,797 INFO [train.py:898] (0/4) Epoch 2, batch 3300, loss[loss=0.3454, simple_loss=0.3893, pruned_loss=0.1508, over 12951.00 frames. ], tot_loss[loss=0.294, simple_loss=0.3541, pruned_loss=0.117, over 3581725.07 frames. ], batch size: 129, lr: 3.75e-02, grad_scale: 8.0 +2023-03-08 15:53:00,106 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6939.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 15:53:02,225 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0865, 4.8968, 2.2021, 4.8144, 4.9643, 2.3339, 4.3106, 3.6754], + device='cuda:0'), covar=tensor([0.0053, 0.0297, 0.1882, 0.0178, 0.0078, 0.1462, 0.0435, 0.0737], + device='cuda:0'), in_proj_covar=tensor([0.0065, 0.0066, 0.0146, 0.0096, 0.0057, 0.0131, 0.0129, 0.0120], + device='cuda:0'), out_proj_covar=tensor([6.9438e-05, 8.7664e-05, 1.4622e-04, 1.0170e-04, 6.1453e-05, 1.3596e-04, + 1.3788e-04, 1.3541e-04], device='cuda:0') +2023-03-08 15:53:16,969 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3492, 5.3022, 4.6364, 5.4331, 5.3480, 4.8459, 5.3271, 4.7837], + device='cuda:0'), covar=tensor([0.0340, 0.0284, 0.1916, 0.0525, 0.0300, 0.0416, 0.0295, 0.0479], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0207, 0.0342, 0.0167, 0.0167, 0.0201, 0.0213, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 15:53:36,593 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=6970.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:53:36,863 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=6970.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:53:54,223 INFO [train.py:898] (0/4) Epoch 2, batch 3350, loss[loss=0.2814, simple_loss=0.3488, pruned_loss=0.1069, over 18623.00 frames. ], tot_loss[loss=0.2936, simple_loss=0.3539, pruned_loss=0.1167, over 3578483.47 frames. ], batch size: 52, lr: 3.74e-02, grad_scale: 8.0 +2023-03-08 15:53:54,369 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=6985.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:54:32,863 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.449e+02 5.349e+02 6.526e+02 8.040e+02 1.400e+03, threshold=1.305e+03, percent-clipped=1.0 +2023-03-08 15:54:53,424 INFO [train.py:898] (0/4) Epoch 2, batch 3400, loss[loss=0.366, simple_loss=0.4153, pruned_loss=0.1583, over 17972.00 frames. ], tot_loss[loss=0.2941, simple_loss=0.3536, pruned_loss=0.1173, over 3583366.28 frames. ], batch size: 65, lr: 3.73e-02, grad_scale: 8.0 +2023-03-08 15:55:09,221 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7048.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 15:55:52,763 INFO [train.py:898] (0/4) Epoch 2, batch 3450, loss[loss=0.2815, simple_loss=0.3492, pruned_loss=0.1069, over 18509.00 frames. ], tot_loss[loss=0.2931, simple_loss=0.3528, pruned_loss=0.1167, over 3586261.10 frames. ], batch size: 51, lr: 3.72e-02, grad_scale: 8.0 +2023-03-08 15:56:20,882 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7109.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 15:56:30,819 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.505e+02 5.757e+02 6.768e+02 8.417e+02 2.561e+03, threshold=1.354e+03, percent-clipped=6.0 +2023-03-08 15:56:32,039 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7118.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 15:56:37,266 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7123.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:56:51,287 INFO [train.py:898] (0/4) Epoch 2, batch 3500, loss[loss=0.283, simple_loss=0.3356, pruned_loss=0.1152, over 18352.00 frames. ], tot_loss[loss=0.2923, simple_loss=0.3517, pruned_loss=0.1165, over 3570257.86 frames. ], batch size: 46, lr: 3.71e-02, grad_scale: 8.0 +2023-03-08 15:57:11,560 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5906, 4.4611, 4.5990, 4.4715, 4.3325, 4.3923, 4.9848, 4.8539], + device='cuda:0'), covar=tensor([0.0101, 0.0157, 0.0130, 0.0147, 0.0156, 0.0150, 0.0132, 0.0134], + device='cuda:0'), in_proj_covar=tensor([0.0062, 0.0048, 0.0046, 0.0056, 0.0052, 0.0060, 0.0054, 0.0047], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0001, 0.0001, 0.0001, 0.0002, 0.0001, 0.0001], + device='cuda:0') +2023-03-08 15:57:23,525 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7163.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:57:26,593 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=7166.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:57:31,613 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=7171.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:57:46,604 INFO [train.py:898] (0/4) Epoch 2, batch 3550, loss[loss=0.2705, simple_loss=0.3319, pruned_loss=0.1045, over 18546.00 frames. ], tot_loss[loss=0.2922, simple_loss=0.3515, pruned_loss=0.1164, over 3567630.16 frames. ], batch size: 49, lr: 3.71e-02, grad_scale: 8.0 +2023-03-08 15:58:22,323 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.147e+02 5.640e+02 6.762e+02 8.858e+02 1.958e+03, threshold=1.352e+03, percent-clipped=2.0 +2023-03-08 15:58:30,200 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7224.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 15:58:40,531 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7234.0, num_to_drop=1, layers_to_drop={3} +2023-03-08 15:58:41,453 INFO [train.py:898] (0/4) Epoch 2, batch 3600, loss[loss=0.2665, simple_loss=0.328, pruned_loss=0.1024, over 18242.00 frames. ], tot_loss[loss=0.2944, simple_loss=0.3535, pruned_loss=0.1177, over 3568051.27 frames. ], batch size: 45, lr: 3.70e-02, grad_scale: 8.0 +2023-03-08 15:59:06,138 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5461, 4.0038, 4.0669, 3.8189, 3.7870, 3.9877, 3.5545, 3.8638], + device='cuda:0'), covar=tensor([0.0320, 0.0433, 0.0282, 0.0355, 0.0417, 0.0233, 0.0843, 0.0304], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0118, 0.0103, 0.0090, 0.0112, 0.0111, 0.0150, 0.0104], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 15:59:12,881 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7265.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 15:59:16,858 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-2.pt +2023-03-08 15:59:46,465 INFO [train.py:898] (0/4) Epoch 3, batch 0, loss[loss=0.2492, simple_loss=0.3079, pruned_loss=0.09521, over 18435.00 frames. ], tot_loss[loss=0.2492, simple_loss=0.3079, pruned_loss=0.09521, over 18435.00 frames. ], batch size: 43, lr: 3.51e-02, grad_scale: 8.0 +2023-03-08 15:59:46,474 INFO [train.py:923] (0/4) Computing validation loss +2023-03-08 15:59:58,151 INFO [train.py:932] (0/4) Epoch 3, validation: loss=0.2228, simple_loss=0.3215, pruned_loss=0.06204, over 944034.00 frames. +2023-03-08 15:59:58,152 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19544MB +2023-03-08 16:00:05,143 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7275.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:00:16,707 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7285.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:00:44,133 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.04 vs. limit=2.0 +2023-03-08 16:00:48,363 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7760, 4.8832, 4.8818, 4.7411, 4.7394, 5.2782, 4.7762, 4.6912], + device='cuda:0'), covar=tensor([0.0663, 0.0629, 0.0550, 0.0529, 0.1084, 0.0595, 0.0597, 0.1382], + device='cuda:0'), in_proj_covar=tensor([0.0182, 0.0141, 0.0142, 0.0130, 0.0189, 0.0194, 0.0133, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 16:00:54,864 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.526e+02 5.312e+02 6.673e+02 8.366e+02 1.427e+03, threshold=1.335e+03, percent-clipped=1.0 +2023-03-08 16:00:57,226 INFO [train.py:898] (0/4) Epoch 3, batch 50, loss[loss=0.3026, simple_loss=0.3636, pruned_loss=0.1208, over 15801.00 frames. ], tot_loss[loss=0.2895, simple_loss=0.3522, pruned_loss=0.1134, over 813460.28 frames. ], batch size: 94, lr: 3.50e-02, grad_scale: 8.0 +2023-03-08 16:01:14,259 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=7333.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:01:18,007 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7336.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:01:57,066 INFO [train.py:898] (0/4) Epoch 3, batch 100, loss[loss=0.2453, simple_loss=0.3139, pruned_loss=0.08841, over 18501.00 frames. ], tot_loss[loss=0.2868, simple_loss=0.3498, pruned_loss=0.1119, over 1432236.29 frames. ], batch size: 47, lr: 3.49e-02, grad_scale: 8.0 +2023-03-08 16:02:39,675 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7404.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 16:02:43,227 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7407.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:02:54,249 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.382e+02 5.439e+02 6.621e+02 7.725e+02 1.513e+03, threshold=1.324e+03, percent-clipped=3.0 +2023-03-08 16:02:56,567 INFO [train.py:898] (0/4) Epoch 3, batch 150, loss[loss=0.285, simple_loss=0.3521, pruned_loss=0.109, over 18493.00 frames. ], tot_loss[loss=0.2862, simple_loss=0.3489, pruned_loss=0.1118, over 1909609.93 frames. ], batch size: 53, lr: 3.48e-02, grad_scale: 8.0 +2023-03-08 16:03:05,376 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.75 vs. limit=2.0 +2023-03-08 16:03:54,935 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7468.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:03:55,727 INFO [train.py:898] (0/4) Epoch 3, batch 200, loss[loss=0.3238, simple_loss=0.374, pruned_loss=0.1368, over 17091.00 frames. ], tot_loss[loss=0.2861, simple_loss=0.3493, pruned_loss=0.1115, over 2287930.45 frames. ], batch size: 78, lr: 3.47e-02, grad_scale: 8.0 +2023-03-08 16:04:51,796 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.337e+02 5.369e+02 6.599e+02 8.256e+02 1.502e+03, threshold=1.320e+03, percent-clipped=3.0 +2023-03-08 16:04:54,100 INFO [train.py:898] (0/4) Epoch 3, batch 250, loss[loss=0.2789, simple_loss=0.3539, pruned_loss=0.1019, over 18378.00 frames. ], tot_loss[loss=0.2874, simple_loss=0.3504, pruned_loss=0.1122, over 2578041.58 frames. ], batch size: 55, lr: 3.47e-02, grad_scale: 8.0 +2023-03-08 16:04:54,327 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7519.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 16:05:10,707 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7534.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 16:05:47,456 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7565.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:05:51,577 INFO [train.py:898] (0/4) Epoch 3, batch 300, loss[loss=0.2563, simple_loss=0.3155, pruned_loss=0.09859, over 18389.00 frames. ], tot_loss[loss=0.2853, simple_loss=0.3483, pruned_loss=0.1112, over 2803678.26 frames. ], batch size: 42, lr: 3.46e-02, grad_scale: 8.0 +2023-03-08 16:06:06,000 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=7582.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 16:06:20,296 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3075, 4.2167, 4.1443, 3.4200, 3.2288, 2.5055, 1.8347, 1.7809], + device='cuda:0'), covar=tensor([0.0151, 0.0145, 0.0111, 0.0131, 0.0246, 0.0367, 0.0491, 0.0597], + device='cuda:0'), in_proj_covar=tensor([0.0025, 0.0024, 0.0022, 0.0022, 0.0035, 0.0022, 0.0035, 0.0035], + device='cuda:0'), out_proj_covar=tensor([1.1378e-04, 1.1822e-04, 9.0448e-05, 1.0024e-04, 1.5778e-04, 9.7587e-05, + 1.4036e-04, 1.4428e-04], device='cuda:0') +2023-03-08 16:06:26,061 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3296, 4.3678, 4.4233, 3.6085, 3.2046, 2.6828, 1.9848, 1.7563], + device='cuda:0'), covar=tensor([0.0156, 0.0129, 0.0068, 0.0120, 0.0289, 0.0288, 0.0505, 0.0604], + device='cuda:0'), in_proj_covar=tensor([0.0025, 0.0023, 0.0022, 0.0022, 0.0034, 0.0021, 0.0035, 0.0035], + device='cuda:0'), out_proj_covar=tensor([1.1280e-04, 1.1693e-04, 8.9859e-05, 9.9131e-05, 1.5661e-04, 9.6534e-05, + 1.3920e-04, 1.4328e-04], device='cuda:0') +2023-03-08 16:06:40,126 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9809, 3.7835, 2.4591, 3.7699, 4.0304, 2.5239, 3.3562, 3.3648], + device='cuda:0'), covar=tensor([0.0085, 0.0452, 0.1293, 0.0249, 0.0083, 0.1193, 0.0538, 0.0614], + device='cuda:0'), in_proj_covar=tensor([0.0070, 0.0080, 0.0150, 0.0109, 0.0062, 0.0143, 0.0138, 0.0135], + device='cuda:0'), out_proj_covar=tensor([7.7911e-05, 1.0863e-04, 1.5681e-04, 1.1794e-04, 6.9283e-05, 1.5306e-04, + 1.5179e-04, 1.5558e-04], device='cuda:0') +2023-03-08 16:06:43,886 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=7613.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:06:48,187 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.523e+02 5.233e+02 6.862e+02 8.770e+02 1.930e+03, threshold=1.372e+03, percent-clipped=4.0 +2023-03-08 16:06:50,538 INFO [train.py:898] (0/4) Epoch 3, batch 350, loss[loss=0.2712, simple_loss=0.3324, pruned_loss=0.105, over 18251.00 frames. ], tot_loss[loss=0.2858, simple_loss=0.3485, pruned_loss=0.1116, over 2974175.66 frames. ], batch size: 45, lr: 3.45e-02, grad_scale: 8.0 +2023-03-08 16:06:53,132 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0770, 4.9541, 4.2973, 5.0955, 4.9342, 4.4992, 4.8859, 4.4529], + device='cuda:0'), covar=tensor([0.0379, 0.0453, 0.1930, 0.0472, 0.0472, 0.0415, 0.0390, 0.0596], + device='cuda:0'), in_proj_covar=tensor([0.0206, 0.0227, 0.0363, 0.0177, 0.0175, 0.0208, 0.0222, 0.0256], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0005, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 16:06:56,745 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7624.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:07:04,384 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7631.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:07:49,574 INFO [train.py:898] (0/4) Epoch 3, batch 400, loss[loss=0.2403, simple_loss=0.3098, pruned_loss=0.08537, over 17693.00 frames. ], tot_loss[loss=0.2839, simple_loss=0.347, pruned_loss=0.1103, over 3107248.10 frames. ], batch size: 39, lr: 3.44e-02, grad_scale: 8.0 +2023-03-08 16:08:07,799 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7685.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:08:29,756 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7704.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 16:08:47,268 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.778e+02 4.774e+02 5.790e+02 7.165e+02 2.375e+03, threshold=1.158e+03, percent-clipped=2.0 +2023-03-08 16:08:48,397 INFO [train.py:898] (0/4) Epoch 3, batch 450, loss[loss=0.2422, simple_loss=0.3129, pruned_loss=0.08578, over 18146.00 frames. ], tot_loss[loss=0.2841, simple_loss=0.3472, pruned_loss=0.1106, over 3218645.74 frames. ], batch size: 44, lr: 3.44e-02, grad_scale: 8.0 +2023-03-08 16:09:15,811 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7743.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:09:26,347 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=7752.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 16:09:38,520 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7763.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:09:46,279 INFO [train.py:898] (0/4) Epoch 3, batch 500, loss[loss=0.2956, simple_loss=0.3594, pruned_loss=0.1159, over 18261.00 frames. ], tot_loss[loss=0.284, simple_loss=0.347, pruned_loss=0.1105, over 3307417.99 frames. ], batch size: 60, lr: 3.43e-02, grad_scale: 8.0 +2023-03-08 16:10:20,036 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-08 16:10:26,080 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=7802.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:10:28,414 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7804.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:10:45,272 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.544e+02 5.001e+02 6.338e+02 8.345e+02 2.616e+03, threshold=1.268e+03, percent-clipped=10.0 +2023-03-08 16:10:46,496 INFO [train.py:898] (0/4) Epoch 3, batch 550, loss[loss=0.2771, simple_loss=0.3428, pruned_loss=0.1057, over 18537.00 frames. ], tot_loss[loss=0.2835, simple_loss=0.3467, pruned_loss=0.1102, over 3362361.09 frames. ], batch size: 49, lr: 3.42e-02, grad_scale: 8.0 +2023-03-08 16:10:46,757 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7819.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 16:11:06,353 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9718, 2.3375, 4.4707, 2.9062, 3.4509, 4.3689, 4.3013, 3.9368], + device='cuda:0'), covar=tensor([0.0237, 0.0619, 0.0102, 0.0559, 0.1269, 0.0039, 0.0171, 0.0187], + device='cuda:0'), in_proj_covar=tensor([0.0085, 0.0098, 0.0055, 0.0102, 0.0179, 0.0066, 0.0080, 0.0086], + device='cuda:0'), out_proj_covar=tensor([6.9208e-05, 7.8597e-05, 4.4868e-05, 8.0572e-05, 1.3998e-04, 4.4784e-05, + 6.3190e-05, 6.7923e-05], device='cuda:0') +2023-03-08 16:11:37,626 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=7863.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:11:41,786 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=7867.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:11:44,717 INFO [train.py:898] (0/4) Epoch 3, batch 600, loss[loss=0.2772, simple_loss=0.3293, pruned_loss=0.1125, over 17632.00 frames. ], tot_loss[loss=0.2822, simple_loss=0.3451, pruned_loss=0.1097, over 3407402.14 frames. ], batch size: 39, lr: 3.41e-02, grad_scale: 8.0 +2023-03-08 16:12:32,237 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1532, 4.9767, 5.1106, 4.6199, 4.6626, 4.8436, 4.3371, 4.7094], + device='cuda:0'), covar=tensor([0.0318, 0.0332, 0.0177, 0.0280, 0.0422, 0.0240, 0.0848, 0.0317], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0122, 0.0105, 0.0092, 0.0119, 0.0117, 0.0160, 0.0108], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-08 16:12:42,165 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.737e+02 5.313e+02 6.628e+02 8.433e+02 2.088e+03, threshold=1.326e+03, percent-clipped=7.0 +2023-03-08 16:12:43,820 INFO [train.py:898] (0/4) Epoch 3, batch 650, loss[loss=0.3347, simple_loss=0.3881, pruned_loss=0.1407, over 15883.00 frames. ], tot_loss[loss=0.2843, simple_loss=0.3468, pruned_loss=0.1109, over 3435191.60 frames. ], batch size: 94, lr: 3.40e-02, grad_scale: 8.0 +2023-03-08 16:12:59,168 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=7931.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:13:42,825 INFO [train.py:898] (0/4) Epoch 3, batch 700, loss[loss=0.3045, simple_loss=0.3679, pruned_loss=0.1206, over 18450.00 frames. ], tot_loss[loss=0.2831, simple_loss=0.3457, pruned_loss=0.1102, over 3475652.53 frames. ], batch size: 59, lr: 3.40e-02, grad_scale: 8.0 +2023-03-08 16:13:55,625 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=7979.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:13:57,341 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=7980.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:14:20,036 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-8000.pt +2023-03-08 16:14:45,127 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.136e+02 5.602e+02 6.674e+02 7.907e+02 1.811e+03, threshold=1.335e+03, percent-clipped=1.0 +2023-03-08 16:14:46,260 INFO [train.py:898] (0/4) Epoch 3, batch 750, loss[loss=0.262, simple_loss=0.3309, pruned_loss=0.09651, over 18261.00 frames. ], tot_loss[loss=0.2832, simple_loss=0.3461, pruned_loss=0.1101, over 3503271.51 frames. ], batch size: 45, lr: 3.39e-02, grad_scale: 8.0 +2023-03-08 16:15:33,413 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2655, 5.2643, 4.4620, 5.2792, 5.2813, 4.7878, 5.1857, 4.6181], + device='cuda:0'), covar=tensor([0.0353, 0.0339, 0.2029, 0.0448, 0.0274, 0.0339, 0.0313, 0.0589], + device='cuda:0'), in_proj_covar=tensor([0.0211, 0.0230, 0.0372, 0.0180, 0.0173, 0.0204, 0.0225, 0.0255], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0005, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 16:15:38,401 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8063.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:15:44,999 INFO [train.py:898] (0/4) Epoch 3, batch 800, loss[loss=0.307, simple_loss=0.3697, pruned_loss=0.1222, over 18298.00 frames. ], tot_loss[loss=0.2837, simple_loss=0.3465, pruned_loss=0.1105, over 3520584.48 frames. ], batch size: 57, lr: 3.38e-02, grad_scale: 8.0 +2023-03-08 16:16:21,781 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=8099.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:16:35,604 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=8111.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:16:43,308 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.074e+02 5.395e+02 6.687e+02 8.203e+02 1.547e+03, threshold=1.337e+03, percent-clipped=3.0 +2023-03-08 16:16:44,974 INFO [train.py:898] (0/4) Epoch 3, batch 850, loss[loss=0.2555, simple_loss=0.3145, pruned_loss=0.09823, over 18427.00 frames. ], tot_loss[loss=0.2819, simple_loss=0.3452, pruned_loss=0.1093, over 3543052.94 frames. ], batch size: 43, lr: 3.37e-02, grad_scale: 8.0 +2023-03-08 16:16:59,226 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-08 16:17:10,694 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-08 16:17:31,430 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=8158.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:17:43,831 INFO [train.py:898] (0/4) Epoch 3, batch 900, loss[loss=0.2543, simple_loss=0.3176, pruned_loss=0.09552, over 18150.00 frames. ], tot_loss[loss=0.2826, simple_loss=0.3458, pruned_loss=0.1097, over 3547597.03 frames. ], batch size: 44, lr: 3.37e-02, grad_scale: 8.0 +2023-03-08 16:18:42,377 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.447e+02 5.230e+02 6.588e+02 8.350e+02 1.625e+03, threshold=1.318e+03, percent-clipped=3.0 +2023-03-08 16:18:43,442 INFO [train.py:898] (0/4) Epoch 3, batch 950, loss[loss=0.2804, simple_loss=0.3465, pruned_loss=0.1071, over 18397.00 frames. ], tot_loss[loss=0.2829, simple_loss=0.3465, pruned_loss=0.1096, over 3553997.80 frames. ], batch size: 50, lr: 3.36e-02, grad_scale: 8.0 +2023-03-08 16:18:55,815 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-03-08 16:19:38,248 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8265.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:19:42,215 INFO [train.py:898] (0/4) Epoch 3, batch 1000, loss[loss=0.2627, simple_loss=0.3249, pruned_loss=0.1002, over 18279.00 frames. ], tot_loss[loss=0.2825, simple_loss=0.3465, pruned_loss=0.1092, over 3574975.62 frames. ], batch size: 49, lr: 3.35e-02, grad_scale: 8.0 +2023-03-08 16:19:55,431 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8280.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:19:57,151 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-03-08 16:20:09,612 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-03-08 16:20:40,390 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.475e+02 5.034e+02 6.200e+02 8.356e+02 1.633e+03, threshold=1.240e+03, percent-clipped=2.0 +2023-03-08 16:20:41,571 INFO [train.py:898] (0/4) Epoch 3, batch 1050, loss[loss=0.2499, simple_loss=0.3185, pruned_loss=0.09069, over 18360.00 frames. ], tot_loss[loss=0.281, simple_loss=0.3452, pruned_loss=0.1084, over 3567260.80 frames. ], batch size: 46, lr: 3.34e-02, grad_scale: 8.0 +2023-03-08 16:20:49,578 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=8326.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:20:51,627 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=8328.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:21:40,128 INFO [train.py:898] (0/4) Epoch 3, batch 1100, loss[loss=0.2714, simple_loss=0.3407, pruned_loss=0.101, over 18304.00 frames. ], tot_loss[loss=0.2801, simple_loss=0.3445, pruned_loss=0.1079, over 3579080.36 frames. ], batch size: 57, lr: 3.34e-02, grad_scale: 8.0 +2023-03-08 16:21:42,834 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0747, 4.6044, 1.9436, 4.8257, 4.9880, 2.3265, 4.1623, 3.6194], + device='cuda:0'), covar=tensor([0.0080, 0.0733, 0.1902, 0.0225, 0.0057, 0.1632, 0.0539, 0.0925], + device='cuda:0'), in_proj_covar=tensor([0.0071, 0.0088, 0.0154, 0.0118, 0.0059, 0.0145, 0.0148, 0.0138], + device='cuda:0'), out_proj_covar=tensor([8.1426e-05, 1.2246e-04, 1.6624e-04, 1.3368e-04, 7.0793e-05, 1.6212e-04, + 1.6656e-04, 1.6396e-04], device='cuda:0') +2023-03-08 16:21:58,730 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.60 vs. limit=2.0 +2023-03-08 16:22:16,056 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8399.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:22:16,371 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-08 16:22:38,112 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.304e+02 5.764e+02 6.854e+02 8.468e+02 1.709e+03, threshold=1.371e+03, percent-clipped=7.0 +2023-03-08 16:22:39,110 INFO [train.py:898] (0/4) Epoch 3, batch 1150, loss[loss=0.2739, simple_loss=0.3238, pruned_loss=0.112, over 18395.00 frames. ], tot_loss[loss=0.2819, simple_loss=0.3459, pruned_loss=0.109, over 3570251.64 frames. ], batch size: 42, lr: 3.33e-02, grad_scale: 8.0 +2023-03-08 16:23:12,003 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=8447.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:23:15,257 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-03-08 16:23:26,454 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8458.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:23:38,491 INFO [train.py:898] (0/4) Epoch 3, batch 1200, loss[loss=0.2843, simple_loss=0.3555, pruned_loss=0.1066, over 18101.00 frames. ], tot_loss[loss=0.2816, simple_loss=0.3452, pruned_loss=0.109, over 3565622.81 frames. ], batch size: 62, lr: 3.32e-02, grad_scale: 8.0 +2023-03-08 16:24:22,270 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=8506.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:24:36,203 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.117e+02 5.075e+02 6.913e+02 8.856e+02 3.555e+03, threshold=1.383e+03, percent-clipped=10.0 +2023-03-08 16:24:37,335 INFO [train.py:898] (0/4) Epoch 3, batch 1250, loss[loss=0.2796, simple_loss=0.3533, pruned_loss=0.103, over 18350.00 frames. ], tot_loss[loss=0.2806, simple_loss=0.3443, pruned_loss=0.1084, over 3568540.34 frames. ], batch size: 55, lr: 3.31e-02, grad_scale: 8.0 +2023-03-08 16:25:36,293 INFO [train.py:898] (0/4) Epoch 3, batch 1300, loss[loss=0.2421, simple_loss=0.3108, pruned_loss=0.08667, over 18414.00 frames. ], tot_loss[loss=0.2804, simple_loss=0.3444, pruned_loss=0.1082, over 3567977.94 frames. ], batch size: 48, lr: 3.31e-02, grad_scale: 8.0 +2023-03-08 16:26:01,860 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2311, 5.9245, 5.4876, 5.6621, 5.2603, 5.6767, 6.0172, 5.8663], + device='cuda:0'), covar=tensor([0.1030, 0.0488, 0.0314, 0.0596, 0.1440, 0.0493, 0.0353, 0.0498], + device='cuda:0'), in_proj_covar=tensor([0.0308, 0.0257, 0.0204, 0.0274, 0.0384, 0.0274, 0.0271, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-08 16:26:03,012 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3762, 5.1103, 5.1585, 4.8551, 4.8720, 4.8681, 4.4171, 4.8006], + device='cuda:0'), covar=tensor([0.0308, 0.0270, 0.0169, 0.0195, 0.0348, 0.0216, 0.0949, 0.0267], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0126, 0.0107, 0.0096, 0.0120, 0.0122, 0.0168, 0.0112], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-08 16:26:35,143 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.185e+02 5.027e+02 6.070e+02 7.698e+02 1.470e+03, threshold=1.214e+03, percent-clipped=1.0 +2023-03-08 16:26:36,239 INFO [train.py:898] (0/4) Epoch 3, batch 1350, loss[loss=0.2482, simple_loss=0.3097, pruned_loss=0.09333, over 18140.00 frames. ], tot_loss[loss=0.2803, simple_loss=0.3444, pruned_loss=0.1081, over 3558951.24 frames. ], batch size: 44, lr: 3.30e-02, grad_scale: 8.0 +2023-03-08 16:26:38,814 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=8621.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:27:35,129 INFO [train.py:898] (0/4) Epoch 3, batch 1400, loss[loss=0.303, simple_loss=0.3682, pruned_loss=0.1189, over 17944.00 frames. ], tot_loss[loss=0.2794, simple_loss=0.3432, pruned_loss=0.1077, over 3559443.64 frames. ], batch size: 65, lr: 3.29e-02, grad_scale: 8.0 +2023-03-08 16:28:25,612 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.01 vs. limit=2.0 +2023-03-08 16:28:32,214 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.853e+02 5.972e+02 7.488e+02 8.927e+02 1.917e+03, threshold=1.498e+03, percent-clipped=2.0 +2023-03-08 16:28:33,370 INFO [train.py:898] (0/4) Epoch 3, batch 1450, loss[loss=0.2489, simple_loss=0.3148, pruned_loss=0.09149, over 18287.00 frames. ], tot_loss[loss=0.2793, simple_loss=0.3432, pruned_loss=0.1077, over 3559492.25 frames. ], batch size: 45, lr: 3.29e-02, grad_scale: 8.0 +2023-03-08 16:28:34,889 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9078, 2.3211, 2.2033, 2.5537, 2.7627, 2.9102, 2.2177, 2.4946], + device='cuda:0'), covar=tensor([0.0262, 0.0278, 0.0963, 0.0396, 0.0381, 0.0255, 0.0585, 0.0265], + device='cuda:0'), in_proj_covar=tensor([0.0066, 0.0053, 0.0104, 0.0075, 0.0067, 0.0049, 0.0070, 0.0061], + device='cuda:0'), out_proj_covar=tensor([1.1201e-04, 9.2884e-05, 1.6772e-04, 1.1899e-04, 1.1379e-04, 7.8782e-05, + 1.1389e-04, 9.5388e-05], device='cuda:0') +2023-03-08 16:29:21,380 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8760.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:29:32,457 INFO [train.py:898] (0/4) Epoch 3, batch 1500, loss[loss=0.2122, simple_loss=0.2835, pruned_loss=0.07046, over 18380.00 frames. ], tot_loss[loss=0.2761, simple_loss=0.3408, pruned_loss=0.1057, over 3580113.98 frames. ], batch size: 42, lr: 3.28e-02, grad_scale: 8.0 +2023-03-08 16:30:15,268 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.27 vs. limit=5.0 +2023-03-08 16:30:29,509 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.789e+02 5.215e+02 6.562e+02 8.622e+02 2.061e+03, threshold=1.312e+03, percent-clipped=5.0 +2023-03-08 16:30:30,742 INFO [train.py:898] (0/4) Epoch 3, batch 1550, loss[loss=0.2601, simple_loss=0.3294, pruned_loss=0.09546, over 18406.00 frames. ], tot_loss[loss=0.2761, simple_loss=0.3405, pruned_loss=0.1059, over 3586089.13 frames. ], batch size: 48, lr: 3.27e-02, grad_scale: 8.0 +2023-03-08 16:30:32,288 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5108, 2.9149, 2.5512, 2.9511, 3.0326, 3.4140, 2.5669, 3.0795], + device='cuda:0'), covar=tensor([0.0524, 0.0297, 0.0997, 0.0632, 0.0598, 0.0232, 0.0555, 0.0214], + device='cuda:0'), in_proj_covar=tensor([0.0070, 0.0052, 0.0108, 0.0077, 0.0068, 0.0049, 0.0072, 0.0062], + device='cuda:0'), out_proj_covar=tensor([1.1952e-04, 9.1552e-05, 1.7391e-04, 1.2100e-04, 1.1557e-04, 7.9061e-05, + 1.1814e-04, 9.7290e-05], device='cuda:0') +2023-03-08 16:30:34,036 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=8821.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:30:36,398 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0528, 2.9334, 2.3553, 2.2345, 2.7141, 2.2447, 2.0819, 2.9340], + device='cuda:0'), covar=tensor([0.0096, 0.0129, 0.0348, 0.0199, 0.0190, 0.0270, 0.0360, 0.0204], + device='cuda:0'), in_proj_covar=tensor([0.0040, 0.0044, 0.0048, 0.0058, 0.0042, 0.0064, 0.0077, 0.0044], + device='cuda:0'), out_proj_covar=tensor([5.8355e-05, 6.5177e-05, 7.5869e-05, 8.3897e-05, 6.3544e-05, 9.8854e-05, + 1.2251e-04, 7.0045e-05], device='cuda:0') +2023-03-08 16:31:25,901 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.69 vs. limit=5.0 +2023-03-08 16:31:28,683 INFO [train.py:898] (0/4) Epoch 3, batch 1600, loss[loss=0.2923, simple_loss=0.3516, pruned_loss=0.1165, over 18103.00 frames. ], tot_loss[loss=0.2757, simple_loss=0.3401, pruned_loss=0.1057, over 3592817.94 frames. ], batch size: 62, lr: 3.26e-02, grad_scale: 8.0 +2023-03-08 16:31:58,996 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2303, 5.0346, 5.0639, 4.7863, 4.7281, 4.8377, 4.3540, 4.8397], + device='cuda:0'), covar=tensor([0.0378, 0.0334, 0.0212, 0.0259, 0.0416, 0.0244, 0.0971, 0.0280], + device='cuda:0'), in_proj_covar=tensor([0.0097, 0.0133, 0.0113, 0.0100, 0.0128, 0.0128, 0.0175, 0.0114], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003], + device='cuda:0') +2023-03-08 16:32:26,433 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.256e+02 5.119e+02 6.262e+02 8.326e+02 2.097e+03, threshold=1.252e+03, percent-clipped=4.0 +2023-03-08 16:32:27,487 INFO [train.py:898] (0/4) Epoch 3, batch 1650, loss[loss=0.2665, simple_loss=0.3176, pruned_loss=0.1077, over 18493.00 frames. ], tot_loss[loss=0.2767, simple_loss=0.3411, pruned_loss=0.1061, over 3583788.04 frames. ], batch size: 44, lr: 3.26e-02, grad_scale: 8.0 +2023-03-08 16:32:30,135 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=8921.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:32:37,650 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8927.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:32:47,253 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-08 16:33:23,674 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0750, 4.0918, 3.6483, 4.0719, 4.0862, 3.6783, 4.0412, 3.5717], + device='cuda:0'), covar=tensor([0.0337, 0.0447, 0.1323, 0.0443, 0.0325, 0.0409, 0.0353, 0.0663], + device='cuda:0'), in_proj_covar=tensor([0.0216, 0.0237, 0.0375, 0.0191, 0.0176, 0.0219, 0.0234, 0.0277], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0005, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 16:33:26,960 INFO [train.py:898] (0/4) Epoch 3, batch 1700, loss[loss=0.2757, simple_loss=0.338, pruned_loss=0.1067, over 17080.00 frames. ], tot_loss[loss=0.2756, simple_loss=0.3401, pruned_loss=0.1056, over 3578982.27 frames. ], batch size: 78, lr: 3.25e-02, grad_scale: 8.0 +2023-03-08 16:33:27,090 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=8969.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:33:48,840 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=8986.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:33:51,241 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=8988.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:34:19,896 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9013.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:34:25,793 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.324e+02 5.263e+02 6.332e+02 7.938e+02 1.916e+03, threshold=1.266e+03, percent-clipped=4.0 +2023-03-08 16:34:26,991 INFO [train.py:898] (0/4) Epoch 3, batch 1750, loss[loss=0.2222, simple_loss=0.2901, pruned_loss=0.0772, over 18485.00 frames. ], tot_loss[loss=0.2751, simple_loss=0.3397, pruned_loss=0.1052, over 3571390.69 frames. ], batch size: 44, lr: 3.24e-02, grad_scale: 8.0 +2023-03-08 16:34:54,913 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4249, 3.3789, 1.4364, 4.2801, 2.5083, 4.3792, 2.3753, 3.8902], + device='cuda:0'), covar=tensor([0.0483, 0.0829, 0.1743, 0.0219, 0.1013, 0.0065, 0.1034, 0.0301], + device='cuda:0'), in_proj_covar=tensor([0.0107, 0.0152, 0.0142, 0.0101, 0.0135, 0.0070, 0.0139, 0.0126], + device='cuda:0'), out_proj_covar=tensor([1.5181e-04, 1.9115e-04, 1.7532e-04, 1.6387e-04, 1.7557e-04, 9.6248e-05, + 1.6834e-04, 1.5749e-04], device='cuda:0') +2023-03-08 16:35:00,442 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9047.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:35:13,098 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.06 vs. limit=5.0 +2023-03-08 16:35:25,188 INFO [train.py:898] (0/4) Epoch 3, batch 1800, loss[loss=0.2979, simple_loss=0.3667, pruned_loss=0.1146, over 17230.00 frames. ], tot_loss[loss=0.2744, simple_loss=0.3393, pruned_loss=0.1047, over 3581146.70 frames. ], batch size: 78, lr: 3.24e-02, grad_scale: 8.0 +2023-03-08 16:35:31,922 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9074.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:36:20,713 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9116.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:36:22,836 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.642e+02 5.266e+02 6.349e+02 9.123e+02 1.669e+03, threshold=1.270e+03, percent-clipped=4.0 +2023-03-08 16:36:24,030 INFO [train.py:898] (0/4) Epoch 3, batch 1850, loss[loss=0.2615, simple_loss=0.3343, pruned_loss=0.09435, over 18113.00 frames. ], tot_loss[loss=0.2736, simple_loss=0.3389, pruned_loss=0.1042, over 3586903.08 frames. ], batch size: 62, lr: 3.23e-02, grad_scale: 8.0 +2023-03-08 16:36:25,656 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0226, 4.6602, 2.0553, 4.7357, 5.1870, 2.0395, 4.0452, 3.8428], + device='cuda:0'), covar=tensor([0.0075, 0.0436, 0.1856, 0.0260, 0.0056, 0.1840, 0.0589, 0.0731], + device='cuda:0'), in_proj_covar=tensor([0.0073, 0.0103, 0.0162, 0.0129, 0.0066, 0.0151, 0.0157, 0.0151], + device='cuda:0'), out_proj_covar=tensor([8.8249e-05, 1.4663e-04, 1.8112e-04, 1.5086e-04, 8.1452e-05, 1.7450e-04, + 1.8104e-04, 1.8358e-04], device='cuda:0') +2023-03-08 16:37:22,866 INFO [train.py:898] (0/4) Epoch 3, batch 1900, loss[loss=0.3041, simple_loss=0.3655, pruned_loss=0.1214, over 18216.00 frames. ], tot_loss[loss=0.2749, simple_loss=0.3405, pruned_loss=0.1047, over 3588615.21 frames. ], batch size: 60, lr: 3.22e-02, grad_scale: 8.0 +2023-03-08 16:37:42,937 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=8.09 vs. limit=5.0 +2023-03-08 16:37:55,840 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7442, 3.2891, 4.1206, 4.5010, 2.2382, 4.5024, 3.7693, 2.2738], + device='cuda:0'), covar=tensor([0.0157, 0.0352, 0.0091, 0.0055, 0.1125, 0.0055, 0.0186, 0.0913], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0095, 0.0067, 0.0068, 0.0137, 0.0075, 0.0074, 0.0134], + device='cuda:0'), out_proj_covar=tensor([8.6033e-05, 9.3199e-05, 7.1823e-05, 6.7397e-05, 1.3257e-04, 7.2050e-05, + 8.2650e-05, 1.3470e-04], device='cuda:0') +2023-03-08 16:38:20,290 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.837e+02 5.068e+02 6.125e+02 7.856e+02 1.754e+03, threshold=1.225e+03, percent-clipped=6.0 +2023-03-08 16:38:21,520 INFO [train.py:898] (0/4) Epoch 3, batch 1950, loss[loss=0.2764, simple_loss=0.349, pruned_loss=0.1019, over 18019.00 frames. ], tot_loss[loss=0.2746, simple_loss=0.3403, pruned_loss=0.1045, over 3592842.08 frames. ], batch size: 65, lr: 3.22e-02, grad_scale: 8.0 +2023-03-08 16:39:21,323 INFO [train.py:898] (0/4) Epoch 3, batch 2000, loss[loss=0.2585, simple_loss=0.3355, pruned_loss=0.09078, over 18488.00 frames. ], tot_loss[loss=0.2739, simple_loss=0.3397, pruned_loss=0.1041, over 3586022.52 frames. ], batch size: 53, lr: 3.21e-02, grad_scale: 8.0 +2023-03-08 16:39:38,204 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9283.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:40:11,690 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1948, 4.6798, 4.4360, 4.4606, 4.2229, 4.5521, 4.7445, 4.5785], + device='cuda:0'), covar=tensor([0.1030, 0.0595, 0.1008, 0.0585, 0.1548, 0.0506, 0.0460, 0.0639], + device='cuda:0'), in_proj_covar=tensor([0.0310, 0.0260, 0.0203, 0.0270, 0.0387, 0.0279, 0.0276, 0.0249], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-08 16:40:19,549 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.362e+02 5.058e+02 6.498e+02 8.656e+02 1.894e+03, threshold=1.300e+03, percent-clipped=4.0 +2023-03-08 16:40:20,760 INFO [train.py:898] (0/4) Epoch 3, batch 2050, loss[loss=0.2286, simple_loss=0.2912, pruned_loss=0.08295, over 17680.00 frames. ], tot_loss[loss=0.2736, simple_loss=0.3392, pruned_loss=0.104, over 3582629.58 frames. ], batch size: 39, lr: 3.20e-02, grad_scale: 8.0 +2023-03-08 16:40:31,500 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5548, 3.1746, 1.4562, 4.0364, 2.7014, 4.2069, 1.7247, 3.6505], + device='cuda:0'), covar=tensor([0.0491, 0.0992, 0.1985, 0.0340, 0.1180, 0.0072, 0.1574, 0.0385], + device='cuda:0'), in_proj_covar=tensor([0.0115, 0.0160, 0.0147, 0.0109, 0.0143, 0.0073, 0.0145, 0.0131], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 16:40:42,808 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-08 16:40:47,859 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9342.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:41:14,679 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1746, 4.2250, 4.4110, 3.4990, 3.1114, 3.2240, 1.7553, 1.8313], + device='cuda:0'), covar=tensor([0.0227, 0.0213, 0.0117, 0.0148, 0.0362, 0.0137, 0.0628, 0.0770], + device='cuda:0'), in_proj_covar=tensor([0.0027, 0.0027, 0.0023, 0.0027, 0.0040, 0.0022, 0.0039, 0.0045], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0001, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 16:41:20,232 INFO [train.py:898] (0/4) Epoch 3, batch 2100, loss[loss=0.2698, simple_loss=0.3331, pruned_loss=0.1033, over 18458.00 frames. ], tot_loss[loss=0.2727, simple_loss=0.3389, pruned_loss=0.1033, over 3591374.84 frames. ], batch size: 59, lr: 3.20e-02, grad_scale: 8.0 +2023-03-08 16:41:20,402 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9369.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:41:23,991 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9956, 5.7419, 5.2906, 5.3965, 5.0886, 5.4642, 5.8980, 5.7125], + device='cuda:0'), covar=tensor([0.1326, 0.0665, 0.0460, 0.0798, 0.1855, 0.0592, 0.0467, 0.0619], + device='cuda:0'), in_proj_covar=tensor([0.0321, 0.0271, 0.0209, 0.0282, 0.0407, 0.0293, 0.0290, 0.0251], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0005, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-08 16:42:13,672 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6465, 3.3761, 1.2654, 4.1407, 2.6121, 4.4558, 1.9547, 3.9461], + device='cuda:0'), covar=tensor([0.0482, 0.0890, 0.2184, 0.0343, 0.1290, 0.0072, 0.1579, 0.0333], + device='cuda:0'), in_proj_covar=tensor([0.0115, 0.0157, 0.0146, 0.0107, 0.0142, 0.0073, 0.0142, 0.0129], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 16:42:13,677 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9414.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:42:15,877 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9416.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:42:17,748 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.310e+02 4.905e+02 6.085e+02 7.510e+02 1.544e+03, threshold=1.217e+03, percent-clipped=2.0 +2023-03-08 16:42:18,963 INFO [train.py:898] (0/4) Epoch 3, batch 2150, loss[loss=0.267, simple_loss=0.3344, pruned_loss=0.09977, over 18293.00 frames. ], tot_loss[loss=0.2737, simple_loss=0.3396, pruned_loss=0.1038, over 3582968.96 frames. ], batch size: 54, lr: 3.19e-02, grad_scale: 8.0 +2023-03-08 16:43:12,946 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=9464.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:43:18,551 INFO [train.py:898] (0/4) Epoch 3, batch 2200, loss[loss=0.2733, simple_loss=0.3478, pruned_loss=0.09939, over 18300.00 frames. ], tot_loss[loss=0.2721, simple_loss=0.3386, pruned_loss=0.1028, over 3586587.12 frames. ], batch size: 54, lr: 3.18e-02, grad_scale: 8.0 +2023-03-08 16:43:19,427 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-08 16:43:25,642 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9475.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:43:31,814 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-08 16:44:16,065 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.011e+02 5.021e+02 6.106e+02 7.632e+02 1.377e+03, threshold=1.221e+03, percent-clipped=3.0 +2023-03-08 16:44:17,244 INFO [train.py:898] (0/4) Epoch 3, batch 2250, loss[loss=0.2896, simple_loss=0.358, pruned_loss=0.1106, over 18351.00 frames. ], tot_loss[loss=0.2723, simple_loss=0.3386, pruned_loss=0.103, over 3586884.06 frames. ], batch size: 56, lr: 3.18e-02, grad_scale: 8.0 +2023-03-08 16:45:16,165 INFO [train.py:898] (0/4) Epoch 3, batch 2300, loss[loss=0.3025, simple_loss=0.3602, pruned_loss=0.1224, over 15989.00 frames. ], tot_loss[loss=0.2732, simple_loss=0.3392, pruned_loss=0.1035, over 3581193.79 frames. ], batch size: 94, lr: 3.17e-02, grad_scale: 8.0 +2023-03-08 16:45:32,480 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9583.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:45:33,749 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9584.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:46:14,651 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.768e+02 4.849e+02 6.155e+02 7.957e+02 1.846e+03, threshold=1.231e+03, percent-clipped=3.0 +2023-03-08 16:46:15,811 INFO [train.py:898] (0/4) Epoch 3, batch 2350, loss[loss=0.3526, simple_loss=0.396, pruned_loss=0.1546, over 12559.00 frames. ], tot_loss[loss=0.2735, simple_loss=0.3393, pruned_loss=0.1038, over 3570975.88 frames. ], batch size: 129, lr: 3.16e-02, grad_scale: 8.0 +2023-03-08 16:46:29,744 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=9631.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:46:42,481 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9642.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:46:46,079 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9645.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:47:00,822 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9658.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:47:14,967 INFO [train.py:898] (0/4) Epoch 3, batch 2400, loss[loss=0.2887, simple_loss=0.3616, pruned_loss=0.108, over 18057.00 frames. ], tot_loss[loss=0.2723, simple_loss=0.3384, pruned_loss=0.1031, over 3567411.86 frames. ], batch size: 62, lr: 3.16e-02, grad_scale: 8.0 +2023-03-08 16:47:15,279 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=9669.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:47:39,282 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=9690.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:47:48,687 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-08 16:48:10,791 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=9717.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:48:13,580 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.796e+02 5.501e+02 6.533e+02 8.034e+02 1.544e+03, threshold=1.307e+03, percent-clipped=2.0 +2023-03-08 16:48:13,606 INFO [train.py:898] (0/4) Epoch 3, batch 2450, loss[loss=0.2374, simple_loss=0.3165, pruned_loss=0.07908, over 18499.00 frames. ], tot_loss[loss=0.2728, simple_loss=0.3392, pruned_loss=0.1032, over 3563488.06 frames. ], batch size: 51, lr: 3.15e-02, grad_scale: 8.0 +2023-03-08 16:48:13,968 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9719.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:48:34,875 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.17 vs. limit=5.0 +2023-03-08 16:48:40,940 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.61 vs. limit=2.0 +2023-03-08 16:48:41,432 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=9743.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:49:11,285 INFO [train.py:898] (0/4) Epoch 3, batch 2500, loss[loss=0.2244, simple_loss=0.2921, pruned_loss=0.07839, over 18394.00 frames. ], tot_loss[loss=0.2719, simple_loss=0.338, pruned_loss=0.1029, over 3569397.65 frames. ], batch size: 42, lr: 3.14e-02, grad_scale: 8.0 +2023-03-08 16:49:13,123 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9770.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:49:51,931 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=9804.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:50:09,360 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.524e+02 5.803e+02 7.293e+02 8.388e+02 1.924e+03, threshold=1.459e+03, percent-clipped=4.0 +2023-03-08 16:50:09,385 INFO [train.py:898] (0/4) Epoch 3, batch 2550, loss[loss=0.2266, simple_loss=0.3065, pruned_loss=0.07337, over 18494.00 frames. ], tot_loss[loss=0.2711, simple_loss=0.3374, pruned_loss=0.1024, over 3585250.39 frames. ], batch size: 47, lr: 3.14e-02, grad_scale: 8.0 +2023-03-08 16:50:27,396 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0252, 4.0684, 4.3109, 3.4318, 2.8669, 2.4136, 1.8813, 2.0220], + device='cuda:0'), covar=tensor([0.0249, 0.0197, 0.0038, 0.0153, 0.0375, 0.0365, 0.0738, 0.0817], + device='cuda:0'), in_proj_covar=tensor([0.0026, 0.0027, 0.0023, 0.0028, 0.0041, 0.0023, 0.0044, 0.0048], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0001, 0.0001, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 16:51:06,867 INFO [train.py:898] (0/4) Epoch 3, batch 2600, loss[loss=0.3319, simple_loss=0.3777, pruned_loss=0.1431, over 12709.00 frames. ], tot_loss[loss=0.2705, simple_loss=0.337, pruned_loss=0.102, over 3579629.91 frames. ], batch size: 130, lr: 3.13e-02, grad_scale: 8.0 +2023-03-08 16:51:21,586 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.02 vs. limit=2.0 +2023-03-08 16:52:05,255 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.617e+02 5.195e+02 6.356e+02 7.709e+02 1.531e+03, threshold=1.271e+03, percent-clipped=3.0 +2023-03-08 16:52:05,281 INFO [train.py:898] (0/4) Epoch 3, batch 2650, loss[loss=0.2666, simple_loss=0.3385, pruned_loss=0.09731, over 17815.00 frames. ], tot_loss[loss=0.2706, simple_loss=0.3374, pruned_loss=0.1019, over 3587648.59 frames. ], batch size: 70, lr: 3.13e-02, grad_scale: 8.0 +2023-03-08 16:52:10,032 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.90 vs. limit=2.0 +2023-03-08 16:52:20,807 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.36 vs. limit=5.0 +2023-03-08 16:52:31,466 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=9940.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:52:55,057 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0402, 5.3276, 2.9092, 4.9908, 5.0881, 5.3404, 5.1834, 2.5156], + device='cuda:0'), covar=tensor([0.0175, 0.0039, 0.0696, 0.0062, 0.0053, 0.0042, 0.0082, 0.1151], + device='cuda:0'), in_proj_covar=tensor([0.0055, 0.0038, 0.0070, 0.0044, 0.0048, 0.0041, 0.0049, 0.0081], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 16:53:03,797 INFO [train.py:898] (0/4) Epoch 3, batch 2700, loss[loss=0.2479, simple_loss=0.3076, pruned_loss=0.09412, over 18479.00 frames. ], tot_loss[loss=0.2703, simple_loss=0.3373, pruned_loss=0.1017, over 3594193.05 frames. ], batch size: 44, lr: 3.12e-02, grad_scale: 8.0 +2023-03-08 16:53:41,038 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-10000.pt +2023-03-08 16:54:01,194 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10014.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:54:06,411 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.004e+02 5.454e+02 7.297e+02 9.817e+02 2.367e+03, threshold=1.459e+03, percent-clipped=11.0 +2023-03-08 16:54:06,435 INFO [train.py:898] (0/4) Epoch 3, batch 2750, loss[loss=0.2797, simple_loss=0.3489, pruned_loss=0.1053, over 18492.00 frames. ], tot_loss[loss=0.2698, simple_loss=0.3365, pruned_loss=0.1015, over 3594336.29 frames. ], batch size: 53, lr: 3.11e-02, grad_scale: 8.0 +2023-03-08 16:54:23,651 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2847, 5.2297, 4.5301, 5.2310, 5.1889, 4.5874, 5.2324, 4.5139], + device='cuda:0'), covar=tensor([0.0326, 0.0367, 0.1787, 0.0515, 0.0317, 0.0497, 0.0264, 0.0540], + device='cuda:0'), in_proj_covar=tensor([0.0233, 0.0249, 0.0396, 0.0198, 0.0184, 0.0228, 0.0245, 0.0287], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0005, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 16:54:31,093 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10039.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:54:56,973 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10062.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 16:55:04,436 INFO [train.py:898] (0/4) Epoch 3, batch 2800, loss[loss=0.2117, simple_loss=0.2817, pruned_loss=0.07088, over 17615.00 frames. ], tot_loss[loss=0.2702, simple_loss=0.3371, pruned_loss=0.1017, over 3579872.73 frames. ], batch size: 39, lr: 3.11e-02, grad_scale: 8.0 +2023-03-08 16:55:05,867 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10070.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:55:38,837 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.50 vs. limit=5.0 +2023-03-08 16:55:40,454 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10099.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:55:41,711 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10100.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 16:56:01,592 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=10118.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:56:02,533 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.762e+02 4.646e+02 5.829e+02 7.162e+02 1.376e+03, threshold=1.166e+03, percent-clipped=0.0 +2023-03-08 16:56:02,558 INFO [train.py:898] (0/4) Epoch 3, batch 2850, loss[loss=0.2387, simple_loss=0.3105, pruned_loss=0.08351, over 18546.00 frames. ], tot_loss[loss=0.2707, simple_loss=0.3371, pruned_loss=0.1021, over 3566280.50 frames. ], batch size: 49, lr: 3.10e-02, grad_scale: 8.0 +2023-03-08 16:56:02,859 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5224, 5.1664, 5.2977, 4.9647, 4.9401, 5.1736, 4.4975, 5.0046], + device='cuda:0'), covar=tensor([0.0355, 0.0364, 0.0199, 0.0237, 0.0387, 0.0226, 0.1093, 0.0300], + device='cuda:0'), in_proj_covar=tensor([0.0108, 0.0142, 0.0122, 0.0108, 0.0134, 0.0140, 0.0194, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0003, 0.0003, 0.0004, 0.0004, 0.0005, 0.0003], + device='cuda:0') +2023-03-08 16:56:07,163 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10123.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 16:56:26,843 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.1996, 4.1787, 5.2896, 4.2942, 3.1550, 2.5397, 4.6557, 5.3216], + device='cuda:0'), covar=tensor([0.1021, 0.0612, 0.0042, 0.0228, 0.0760, 0.1093, 0.0192, 0.0019], + device='cuda:0'), in_proj_covar=tensor([0.0116, 0.0081, 0.0051, 0.0096, 0.0134, 0.0137, 0.0100, 0.0047], + device='cuda:0'), out_proj_covar=tensor([1.8513e-04, 1.4670e-04, 8.2785e-05, 1.5581e-04, 2.0296e-04, 2.1127e-04, + 1.6099e-04, 7.6083e-05], device='cuda:0') +2023-03-08 16:56:36,909 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.68 vs. limit=2.0 +2023-03-08 16:57:01,501 INFO [train.py:898] (0/4) Epoch 3, batch 2900, loss[loss=0.3602, simple_loss=0.3947, pruned_loss=0.1628, over 12891.00 frames. ], tot_loss[loss=0.2694, simple_loss=0.3363, pruned_loss=0.1012, over 3576240.64 frames. ], batch size: 130, lr: 3.09e-02, grad_scale: 4.0 +2023-03-08 16:57:29,764 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2328, 4.1293, 4.0699, 3.0344, 3.1838, 2.9511, 2.0350, 2.0053], + device='cuda:0'), covar=tensor([0.0166, 0.0166, 0.0104, 0.0193, 0.0264, 0.0172, 0.0644, 0.0764], + device='cuda:0'), in_proj_covar=tensor([0.0028, 0.0028, 0.0023, 0.0029, 0.0043, 0.0023, 0.0046, 0.0049], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 16:57:40,763 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-08 16:58:00,060 INFO [train.py:898] (0/4) Epoch 3, batch 2950, loss[loss=0.2385, simple_loss=0.3111, pruned_loss=0.0829, over 18559.00 frames. ], tot_loss[loss=0.2697, simple_loss=0.3364, pruned_loss=0.1015, over 3577275.68 frames. ], batch size: 49, lr: 3.09e-02, grad_scale: 4.0 +2023-03-08 16:58:01,191 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.457e+02 4.931e+02 6.360e+02 7.565e+02 1.819e+03, threshold=1.272e+03, percent-clipped=6.0 +2023-03-08 16:58:25,515 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10240.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 16:58:48,161 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9497, 3.6825, 3.9195, 3.0219, 2.8359, 2.9398, 2.0578, 1.8120], + device='cuda:0'), covar=tensor([0.0219, 0.0195, 0.0057, 0.0200, 0.0373, 0.0203, 0.0660, 0.0834], + device='cuda:0'), in_proj_covar=tensor([0.0028, 0.0029, 0.0024, 0.0029, 0.0045, 0.0024, 0.0047, 0.0051], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 16:58:59,058 INFO [train.py:898] (0/4) Epoch 3, batch 3000, loss[loss=0.283, simple_loss=0.3462, pruned_loss=0.1099, over 17901.00 frames. ], tot_loss[loss=0.2688, simple_loss=0.3356, pruned_loss=0.101, over 3579550.77 frames. ], batch size: 70, lr: 3.08e-02, grad_scale: 4.0 +2023-03-08 16:58:59,060 INFO [train.py:923] (0/4) Computing validation loss +2023-03-08 16:59:10,964 INFO [train.py:932] (0/4) Epoch 3, validation: loss=0.2015, simple_loss=0.3025, pruned_loss=0.05021, over 944034.00 frames. +2023-03-08 16:59:10,964 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-08 16:59:34,258 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=10288.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:00:03,638 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10314.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:00:08,962 INFO [train.py:898] (0/4) Epoch 3, batch 3050, loss[loss=0.2544, simple_loss=0.3282, pruned_loss=0.09026, over 17350.00 frames. ], tot_loss[loss=0.2695, simple_loss=0.3362, pruned_loss=0.1014, over 3575599.91 frames. ], batch size: 78, lr: 3.08e-02, grad_scale: 4.0 +2023-03-08 17:00:10,063 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.678e+02 5.040e+02 6.054e+02 8.050e+02 1.536e+03, threshold=1.211e+03, percent-clipped=3.0 +2023-03-08 17:00:19,611 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-08 17:00:59,160 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=10362.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:01:07,132 INFO [train.py:898] (0/4) Epoch 3, batch 3100, loss[loss=0.2809, simple_loss=0.3495, pruned_loss=0.1061, over 18367.00 frames. ], tot_loss[loss=0.2687, simple_loss=0.3353, pruned_loss=0.101, over 3582825.47 frames. ], batch size: 55, lr: 3.07e-02, grad_scale: 4.0 +2023-03-08 17:01:38,741 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10395.0, num_to_drop=1, layers_to_drop={3} +2023-03-08 17:01:43,392 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10399.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:01:54,969 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10409.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:02:05,055 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10418.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 17:02:05,995 INFO [train.py:898] (0/4) Epoch 3, batch 3150, loss[loss=0.2814, simple_loss=0.3492, pruned_loss=0.1068, over 18387.00 frames. ], tot_loss[loss=0.2671, simple_loss=0.334, pruned_loss=0.1001, over 3581831.62 frames. ], batch size: 56, lr: 3.06e-02, grad_scale: 4.0 +2023-03-08 17:02:07,188 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.149e+02 4.695e+02 5.685e+02 7.343e+02 1.972e+03, threshold=1.137e+03, percent-clipped=4.0 +2023-03-08 17:02:40,173 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=10447.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:02:48,078 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10454.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:03:04,449 INFO [train.py:898] (0/4) Epoch 3, batch 3200, loss[loss=0.2589, simple_loss=0.3325, pruned_loss=0.09266, over 18630.00 frames. ], tot_loss[loss=0.2682, simple_loss=0.3351, pruned_loss=0.1007, over 3579607.55 frames. ], batch size: 52, lr: 3.06e-02, grad_scale: 8.0 +2023-03-08 17:03:05,096 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-08 17:03:05,976 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10470.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:03:31,887 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8047, 3.8760, 4.8572, 3.1226, 3.9784, 3.2945, 3.5504, 1.9932], + device='cuda:0'), covar=tensor([0.0457, 0.0295, 0.0046, 0.0316, 0.0347, 0.0898, 0.0598, 0.1221], + device='cuda:0'), in_proj_covar=tensor([0.0116, 0.0115, 0.0062, 0.0100, 0.0129, 0.0173, 0.0100, 0.0143], + device='cuda:0'), out_proj_covar=tensor([1.1748e-04, 1.1926e-04, 6.5348e-05, 1.0273e-04, 1.3540e-04, 1.7072e-04, + 1.1638e-04, 1.4055e-04], device='cuda:0') +2023-03-08 17:03:59,281 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10515.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 17:04:02,761 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1699, 4.3266, 4.8985, 3.0353, 4.1426, 3.1864, 3.7325, 1.9034], + device='cuda:0'), covar=tensor([0.0405, 0.0237, 0.0040, 0.0362, 0.0354, 0.0985, 0.0691, 0.1226], + device='cuda:0'), in_proj_covar=tensor([0.0115, 0.0114, 0.0062, 0.0099, 0.0127, 0.0172, 0.0100, 0.0141], + device='cuda:0'), out_proj_covar=tensor([1.1670e-04, 1.1818e-04, 6.5416e-05, 1.0204e-04, 1.3403e-04, 1.7024e-04, + 1.1581e-04, 1.3905e-04], device='cuda:0') +2023-03-08 17:04:03,373 INFO [train.py:898] (0/4) Epoch 3, batch 3250, loss[loss=0.2857, simple_loss=0.3403, pruned_loss=0.1156, over 18468.00 frames. ], tot_loss[loss=0.2677, simple_loss=0.3344, pruned_loss=0.1005, over 3585804.31 frames. ], batch size: 44, lr: 3.05e-02, grad_scale: 8.0 +2023-03-08 17:04:04,539 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.357e+02 4.994e+02 6.141e+02 8.166e+02 2.371e+03, threshold=1.228e+03, percent-clipped=6.0 +2023-03-08 17:04:10,818 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=2.06 vs. limit=2.0 +2023-03-08 17:04:29,071 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5828, 4.4351, 4.5463, 4.2089, 4.2550, 4.3756, 4.6732, 4.7193], + device='cuda:0'), covar=tensor([0.0071, 0.0120, 0.0101, 0.0122, 0.0129, 0.0105, 0.0142, 0.0140], + device='cuda:0'), in_proj_covar=tensor([0.0063, 0.0050, 0.0046, 0.0058, 0.0054, 0.0065, 0.0055, 0.0054], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 17:05:01,931 INFO [train.py:898] (0/4) Epoch 3, batch 3300, loss[loss=0.2555, simple_loss=0.3218, pruned_loss=0.09462, over 18386.00 frames. ], tot_loss[loss=0.2669, simple_loss=0.3339, pruned_loss=0.09997, over 3589620.42 frames. ], batch size: 48, lr: 3.05e-02, grad_scale: 8.0 +2023-03-08 17:05:30,415 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3430, 5.0917, 5.0945, 4.7466, 4.7078, 4.9125, 4.3009, 4.7770], + device='cuda:0'), covar=tensor([0.0333, 0.0323, 0.0219, 0.0269, 0.0555, 0.0240, 0.1337, 0.0350], + device='cuda:0'), in_proj_covar=tensor([0.0105, 0.0141, 0.0124, 0.0112, 0.0140, 0.0136, 0.0197, 0.0125], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0003, 0.0003, 0.0004, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-08 17:05:30,922 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.59 vs. limit=2.0 +2023-03-08 17:06:01,135 INFO [train.py:898] (0/4) Epoch 3, batch 3350, loss[loss=0.2977, simple_loss=0.3567, pruned_loss=0.1193, over 17929.00 frames. ], tot_loss[loss=0.2671, simple_loss=0.3342, pruned_loss=0.1, over 3588345.49 frames. ], batch size: 70, lr: 3.04e-02, grad_scale: 8.0 +2023-03-08 17:06:02,282 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.520e+02 5.384e+02 6.304e+02 8.125e+02 1.835e+03, threshold=1.261e+03, percent-clipped=2.0 +2023-03-08 17:06:06,245 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.47 vs. limit=5.0 +2023-03-08 17:06:59,029 INFO [train.py:898] (0/4) Epoch 3, batch 3400, loss[loss=0.2646, simple_loss=0.331, pruned_loss=0.09907, over 18275.00 frames. ], tot_loss[loss=0.2678, simple_loss=0.3348, pruned_loss=0.1004, over 3592783.22 frames. ], batch size: 49, lr: 3.03e-02, grad_scale: 8.0 +2023-03-08 17:07:21,182 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.97 vs. limit=2.0 +2023-03-08 17:07:30,293 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10695.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:07:56,722 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=10718.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 17:07:57,571 INFO [train.py:898] (0/4) Epoch 3, batch 3450, loss[loss=0.2897, simple_loss=0.3629, pruned_loss=0.1083, over 18397.00 frames. ], tot_loss[loss=0.2662, simple_loss=0.3337, pruned_loss=0.09938, over 3595829.66 frames. ], batch size: 52, lr: 3.03e-02, grad_scale: 8.0 +2023-03-08 17:07:58,709 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.826e+02 5.643e+02 6.682e+02 8.838e+02 1.430e+03, threshold=1.336e+03, percent-clipped=8.0 +2023-03-08 17:07:59,142 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3841, 4.4702, 2.3529, 4.8420, 5.4812, 2.5578, 4.1598, 3.9296], + device='cuda:0'), covar=tensor([0.0050, 0.1073, 0.1503, 0.0281, 0.0024, 0.1251, 0.0492, 0.0638], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0116, 0.0165, 0.0140, 0.0062, 0.0155, 0.0166, 0.0153], + device='cuda:0'), out_proj_covar=tensor([9.7429e-05, 1.6890e-04, 1.9487e-04, 1.7269e-04, 8.1409e-05, 1.8955e-04, + 2.0114e-04, 1.9572e-04], device='cuda:0') +2023-03-08 17:08:24,805 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=10743.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:08:26,048 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10744.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:08:47,128 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9389, 4.2304, 2.1595, 4.4035, 5.0508, 2.2964, 3.8458, 3.5519], + device='cuda:0'), covar=tensor([0.0056, 0.0871, 0.1659, 0.0364, 0.0034, 0.1503, 0.0575, 0.0801], + device='cuda:0'), in_proj_covar=tensor([0.0072, 0.0114, 0.0161, 0.0139, 0.0061, 0.0152, 0.0161, 0.0151], + device='cuda:0'), out_proj_covar=tensor([9.4288e-05, 1.6571e-04, 1.9061e-04, 1.7135e-04, 8.0031e-05, 1.8570e-04, + 1.9580e-04, 1.9245e-04], device='cuda:0') +2023-03-08 17:08:51,324 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10765.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:08:52,374 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=10766.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 17:08:55,409 INFO [train.py:898] (0/4) Epoch 3, batch 3500, loss[loss=0.2219, simple_loss=0.2794, pruned_loss=0.08222, over 18428.00 frames. ], tot_loss[loss=0.265, simple_loss=0.3327, pruned_loss=0.09861, over 3607274.02 frames. ], batch size: 43, lr: 3.02e-02, grad_scale: 8.0 +2023-03-08 17:09:01,518 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10774.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:09:27,907 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.99 vs. limit=2.0 +2023-03-08 17:09:36,591 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10805.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:09:41,039 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3996, 3.2261, 1.4852, 4.1872, 2.8987, 4.3290, 1.8265, 3.3355], + device='cuda:0'), covar=tensor([0.0541, 0.0924, 0.1922, 0.0300, 0.0994, 0.0086, 0.1390, 0.0471], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0169, 0.0153, 0.0118, 0.0147, 0.0079, 0.0150, 0.0136], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 17:09:41,865 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=10810.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 17:09:51,964 INFO [train.py:898] (0/4) Epoch 3, batch 3550, loss[loss=0.2574, simple_loss=0.3352, pruned_loss=0.08977, over 18252.00 frames. ], tot_loss[loss=0.2652, simple_loss=0.3332, pruned_loss=0.09859, over 3602859.39 frames. ], batch size: 60, lr: 3.02e-02, grad_scale: 8.0 +2023-03-08 17:09:52,945 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.843e+02 4.878e+02 6.150e+02 7.630e+02 1.368e+03, threshold=1.230e+03, percent-clipped=1.0 +2023-03-08 17:09:55,340 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6914, 5.1450, 5.2430, 5.0543, 4.9132, 5.1153, 4.3283, 4.9496], + device='cuda:0'), covar=tensor([0.0202, 0.0386, 0.0210, 0.0215, 0.0487, 0.0205, 0.1315, 0.0283], + device='cuda:0'), in_proj_covar=tensor([0.0108, 0.0146, 0.0128, 0.0117, 0.0145, 0.0140, 0.0203, 0.0128], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-08 17:10:09,060 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10835.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:10:22,120 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4503, 5.2538, 4.7880, 5.4201, 5.3298, 4.6734, 5.2342, 4.7118], + device='cuda:0'), covar=tensor([0.0332, 0.0408, 0.1742, 0.0493, 0.0440, 0.0487, 0.0366, 0.0738], + device='cuda:0'), in_proj_covar=tensor([0.0240, 0.0263, 0.0415, 0.0211, 0.0196, 0.0245, 0.0265, 0.0304], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0005, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 17:10:38,719 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-08 17:10:46,352 INFO [train.py:898] (0/4) Epoch 3, batch 3600, loss[loss=0.2335, simple_loss=0.3024, pruned_loss=0.08233, over 18509.00 frames. ], tot_loss[loss=0.2654, simple_loss=0.3335, pruned_loss=0.0987, over 3600879.74 frames. ], batch size: 47, lr: 3.01e-02, grad_scale: 8.0 +2023-03-08 17:11:03,929 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=10885.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 17:11:21,742 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-3.pt +2023-03-08 17:11:51,263 INFO [train.py:898] (0/4) Epoch 4, batch 0, loss[loss=0.2711, simple_loss=0.3486, pruned_loss=0.09679, over 18482.00 frames. ], tot_loss[loss=0.2711, simple_loss=0.3486, pruned_loss=0.09679, over 18482.00 frames. ], batch size: 59, lr: 2.81e-02, grad_scale: 8.0 +2023-03-08 17:11:51,265 INFO [train.py:923] (0/4) Computing validation loss +2023-03-08 17:12:03,131 INFO [train.py:932] (0/4) Epoch 4, validation: loss=0.2018, simple_loss=0.3032, pruned_loss=0.05022, over 944034.00 frames. +2023-03-08 17:12:03,132 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-08 17:12:22,935 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.062e+02 5.140e+02 6.071e+02 7.420e+02 1.697e+03, threshold=1.214e+03, percent-clipped=4.0 +2023-03-08 17:12:55,024 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=10946.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 17:13:02,668 INFO [train.py:898] (0/4) Epoch 4, batch 50, loss[loss=0.2821, simple_loss=0.3439, pruned_loss=0.1101, over 18295.00 frames. ], tot_loss[loss=0.2682, simple_loss=0.3334, pruned_loss=0.1015, over 806288.49 frames. ], batch size: 57, lr: 2.81e-02, grad_scale: 8.0 +2023-03-08 17:14:00,535 INFO [train.py:898] (0/4) Epoch 4, batch 100, loss[loss=0.2424, simple_loss=0.3208, pruned_loss=0.08198, over 18469.00 frames. ], tot_loss[loss=0.2615, simple_loss=0.3293, pruned_loss=0.09681, over 1435720.68 frames. ], batch size: 59, lr: 2.80e-02, grad_scale: 8.0 +2023-03-08 17:14:08,167 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-03-08 17:14:19,670 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.738e+02 4.868e+02 5.817e+02 7.449e+02 2.139e+03, threshold=1.163e+03, percent-clipped=6.0 +2023-03-08 17:14:39,522 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-03-08 17:14:42,426 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4554, 3.5969, 2.8375, 2.8747, 3.1772, 2.8447, 2.6138, 3.5404], + device='cuda:0'), covar=tensor([0.0064, 0.0056, 0.0194, 0.0129, 0.0136, 0.0195, 0.0243, 0.0134], + device='cuda:0'), in_proj_covar=tensor([0.0044, 0.0049, 0.0050, 0.0076, 0.0051, 0.0079, 0.0091, 0.0049], + device='cuda:0'), out_proj_covar=tensor([6.5492e-05, 7.6933e-05, 8.2404e-05, 1.1838e-04, 7.9595e-05, 1.2383e-04, + 1.5072e-04, 7.9375e-05], device='cuda:0') +2023-03-08 17:14:58,824 INFO [train.py:898] (0/4) Epoch 4, batch 150, loss[loss=0.2752, simple_loss=0.3478, pruned_loss=0.1013, over 18131.00 frames. ], tot_loss[loss=0.2556, simple_loss=0.3247, pruned_loss=0.09327, over 1911308.30 frames. ], batch size: 62, lr: 2.80e-02, grad_scale: 8.0 +2023-03-08 17:15:12,367 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11065.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:15:16,845 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6859, 5.3334, 5.2556, 5.0317, 4.9852, 5.1633, 4.4720, 5.0585], + device='cuda:0'), covar=tensor([0.0196, 0.0196, 0.0160, 0.0200, 0.0291, 0.0188, 0.0897, 0.0245], + device='cuda:0'), in_proj_covar=tensor([0.0108, 0.0141, 0.0128, 0.0117, 0.0139, 0.0140, 0.0199, 0.0128], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0003, 0.0004, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-08 17:15:53,921 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11100.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:15:55,567 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.93 vs. limit=2.0 +2023-03-08 17:15:57,090 INFO [train.py:898] (0/4) Epoch 4, batch 200, loss[loss=0.2352, simple_loss=0.3121, pruned_loss=0.07914, over 18373.00 frames. ], tot_loss[loss=0.2554, simple_loss=0.325, pruned_loss=0.09291, over 2285081.08 frames. ], batch size: 50, lr: 2.79e-02, grad_scale: 8.0 +2023-03-08 17:16:05,181 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11110.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 17:16:08,741 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=11113.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:16:16,250 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.946e+02 4.745e+02 5.920e+02 7.633e+02 1.560e+03, threshold=1.184e+03, percent-clipped=4.0 +2023-03-08 17:16:17,698 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2452, 5.2267, 4.5418, 5.2046, 5.2108, 4.6343, 5.1876, 4.5811], + device='cuda:0'), covar=tensor([0.0341, 0.0325, 0.1618, 0.0540, 0.0319, 0.0347, 0.0294, 0.0593], + device='cuda:0'), in_proj_covar=tensor([0.0252, 0.0273, 0.0427, 0.0217, 0.0203, 0.0245, 0.0263, 0.0315], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0005, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 17:16:27,636 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11130.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:16:55,260 INFO [train.py:898] (0/4) Epoch 4, batch 250, loss[loss=0.2321, simple_loss=0.3081, pruned_loss=0.0781, over 18503.00 frames. ], tot_loss[loss=0.2551, simple_loss=0.325, pruned_loss=0.0926, over 2579667.97 frames. ], batch size: 47, lr: 2.79e-02, grad_scale: 8.0 +2023-03-08 17:17:01,241 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=11158.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:17:19,412 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11174.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:17:52,791 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.04 vs. limit=2.0 +2023-03-08 17:17:54,162 INFO [train.py:898] (0/4) Epoch 4, batch 300, loss[loss=0.2721, simple_loss=0.3367, pruned_loss=0.1037, over 18492.00 frames. ], tot_loss[loss=0.2576, simple_loss=0.3275, pruned_loss=0.09386, over 2803010.64 frames. ], batch size: 53, lr: 2.78e-02, grad_scale: 8.0 +2023-03-08 17:18:04,978 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1131, 5.0749, 4.4334, 5.0742, 5.0518, 4.5458, 5.0194, 4.4850], + device='cuda:0'), covar=tensor([0.0304, 0.0388, 0.1625, 0.0534, 0.0362, 0.0350, 0.0315, 0.0664], + device='cuda:0'), in_proj_covar=tensor([0.0247, 0.0273, 0.0422, 0.0216, 0.0201, 0.0245, 0.0264, 0.0314], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0005, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 17:18:13,960 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.374e+02 4.613e+02 5.593e+02 6.411e+02 1.138e+03, threshold=1.119e+03, percent-clipped=0.0 +2023-03-08 17:18:31,456 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11235.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:18:38,664 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11241.0, num_to_drop=1, layers_to_drop={3} +2023-03-08 17:18:53,615 INFO [train.py:898] (0/4) Epoch 4, batch 350, loss[loss=0.2371, simple_loss=0.3078, pruned_loss=0.08322, over 18375.00 frames. ], tot_loss[loss=0.2577, simple_loss=0.3277, pruned_loss=0.0939, over 2980092.96 frames. ], batch size: 50, lr: 2.78e-02, grad_scale: 8.0 +2023-03-08 17:19:16,051 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3937, 1.8993, 4.5945, 2.6388, 3.2201, 4.7432, 4.5703, 4.1095], + device='cuda:0'), covar=tensor([0.0252, 0.0786, 0.0147, 0.0572, 0.1065, 0.0025, 0.0137, 0.0139], + device='cuda:0'), in_proj_covar=tensor([0.0112, 0.0147, 0.0076, 0.0145, 0.0227, 0.0081, 0.0118, 0.0110], + device='cuda:0'), out_proj_covar=tensor([9.2845e-05, 1.2082e-04, 6.7164e-05, 1.1160e-04, 1.8280e-04, 5.9908e-05, + 9.8729e-05, 8.7966e-05], device='cuda:0') +2023-03-08 17:19:29,910 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.14 vs. limit=5.0 +2023-03-08 17:19:51,471 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6411, 3.4565, 1.6917, 4.3853, 3.1171, 4.5375, 2.0075, 3.8162], + device='cuda:0'), covar=tensor([0.0424, 0.0805, 0.1542, 0.0272, 0.0823, 0.0080, 0.1225, 0.0299], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0173, 0.0154, 0.0119, 0.0149, 0.0085, 0.0155, 0.0137], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 17:19:52,173 INFO [train.py:898] (0/4) Epoch 4, batch 400, loss[loss=0.2774, simple_loss=0.3455, pruned_loss=0.1047, over 18268.00 frames. ], tot_loss[loss=0.2592, simple_loss=0.3289, pruned_loss=0.09475, over 3109333.71 frames. ], batch size: 57, lr: 2.77e-02, grad_scale: 8.0 +2023-03-08 17:19:59,833 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-03-08 17:20:10,946 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.476e+02 4.514e+02 5.687e+02 6.910e+02 1.325e+03, threshold=1.137e+03, percent-clipped=4.0 +2023-03-08 17:20:30,459 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3967, 4.4424, 2.5160, 4.8670, 5.4457, 2.6340, 4.4411, 3.7418], + device='cuda:0'), covar=tensor([0.0042, 0.0898, 0.1340, 0.0257, 0.0028, 0.1265, 0.0373, 0.0752], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0118, 0.0163, 0.0141, 0.0066, 0.0155, 0.0163, 0.0155], + device='cuda:0'), out_proj_covar=tensor([9.6538e-05, 1.7278e-04, 1.9669e-04, 1.7708e-04, 8.8279e-05, 1.9092e-04, + 2.0029e-04, 1.9940e-04], device='cuda:0') +2023-03-08 17:20:50,454 INFO [train.py:898] (0/4) Epoch 4, batch 450, loss[loss=0.2777, simple_loss=0.3536, pruned_loss=0.1009, over 18487.00 frames. ], tot_loss[loss=0.2585, simple_loss=0.3282, pruned_loss=0.09438, over 3230129.89 frames. ], batch size: 51, lr: 2.77e-02, grad_scale: 8.0 +2023-03-08 17:21:17,912 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.3356, 3.8360, 5.0194, 4.2020, 3.1042, 2.8328, 4.4543, 5.0519], + device='cuda:0'), covar=tensor([0.0904, 0.0680, 0.0040, 0.0267, 0.0718, 0.0941, 0.0190, 0.0023], + device='cuda:0'), in_proj_covar=tensor([0.0117, 0.0099, 0.0055, 0.0107, 0.0136, 0.0143, 0.0110, 0.0052], + device='cuda:0'), out_proj_covar=tensor([1.9062e-04, 1.7440e-04, 9.0794e-05, 1.7432e-04, 2.0918e-04, 2.2222e-04, + 1.7535e-04, 8.5033e-05], device='cuda:0') +2023-03-08 17:21:26,218 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.1999, 3.9196, 4.9173, 3.9213, 2.9512, 2.4225, 4.1068, 4.8994], + device='cuda:0'), covar=tensor([0.1039, 0.0720, 0.0046, 0.0273, 0.0858, 0.1184, 0.0254, 0.0025], + device='cuda:0'), in_proj_covar=tensor([0.0119, 0.0100, 0.0055, 0.0108, 0.0138, 0.0144, 0.0111, 0.0053], + device='cuda:0'), out_proj_covar=tensor([1.9306e-04, 1.7627e-04, 9.2001e-05, 1.7609e-04, 2.1124e-04, 2.2481e-04, + 1.7727e-04, 8.5977e-05], device='cuda:0') +2023-03-08 17:21:43,954 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1631, 5.8146, 5.2308, 5.5103, 5.1667, 5.3917, 5.8460, 5.7544], + device='cuda:0'), covar=tensor([0.1071, 0.0557, 0.0459, 0.0685, 0.1762, 0.0590, 0.0510, 0.0595], + device='cuda:0'), in_proj_covar=tensor([0.0347, 0.0285, 0.0226, 0.0309, 0.0435, 0.0307, 0.0325, 0.0271], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0005, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-08 17:21:45,805 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11400.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:21:49,042 INFO [train.py:898] (0/4) Epoch 4, batch 500, loss[loss=0.2406, simple_loss=0.3168, pruned_loss=0.08214, over 18482.00 frames. ], tot_loss[loss=0.2586, simple_loss=0.3287, pruned_loss=0.09423, over 3321973.40 frames. ], batch size: 53, lr: 2.76e-02, grad_scale: 4.0 +2023-03-08 17:22:09,637 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.774e+02 5.147e+02 6.792e+02 8.631e+02 2.583e+03, threshold=1.358e+03, percent-clipped=9.0 +2023-03-08 17:22:20,271 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11430.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:22:40,920 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=11448.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:22:45,873 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.34 vs. limit=5.0 +2023-03-08 17:22:46,352 INFO [train.py:898] (0/4) Epoch 4, batch 550, loss[loss=0.2408, simple_loss=0.3172, pruned_loss=0.08218, over 18485.00 frames. ], tot_loss[loss=0.2562, simple_loss=0.3264, pruned_loss=0.09303, over 3397515.20 frames. ], batch size: 51, lr: 2.76e-02, grad_scale: 4.0 +2023-03-08 17:23:17,097 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=11478.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:23:45,217 INFO [train.py:898] (0/4) Epoch 4, batch 600, loss[loss=0.2417, simple_loss=0.3011, pruned_loss=0.09108, over 18502.00 frames. ], tot_loss[loss=0.2581, simple_loss=0.3279, pruned_loss=0.09417, over 3422260.18 frames. ], batch size: 44, lr: 2.75e-02, grad_scale: 4.0 +2023-03-08 17:24:07,206 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.433e+02 4.507e+02 5.343e+02 7.594e+02 1.826e+03, threshold=1.069e+03, percent-clipped=2.0 +2023-03-08 17:24:17,742 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11530.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:24:19,125 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1749, 4.2794, 2.0023, 4.6901, 5.2390, 2.6238, 4.2597, 3.9594], + device='cuda:0'), covar=tensor([0.0071, 0.0896, 0.1684, 0.0293, 0.0034, 0.1315, 0.0432, 0.0633], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0127, 0.0164, 0.0144, 0.0068, 0.0156, 0.0168, 0.0157], + device='cuda:0'), out_proj_covar=tensor([9.9614e-05, 1.8340e-04, 1.9903e-04, 1.8173e-04, 9.0681e-05, 1.9374e-04, + 2.0692e-04, 2.0350e-04], device='cuda:0') +2023-03-08 17:24:20,229 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11532.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:24:29,973 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11541.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 17:24:43,295 INFO [train.py:898] (0/4) Epoch 4, batch 650, loss[loss=0.2414, simple_loss=0.3028, pruned_loss=0.08995, over 18475.00 frames. ], tot_loss[loss=0.258, simple_loss=0.3274, pruned_loss=0.09428, over 3461328.38 frames. ], batch size: 43, lr: 2.75e-02, grad_scale: 4.0 +2023-03-08 17:25:16,922 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1660, 2.7745, 2.3206, 2.7445, 3.1349, 3.3372, 2.4972, 2.7639], + device='cuda:0'), covar=tensor([0.0418, 0.0537, 0.1289, 0.0454, 0.0393, 0.0302, 0.0571, 0.0372], + device='cuda:0'), in_proj_covar=tensor([0.0086, 0.0062, 0.0126, 0.0092, 0.0073, 0.0051, 0.0086, 0.0079], + device='cuda:0'), out_proj_covar=tensor([1.6192e-04, 1.1876e-04, 2.1829e-04, 1.6163e-04, 1.3773e-04, 9.1097e-05, + 1.5868e-04, 1.4295e-04], device='cuda:0') +2023-03-08 17:25:25,670 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=11589.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 17:25:30,269 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11593.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:25:41,337 INFO [train.py:898] (0/4) Epoch 4, batch 700, loss[loss=0.274, simple_loss=0.3465, pruned_loss=0.1008, over 16963.00 frames. ], tot_loss[loss=0.2568, simple_loss=0.3262, pruned_loss=0.09368, over 3490380.92 frames. ], batch size: 78, lr: 2.74e-02, grad_scale: 4.0 +2023-03-08 17:26:03,621 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.297e+02 5.310e+02 6.723e+02 8.319e+02 1.846e+03, threshold=1.345e+03, percent-clipped=5.0 +2023-03-08 17:26:19,754 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11635.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:26:39,985 INFO [train.py:898] (0/4) Epoch 4, batch 750, loss[loss=0.232, simple_loss=0.2982, pruned_loss=0.0829, over 17698.00 frames. ], tot_loss[loss=0.2567, simple_loss=0.326, pruned_loss=0.0937, over 3514866.26 frames. ], batch size: 39, lr: 2.74e-02, grad_scale: 4.0 +2023-03-08 17:26:52,786 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6731, 4.5405, 4.7369, 4.4011, 4.3545, 4.4251, 4.9284, 4.9040], + device='cuda:0'), covar=tensor([0.0066, 0.0092, 0.0092, 0.0104, 0.0096, 0.0122, 0.0082, 0.0099], + device='cuda:0'), in_proj_covar=tensor([0.0061, 0.0049, 0.0046, 0.0059, 0.0053, 0.0066, 0.0055, 0.0052], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 17:26:56,788 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.70 vs. limit=2.0 +2023-03-08 17:27:31,194 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11696.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:27:31,526 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.89 vs. limit=2.0 +2023-03-08 17:27:34,636 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11699.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:27:38,699 INFO [train.py:898] (0/4) Epoch 4, batch 800, loss[loss=0.3645, simple_loss=0.3979, pruned_loss=0.1655, over 12517.00 frames. ], tot_loss[loss=0.2555, simple_loss=0.325, pruned_loss=0.09297, over 3538862.38 frames. ], batch size: 129, lr: 2.73e-02, grad_scale: 8.0 +2023-03-08 17:28:00,217 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.738e+02 5.059e+02 5.928e+02 8.034e+02 1.891e+03, threshold=1.186e+03, percent-clipped=2.0 +2023-03-08 17:28:15,506 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.62 vs. limit=2.0 +2023-03-08 17:28:37,646 INFO [train.py:898] (0/4) Epoch 4, batch 850, loss[loss=0.3197, simple_loss=0.3675, pruned_loss=0.1359, over 12754.00 frames. ], tot_loss[loss=0.256, simple_loss=0.3253, pruned_loss=0.09339, over 3547208.99 frames. ], batch size: 130, lr: 2.73e-02, grad_scale: 8.0 +2023-03-08 17:28:46,050 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11760.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:29:37,624 INFO [train.py:898] (0/4) Epoch 4, batch 900, loss[loss=0.2706, simple_loss=0.3446, pruned_loss=0.09834, over 18255.00 frames. ], tot_loss[loss=0.2564, simple_loss=0.3258, pruned_loss=0.09345, over 3552934.89 frames. ], batch size: 60, lr: 2.72e-02, grad_scale: 8.0 +2023-03-08 17:29:52,641 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11816.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:29:58,018 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.973e+02 4.760e+02 5.740e+02 6.849e+02 1.554e+03, threshold=1.148e+03, percent-clipped=4.0 +2023-03-08 17:30:11,161 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=11830.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:30:29,148 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11846.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:30:36,936 INFO [train.py:898] (0/4) Epoch 4, batch 950, loss[loss=0.261, simple_loss=0.3401, pruned_loss=0.09093, over 18268.00 frames. ], tot_loss[loss=0.2561, simple_loss=0.3256, pruned_loss=0.09325, over 3556458.11 frames. ], batch size: 57, lr: 2.72e-02, grad_scale: 8.0 +2023-03-08 17:30:51,857 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4590, 5.3937, 4.7328, 5.3070, 5.2508, 4.7510, 5.3031, 4.8530], + device='cuda:0'), covar=tensor([0.0256, 0.0345, 0.1715, 0.0670, 0.0454, 0.0374, 0.0294, 0.0671], + device='cuda:0'), in_proj_covar=tensor([0.0254, 0.0276, 0.0439, 0.0229, 0.0211, 0.0257, 0.0273, 0.0327], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0005, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 17:31:05,065 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11877.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:31:06,030 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=11878.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:31:18,410 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11888.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:31:35,205 INFO [train.py:898] (0/4) Epoch 4, batch 1000, loss[loss=0.2284, simple_loss=0.2943, pruned_loss=0.08125, over 18585.00 frames. ], tot_loss[loss=0.2546, simple_loss=0.3243, pruned_loss=0.09244, over 3569228.54 frames. ], batch size: 45, lr: 2.71e-02, grad_scale: 8.0 +2023-03-08 17:31:40,123 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11907.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:31:54,939 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11920.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:31:55,672 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.070e+02 4.938e+02 5.692e+02 6.896e+02 1.055e+03, threshold=1.138e+03, percent-clipped=0.0 +2023-03-08 17:32:24,774 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5841, 3.1876, 2.7004, 2.9331, 3.2574, 3.5149, 2.9052, 3.1220], + device='cuda:0'), covar=tensor([0.0304, 0.0278, 0.0914, 0.0393, 0.0343, 0.0236, 0.0417, 0.0233], + device='cuda:0'), in_proj_covar=tensor([0.0080, 0.0059, 0.0119, 0.0088, 0.0069, 0.0050, 0.0079, 0.0075], + device='cuda:0'), out_proj_covar=tensor([1.5283e-04, 1.1569e-04, 2.0714e-04, 1.5617e-04, 1.3321e-04, 9.0528e-05, + 1.4696e-04, 1.3730e-04], device='cuda:0') +2023-03-08 17:32:28,234 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11948.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:32:33,399 INFO [train.py:898] (0/4) Epoch 4, batch 1050, loss[loss=0.2301, simple_loss=0.3008, pruned_loss=0.07967, over 18503.00 frames. ], tot_loss[loss=0.2544, simple_loss=0.324, pruned_loss=0.09235, over 3571355.45 frames. ], batch size: 44, lr: 2.71e-02, grad_scale: 8.0 +2023-03-08 17:33:05,181 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=11981.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:33:05,573 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.01 vs. limit=2.0 +2023-03-08 17:33:08,854 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=11984.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:33:17,597 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=11991.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:33:27,644 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-12000.pt +2023-03-08 17:33:35,192 INFO [train.py:898] (0/4) Epoch 4, batch 1100, loss[loss=0.2952, simple_loss=0.3571, pruned_loss=0.1167, over 18083.00 frames. ], tot_loss[loss=0.255, simple_loss=0.3242, pruned_loss=0.09284, over 3576478.27 frames. ], batch size: 62, lr: 2.70e-02, grad_scale: 4.0 +2023-03-08 17:33:42,440 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12009.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:33:56,849 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.224e+02 5.074e+02 6.054e+02 6.990e+02 2.904e+03, threshold=1.211e+03, percent-clipped=5.0 +2023-03-08 17:34:25,351 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12045.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:34:34,309 INFO [train.py:898] (0/4) Epoch 4, batch 1150, loss[loss=0.2263, simple_loss=0.2844, pruned_loss=0.08406, over 18170.00 frames. ], tot_loss[loss=0.254, simple_loss=0.3236, pruned_loss=0.09225, over 3575840.20 frames. ], batch size: 40, lr: 2.70e-02, grad_scale: 4.0 +2023-03-08 17:34:36,786 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12055.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:34:44,021 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8415, 4.7705, 4.8273, 4.6142, 4.7216, 5.3213, 4.8940, 4.7667], + device='cuda:0'), covar=tensor([0.0767, 0.0646, 0.0639, 0.0596, 0.1147, 0.0674, 0.0566, 0.1368], + device='cuda:0'), in_proj_covar=tensor([0.0216, 0.0158, 0.0167, 0.0164, 0.0215, 0.0246, 0.0157, 0.0238], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 17:34:56,949 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-08 17:35:00,833 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=12076.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:35:23,221 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2919, 4.0818, 4.1439, 3.0206, 3.2771, 3.1103, 1.9987, 1.5191], + device='cuda:0'), covar=tensor([0.0233, 0.0159, 0.0091, 0.0263, 0.0305, 0.0209, 0.0784, 0.1004], + device='cuda:0'), in_proj_covar=tensor([0.0030, 0.0030, 0.0025, 0.0034, 0.0050, 0.0027, 0.0051, 0.0054], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 17:35:30,392 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.56 vs. limit=2.0 +2023-03-08 17:35:32,889 INFO [train.py:898] (0/4) Epoch 4, batch 1200, loss[loss=0.2677, simple_loss=0.3388, pruned_loss=0.09827, over 17038.00 frames. ], tot_loss[loss=0.2544, simple_loss=0.3238, pruned_loss=0.0925, over 3565615.78 frames. ], batch size: 78, lr: 2.69e-02, grad_scale: 8.0 +2023-03-08 17:35:54,812 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.657e+02 4.866e+02 5.977e+02 7.705e+02 1.703e+03, threshold=1.195e+03, percent-clipped=4.0 +2023-03-08 17:36:12,324 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12137.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:36:31,682 INFO [train.py:898] (0/4) Epoch 4, batch 1250, loss[loss=0.2572, simple_loss=0.3321, pruned_loss=0.09118, over 18490.00 frames. ], tot_loss[loss=0.2536, simple_loss=0.3232, pruned_loss=0.09204, over 3567875.41 frames. ], batch size: 59, lr: 2.69e-02, grad_scale: 8.0 +2023-03-08 17:36:53,287 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12172.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:37:11,224 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12188.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:37:17,382 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5370, 5.4586, 4.9304, 5.5687, 5.4719, 4.9092, 5.4033, 4.9610], + device='cuda:0'), covar=tensor([0.0329, 0.0298, 0.1632, 0.0466, 0.0299, 0.0363, 0.0356, 0.0658], + device='cuda:0'), in_proj_covar=tensor([0.0265, 0.0285, 0.0441, 0.0236, 0.0216, 0.0267, 0.0284, 0.0339], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0003, 0.0003, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 17:37:29,462 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12202.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:37:30,435 INFO [train.py:898] (0/4) Epoch 4, batch 1300, loss[loss=0.2442, simple_loss=0.3168, pruned_loss=0.08577, over 18357.00 frames. ], tot_loss[loss=0.2535, simple_loss=0.3232, pruned_loss=0.09192, over 3576300.37 frames. ], batch size: 50, lr: 2.68e-02, grad_scale: 4.0 +2023-03-08 17:37:52,910 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.709e+02 4.849e+02 5.902e+02 7.729e+02 1.516e+03, threshold=1.180e+03, percent-clipped=2.0 +2023-03-08 17:38:07,767 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=12236.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:38:23,428 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.02 vs. limit=2.0 +2023-03-08 17:38:25,850 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4999, 5.0089, 5.4920, 5.4060, 5.2231, 6.0576, 5.6022, 5.3124], + device='cuda:0'), covar=tensor([0.0637, 0.0567, 0.0525, 0.0527, 0.1081, 0.0682, 0.0622, 0.1289], + device='cuda:0'), in_proj_covar=tensor([0.0227, 0.0164, 0.0177, 0.0171, 0.0220, 0.0258, 0.0166, 0.0245], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 17:38:29,246 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-08 17:38:29,655 INFO [train.py:898] (0/4) Epoch 4, batch 1350, loss[loss=0.2503, simple_loss=0.3152, pruned_loss=0.09268, over 18285.00 frames. ], tot_loss[loss=0.253, simple_loss=0.3229, pruned_loss=0.09156, over 3582625.99 frames. ], batch size: 49, lr: 2.68e-02, grad_scale: 4.0 +2023-03-08 17:38:41,344 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3100, 5.9711, 5.4329, 5.8045, 5.3015, 5.5108, 6.0471, 5.9906], + device='cuda:0'), covar=tensor([0.1119, 0.0522, 0.0475, 0.0559, 0.1589, 0.0588, 0.0400, 0.0485], + device='cuda:0'), in_proj_covar=tensor([0.0354, 0.0302, 0.0237, 0.0314, 0.0449, 0.0319, 0.0338, 0.0298], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-08 17:38:56,147 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12276.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:39:13,279 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12291.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:39:27,798 INFO [train.py:898] (0/4) Epoch 4, batch 1400, loss[loss=0.2298, simple_loss=0.3022, pruned_loss=0.07875, over 18264.00 frames. ], tot_loss[loss=0.2553, simple_loss=0.3248, pruned_loss=0.09291, over 3570680.36 frames. ], batch size: 47, lr: 2.67e-02, grad_scale: 4.0 +2023-03-08 17:39:29,724 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12304.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:39:30,922 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5731, 5.3098, 5.3103, 4.9660, 4.8285, 5.0502, 4.4291, 5.0384], + device='cuda:0'), covar=tensor([0.0228, 0.0203, 0.0150, 0.0258, 0.0427, 0.0229, 0.1024, 0.0263], + device='cuda:0'), in_proj_covar=tensor([0.0114, 0.0153, 0.0139, 0.0128, 0.0149, 0.0153, 0.0216, 0.0135], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0005, 0.0004], + device='cuda:0') +2023-03-08 17:39:51,202 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.094e+02 5.084e+02 6.027e+02 7.946e+02 1.309e+03, threshold=1.205e+03, percent-clipped=1.0 +2023-03-08 17:40:09,785 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=12339.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:40:11,031 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12340.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:40:15,782 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.19 vs. limit=2.0 +2023-03-08 17:40:26,786 INFO [train.py:898] (0/4) Epoch 4, batch 1450, loss[loss=0.2294, simple_loss=0.2902, pruned_loss=0.08423, over 18430.00 frames. ], tot_loss[loss=0.2562, simple_loss=0.3257, pruned_loss=0.09331, over 3564730.47 frames. ], batch size: 43, lr: 2.67e-02, grad_scale: 4.0 +2023-03-08 17:40:29,893 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12355.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:41:06,602 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3443, 2.6638, 2.5375, 2.7147, 3.0570, 3.4802, 2.6834, 3.1908], + device='cuda:0'), covar=tensor([0.0471, 0.0262, 0.0816, 0.0358, 0.0393, 0.0157, 0.0631, 0.0287], + device='cuda:0'), in_proj_covar=tensor([0.0082, 0.0059, 0.0119, 0.0086, 0.0072, 0.0049, 0.0082, 0.0076], + device='cuda:0'), out_proj_covar=tensor([1.5831e-04, 1.1687e-04, 2.0848e-04, 1.5634e-04, 1.3818e-04, 8.7544e-05, + 1.5563e-04, 1.4294e-04], device='cuda:0') +2023-03-08 17:41:24,334 INFO [train.py:898] (0/4) Epoch 4, batch 1500, loss[loss=0.2991, simple_loss=0.3619, pruned_loss=0.1181, over 18108.00 frames. ], tot_loss[loss=0.2557, simple_loss=0.3252, pruned_loss=0.09314, over 3563685.04 frames. ], batch size: 62, lr: 2.66e-02, grad_scale: 4.0 +2023-03-08 17:41:24,476 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=12403.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:41:48,829 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.804e+02 4.633e+02 5.942e+02 7.166e+02 1.765e+03, threshold=1.188e+03, percent-clipped=4.0 +2023-03-08 17:41:59,112 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12432.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:42:15,027 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=12446.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:42:22,745 INFO [train.py:898] (0/4) Epoch 4, batch 1550, loss[loss=0.2213, simple_loss=0.2885, pruned_loss=0.07705, over 17705.00 frames. ], tot_loss[loss=0.2546, simple_loss=0.3245, pruned_loss=0.09235, over 3571859.91 frames. ], batch size: 39, lr: 2.66e-02, grad_scale: 4.0 +2023-03-08 17:42:45,802 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12472.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:43:11,435 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9839, 4.8342, 4.9773, 4.8622, 4.9464, 5.3929, 5.0427, 4.9077], + device='cuda:0'), covar=tensor([0.0488, 0.0466, 0.0491, 0.0353, 0.0833, 0.0505, 0.0417, 0.1051], + device='cuda:0'), in_proj_covar=tensor([0.0214, 0.0156, 0.0168, 0.0166, 0.0211, 0.0242, 0.0158, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 17:43:19,589 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12502.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:43:20,380 INFO [train.py:898] (0/4) Epoch 4, batch 1600, loss[loss=0.224, simple_loss=0.2864, pruned_loss=0.08084, over 18423.00 frames. ], tot_loss[loss=0.255, simple_loss=0.3249, pruned_loss=0.09256, over 3576119.09 frames. ], batch size: 43, lr: 2.65e-02, grad_scale: 8.0 +2023-03-08 17:43:25,783 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12507.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:43:41,716 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=12520.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:43:44,949 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.642e+02 4.900e+02 6.087e+02 7.463e+02 1.966e+03, threshold=1.217e+03, percent-clipped=9.0 +2023-03-08 17:44:15,882 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=12550.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:44:19,003 INFO [train.py:898] (0/4) Epoch 4, batch 1650, loss[loss=0.2901, simple_loss=0.3447, pruned_loss=0.1177, over 12392.00 frames. ], tot_loss[loss=0.2543, simple_loss=0.3247, pruned_loss=0.09194, over 3575427.12 frames. ], batch size: 129, lr: 2.65e-02, grad_scale: 8.0 +2023-03-08 17:44:33,496 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.62 vs. limit=2.0 +2023-03-08 17:44:48,155 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12576.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:44:58,390 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1665, 2.5118, 2.3602, 2.4912, 3.1408, 3.0890, 2.5560, 2.9385], + device='cuda:0'), covar=tensor([0.0292, 0.0431, 0.0920, 0.0484, 0.0306, 0.0233, 0.0604, 0.0284], + device='cuda:0'), in_proj_covar=tensor([0.0084, 0.0065, 0.0118, 0.0091, 0.0073, 0.0052, 0.0085, 0.0077], + device='cuda:0'), out_proj_covar=tensor([1.6321e-04, 1.2734e-04, 2.0850e-04, 1.6605e-04, 1.4045e-04, 9.4051e-05, + 1.5998e-04, 1.4596e-04], device='cuda:0') +2023-03-08 17:45:18,522 INFO [train.py:898] (0/4) Epoch 4, batch 1700, loss[loss=0.2317, simple_loss=0.2974, pruned_loss=0.08302, over 18383.00 frames. ], tot_loss[loss=0.2541, simple_loss=0.3248, pruned_loss=0.09173, over 3571087.34 frames. ], batch size: 42, lr: 2.65e-02, grad_scale: 8.0 +2023-03-08 17:45:19,896 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12604.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:45:43,559 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.156e+02 4.705e+02 5.777e+02 7.062e+02 1.643e+03, threshold=1.155e+03, percent-clipped=4.0 +2023-03-08 17:45:44,914 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=12624.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:45:52,906 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2850, 5.4091, 3.1306, 5.1133, 4.9878, 5.4327, 5.2306, 2.6760], + device='cuda:0'), covar=tensor([0.0137, 0.0043, 0.0631, 0.0071, 0.0072, 0.0048, 0.0090, 0.1075], + device='cuda:0'), in_proj_covar=tensor([0.0061, 0.0046, 0.0079, 0.0056, 0.0055, 0.0047, 0.0059, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0004, 0.0003, 0.0003, 0.0002, 0.0003, 0.0004], + device='cuda:0') +2023-03-08 17:46:02,680 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12640.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:46:15,921 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=12652.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:46:16,914 INFO [train.py:898] (0/4) Epoch 4, batch 1750, loss[loss=0.1826, simple_loss=0.2627, pruned_loss=0.05122, over 18498.00 frames. ], tot_loss[loss=0.2537, simple_loss=0.3242, pruned_loss=0.09157, over 3575711.99 frames. ], batch size: 44, lr: 2.64e-02, grad_scale: 8.0 +2023-03-08 17:46:33,700 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7801, 1.6258, 4.0627, 2.9747, 3.7444, 5.0191, 4.4881, 4.5223], + device='cuda:0'), covar=tensor([0.0208, 0.0837, 0.0243, 0.0466, 0.0785, 0.0026, 0.0164, 0.0124], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0162, 0.0096, 0.0158, 0.0246, 0.0091, 0.0134, 0.0123], + device='cuda:0'), out_proj_covar=tensor([9.9234e-05, 1.3031e-04, 8.4416e-05, 1.1744e-04, 1.9498e-04, 6.5743e-05, + 1.0883e-04, 9.7987e-05], device='cuda:0') +2023-03-08 17:46:58,851 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=12688.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:47:11,444 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1354, 5.7885, 5.2232, 5.5131, 5.2288, 5.3908, 5.8906, 5.8422], + device='cuda:0'), covar=tensor([0.1249, 0.0586, 0.0433, 0.0671, 0.1516, 0.0590, 0.0481, 0.0503], + device='cuda:0'), in_proj_covar=tensor([0.0353, 0.0299, 0.0226, 0.0315, 0.0448, 0.0312, 0.0346, 0.0293], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0005, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-08 17:47:15,807 INFO [train.py:898] (0/4) Epoch 4, batch 1800, loss[loss=0.2592, simple_loss=0.3356, pruned_loss=0.09143, over 18382.00 frames. ], tot_loss[loss=0.2523, simple_loss=0.3232, pruned_loss=0.09071, over 3570531.17 frames. ], batch size: 50, lr: 2.64e-02, grad_scale: 8.0 +2023-03-08 17:47:39,918 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.294e+02 5.100e+02 5.998e+02 7.451e+02 2.129e+03, threshold=1.200e+03, percent-clipped=3.0 +2023-03-08 17:47:51,185 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=12732.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:48:05,069 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2187, 3.2441, 1.4747, 3.9421, 2.6292, 4.2035, 1.7488, 3.5763], + device='cuda:0'), covar=tensor([0.0611, 0.0883, 0.1837, 0.0336, 0.1104, 0.0106, 0.1556, 0.0370], + device='cuda:0'), in_proj_covar=tensor([0.0140, 0.0179, 0.0159, 0.0135, 0.0154, 0.0093, 0.0161, 0.0143], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 17:48:15,105 INFO [train.py:898] (0/4) Epoch 4, batch 1850, loss[loss=0.2344, simple_loss=0.3165, pruned_loss=0.07614, over 18575.00 frames. ], tot_loss[loss=0.2509, simple_loss=0.3223, pruned_loss=0.08978, over 3582124.31 frames. ], batch size: 54, lr: 2.63e-02, grad_scale: 8.0 +2023-03-08 17:48:26,829 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=12763.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:48:47,864 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=12780.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:49:08,451 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.57 vs. limit=2.0 +2023-03-08 17:49:12,589 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=12802.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:49:13,576 INFO [train.py:898] (0/4) Epoch 4, batch 1900, loss[loss=0.2413, simple_loss=0.3152, pruned_loss=0.08372, over 18397.00 frames. ], tot_loss[loss=0.2525, simple_loss=0.3237, pruned_loss=0.09061, over 3588835.56 frames. ], batch size: 48, lr: 2.63e-02, grad_scale: 8.0 +2023-03-08 17:49:24,085 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3199, 5.3907, 3.2514, 5.1680, 4.9577, 5.3674, 5.1529, 2.6044], + device='cuda:0'), covar=tensor([0.0147, 0.0063, 0.0646, 0.0058, 0.0091, 0.0089, 0.0140, 0.1093], + device='cuda:0'), in_proj_covar=tensor([0.0062, 0.0046, 0.0080, 0.0056, 0.0056, 0.0048, 0.0060, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0004, 0.0003, 0.0003, 0.0002, 0.0003, 0.0004], + device='cuda:0') +2023-03-08 17:49:36,777 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.387e+02 4.784e+02 5.928e+02 6.975e+02 1.301e+03, threshold=1.186e+03, percent-clipped=1.0 +2023-03-08 17:49:37,619 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-03-08 17:49:38,282 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=12824.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 17:50:11,762 INFO [train.py:898] (0/4) Epoch 4, batch 1950, loss[loss=0.2495, simple_loss=0.3275, pruned_loss=0.08576, over 18334.00 frames. ], tot_loss[loss=0.2512, simple_loss=0.3224, pruned_loss=0.08999, over 3595484.67 frames. ], batch size: 56, lr: 2.62e-02, grad_scale: 8.0 +2023-03-08 17:50:20,199 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.60 vs. limit=2.0 +2023-03-08 17:51:10,761 INFO [train.py:898] (0/4) Epoch 4, batch 2000, loss[loss=0.2383, simple_loss=0.3097, pruned_loss=0.08346, over 18348.00 frames. ], tot_loss[loss=0.2513, simple_loss=0.3226, pruned_loss=0.08996, over 3599293.32 frames. ], batch size: 46, lr: 2.62e-02, grad_scale: 8.0 +2023-03-08 17:51:33,456 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.610e+02 4.786e+02 5.613e+02 6.495e+02 1.703e+03, threshold=1.123e+03, percent-clipped=3.0 +2023-03-08 17:52:08,660 INFO [train.py:898] (0/4) Epoch 4, batch 2050, loss[loss=0.2096, simple_loss=0.2803, pruned_loss=0.06944, over 18447.00 frames. ], tot_loss[loss=0.2517, simple_loss=0.323, pruned_loss=0.09021, over 3604163.96 frames. ], batch size: 43, lr: 2.61e-02, grad_scale: 8.0 +2023-03-08 17:53:07,613 INFO [train.py:898] (0/4) Epoch 4, batch 2100, loss[loss=0.2728, simple_loss=0.3496, pruned_loss=0.098, over 18571.00 frames. ], tot_loss[loss=0.2533, simple_loss=0.3244, pruned_loss=0.09113, over 3577565.04 frames. ], batch size: 54, lr: 2.61e-02, grad_scale: 8.0 +2023-03-08 17:53:29,878 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.835e+02 4.911e+02 6.247e+02 7.223e+02 1.206e+03, threshold=1.249e+03, percent-clipped=2.0 +2023-03-08 17:54:06,017 INFO [train.py:898] (0/4) Epoch 4, batch 2150, loss[loss=0.1911, simple_loss=0.2705, pruned_loss=0.05584, over 18548.00 frames. ], tot_loss[loss=0.2523, simple_loss=0.3234, pruned_loss=0.09065, over 3580971.04 frames. ], batch size: 45, lr: 2.61e-02, grad_scale: 8.0 +2023-03-08 17:55:03,485 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=13102.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:55:03,560 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4462, 3.3442, 1.8902, 4.3228, 2.7386, 4.5611, 2.1842, 4.1219], + device='cuda:0'), covar=tensor([0.0464, 0.0811, 0.1665, 0.0297, 0.0990, 0.0095, 0.1183, 0.0227], + device='cuda:0'), in_proj_covar=tensor([0.0141, 0.0180, 0.0159, 0.0138, 0.0157, 0.0097, 0.0163, 0.0144], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 17:55:04,392 INFO [train.py:898] (0/4) Epoch 4, batch 2200, loss[loss=0.2248, simple_loss=0.3021, pruned_loss=0.07379, over 18430.00 frames. ], tot_loss[loss=0.2525, simple_loss=0.3234, pruned_loss=0.09079, over 3580208.61 frames. ], batch size: 48, lr: 2.60e-02, grad_scale: 8.0 +2023-03-08 17:55:22,889 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=13119.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 17:55:27,147 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.452e+02 4.789e+02 5.834e+02 7.042e+02 2.257e+03, threshold=1.167e+03, percent-clipped=4.0 +2023-03-08 17:55:59,415 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=13150.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 17:56:02,802 INFO [train.py:898] (0/4) Epoch 4, batch 2250, loss[loss=0.2275, simple_loss=0.3025, pruned_loss=0.07629, over 18352.00 frames. ], tot_loss[loss=0.2521, simple_loss=0.3232, pruned_loss=0.09045, over 3593710.40 frames. ], batch size: 46, lr: 2.60e-02, grad_scale: 8.0 +2023-03-08 17:57:01,372 INFO [train.py:898] (0/4) Epoch 4, batch 2300, loss[loss=0.2175, simple_loss=0.2862, pruned_loss=0.07442, over 18542.00 frames. ], tot_loss[loss=0.2522, simple_loss=0.3235, pruned_loss=0.0905, over 3605087.69 frames. ], batch size: 45, lr: 2.59e-02, grad_scale: 8.0 +2023-03-08 17:57:24,770 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.647e+02 4.949e+02 6.048e+02 7.391e+02 1.554e+03, threshold=1.210e+03, percent-clipped=6.0 +2023-03-08 17:57:33,172 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9627, 2.9760, 1.8480, 3.3206, 2.3754, 3.3817, 2.0473, 2.7947], + device='cuda:0'), covar=tensor([0.0397, 0.0603, 0.1208, 0.0410, 0.0803, 0.0139, 0.1004, 0.0375], + device='cuda:0'), in_proj_covar=tensor([0.0143, 0.0179, 0.0161, 0.0141, 0.0157, 0.0098, 0.0161, 0.0147], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 17:58:00,615 INFO [train.py:898] (0/4) Epoch 4, batch 2350, loss[loss=0.2535, simple_loss=0.3281, pruned_loss=0.08947, over 17096.00 frames. ], tot_loss[loss=0.2517, simple_loss=0.3227, pruned_loss=0.09035, over 3592074.66 frames. ], batch size: 78, lr: 2.59e-02, grad_scale: 8.0 +2023-03-08 17:58:58,984 INFO [train.py:898] (0/4) Epoch 4, batch 2400, loss[loss=0.3464, simple_loss=0.386, pruned_loss=0.1534, over 12251.00 frames. ], tot_loss[loss=0.2511, simple_loss=0.3221, pruned_loss=0.0901, over 3584437.43 frames. ], batch size: 129, lr: 2.58e-02, grad_scale: 8.0 +2023-03-08 17:59:23,023 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.316e+02 4.704e+02 5.884e+02 7.120e+02 1.441e+03, threshold=1.177e+03, percent-clipped=2.0 +2023-03-08 17:59:58,170 INFO [train.py:898] (0/4) Epoch 4, batch 2450, loss[loss=0.2622, simple_loss=0.3398, pruned_loss=0.09236, over 18122.00 frames. ], tot_loss[loss=0.2522, simple_loss=0.3232, pruned_loss=0.0906, over 3573941.39 frames. ], batch size: 62, lr: 2.58e-02, grad_scale: 8.0 +2023-03-08 18:00:56,432 INFO [train.py:898] (0/4) Epoch 4, batch 2500, loss[loss=0.2277, simple_loss=0.2963, pruned_loss=0.07951, over 18344.00 frames. ], tot_loss[loss=0.2508, simple_loss=0.3217, pruned_loss=0.08999, over 3577642.04 frames. ], batch size: 46, lr: 2.58e-02, grad_scale: 8.0 +2023-03-08 18:01:16,540 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=13419.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:01:20,736 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.373e+02 5.229e+02 6.294e+02 7.709e+02 1.447e+03, threshold=1.259e+03, percent-clipped=4.0 +2023-03-08 18:01:55,117 INFO [train.py:898] (0/4) Epoch 4, batch 2550, loss[loss=0.2104, simple_loss=0.2819, pruned_loss=0.06952, over 18393.00 frames. ], tot_loss[loss=0.2501, simple_loss=0.3212, pruned_loss=0.08949, over 3591254.92 frames. ], batch size: 48, lr: 2.57e-02, grad_scale: 8.0 +2023-03-08 18:02:12,159 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=13467.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:02:19,678 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0664, 1.7813, 3.6620, 3.2204, 3.6972, 4.8727, 4.3335, 4.4233], + device='cuda:0'), covar=tensor([0.0306, 0.0821, 0.0377, 0.0475, 0.0826, 0.0031, 0.0209, 0.0124], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0170, 0.0110, 0.0165, 0.0255, 0.0095, 0.0140, 0.0126], + device='cuda:0'), out_proj_covar=tensor([1.0319e-04, 1.3490e-04, 9.5021e-05, 1.2028e-04, 1.9926e-04, 6.8741e-05, + 1.1146e-04, 9.9101e-05], device='cuda:0') +2023-03-08 18:02:53,611 INFO [train.py:898] (0/4) Epoch 4, batch 2600, loss[loss=0.2696, simple_loss=0.3442, pruned_loss=0.09747, over 18256.00 frames. ], tot_loss[loss=0.2513, simple_loss=0.3222, pruned_loss=0.09018, over 3584508.27 frames. ], batch size: 57, lr: 2.57e-02, grad_scale: 8.0 +2023-03-08 18:03:00,257 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-03-08 18:03:17,315 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.093e+02 4.722e+02 5.886e+02 7.226e+02 1.343e+03, threshold=1.177e+03, percent-clipped=1.0 +2023-03-08 18:03:44,124 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3549, 5.1304, 5.2987, 5.2372, 5.1557, 5.8777, 5.5015, 5.3210], + device='cuda:0'), covar=tensor([0.0696, 0.0545, 0.0623, 0.0480, 0.1286, 0.0710, 0.0507, 0.1519], + device='cuda:0'), in_proj_covar=tensor([0.0226, 0.0171, 0.0177, 0.0170, 0.0220, 0.0255, 0.0164, 0.0245], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 18:03:51,833 INFO [train.py:898] (0/4) Epoch 4, batch 2650, loss[loss=0.284, simple_loss=0.3456, pruned_loss=0.1112, over 18363.00 frames. ], tot_loss[loss=0.2509, simple_loss=0.3219, pruned_loss=0.08998, over 3584785.35 frames. ], batch size: 56, lr: 2.56e-02, grad_scale: 8.0 +2023-03-08 18:04:28,235 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7829, 4.3735, 4.5779, 3.3141, 3.5598, 3.3228, 2.2615, 1.5539], + device='cuda:0'), covar=tensor([0.0149, 0.0226, 0.0040, 0.0233, 0.0262, 0.0163, 0.0780, 0.1073], + device='cuda:0'), in_proj_covar=tensor([0.0034, 0.0035, 0.0028, 0.0040, 0.0056, 0.0031, 0.0058, 0.0062], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 18:04:38,228 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4511, 6.0632, 5.4416, 5.7890, 5.5587, 5.6405, 6.1144, 5.9878], + device='cuda:0'), covar=tensor([0.0978, 0.0545, 0.0371, 0.0592, 0.1437, 0.0567, 0.0404, 0.0596], + device='cuda:0'), in_proj_covar=tensor([0.0369, 0.0314, 0.0237, 0.0325, 0.0465, 0.0334, 0.0356, 0.0304], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-08 18:04:39,975 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=5.00 vs. limit=5.0 +2023-03-08 18:04:50,558 INFO [train.py:898] (0/4) Epoch 4, batch 2700, loss[loss=0.2584, simple_loss=0.3342, pruned_loss=0.09134, over 18261.00 frames. ], tot_loss[loss=0.249, simple_loss=0.3206, pruned_loss=0.08872, over 3594655.96 frames. ], batch size: 57, lr: 2.56e-02, grad_scale: 8.0 +2023-03-08 18:05:14,649 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.306e+02 5.163e+02 6.316e+02 7.971e+02 1.590e+03, threshold=1.263e+03, percent-clipped=4.0 +2023-03-08 18:05:48,660 INFO [train.py:898] (0/4) Epoch 4, batch 2750, loss[loss=0.2731, simple_loss=0.3436, pruned_loss=0.1013, over 18102.00 frames. ], tot_loss[loss=0.2487, simple_loss=0.3202, pruned_loss=0.08863, over 3602762.12 frames. ], batch size: 62, lr: 2.55e-02, grad_scale: 8.0 +2023-03-08 18:06:47,203 INFO [train.py:898] (0/4) Epoch 4, batch 2800, loss[loss=0.1923, simple_loss=0.2629, pruned_loss=0.0609, over 18437.00 frames. ], tot_loss[loss=0.2493, simple_loss=0.3209, pruned_loss=0.08884, over 3600386.01 frames. ], batch size: 43, lr: 2.55e-02, grad_scale: 8.0 +2023-03-08 18:07:11,543 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.506e+02 4.982e+02 6.101e+02 7.616e+02 1.473e+03, threshold=1.220e+03, percent-clipped=5.0 +2023-03-08 18:07:28,123 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=13737.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:07:36,335 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-03-08 18:07:45,676 INFO [train.py:898] (0/4) Epoch 4, batch 2850, loss[loss=0.2205, simple_loss=0.2958, pruned_loss=0.07263, over 18155.00 frames. ], tot_loss[loss=0.2493, simple_loss=0.3205, pruned_loss=0.089, over 3592243.99 frames. ], batch size: 44, lr: 2.55e-02, grad_scale: 8.0 +2023-03-08 18:08:22,256 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8099, 4.7108, 4.8435, 4.6844, 4.6038, 4.6632, 4.9816, 5.0831], + device='cuda:0'), covar=tensor([0.0068, 0.0073, 0.0070, 0.0083, 0.0100, 0.0111, 0.0110, 0.0099], + device='cuda:0'), in_proj_covar=tensor([0.0063, 0.0047, 0.0046, 0.0061, 0.0053, 0.0068, 0.0056, 0.0056], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 18:08:38,518 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=13798.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 18:08:43,997 INFO [train.py:898] (0/4) Epoch 4, batch 2900, loss[loss=0.2291, simple_loss=0.292, pruned_loss=0.08309, over 18469.00 frames. ], tot_loss[loss=0.2509, simple_loss=0.322, pruned_loss=0.08986, over 3586453.78 frames. ], batch size: 44, lr: 2.54e-02, grad_scale: 8.0 +2023-03-08 18:09:07,286 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.361e+02 5.191e+02 6.530e+02 8.692e+02 2.187e+03, threshold=1.306e+03, percent-clipped=7.0 +2023-03-08 18:09:17,808 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1690, 4.8355, 4.8242, 4.6281, 4.4495, 4.6262, 4.0421, 4.6833], + device='cuda:0'), covar=tensor([0.0297, 0.0306, 0.0281, 0.0291, 0.0446, 0.0307, 0.1187, 0.0262], + device='cuda:0'), in_proj_covar=tensor([0.0119, 0.0157, 0.0140, 0.0128, 0.0152, 0.0158, 0.0226, 0.0142], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0005, 0.0004], + device='cuda:0') +2023-03-08 18:09:43,213 INFO [train.py:898] (0/4) Epoch 4, batch 2950, loss[loss=0.3004, simple_loss=0.3593, pruned_loss=0.1207, over 17039.00 frames. ], tot_loss[loss=0.2503, simple_loss=0.3217, pruned_loss=0.08944, over 3581780.73 frames. ], batch size: 78, lr: 2.54e-02, grad_scale: 8.0 +2023-03-08 18:09:49,055 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3566, 5.2954, 4.8225, 5.2969, 5.2137, 4.8446, 5.2625, 4.8325], + device='cuda:0'), covar=tensor([0.0272, 0.0349, 0.1340, 0.0493, 0.0390, 0.0317, 0.0315, 0.0621], + device='cuda:0'), in_proj_covar=tensor([0.0271, 0.0301, 0.0456, 0.0238, 0.0228, 0.0276, 0.0299, 0.0365], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0003, 0.0003, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 18:09:56,283 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.66 vs. limit=2.0 +2023-03-08 18:10:40,648 INFO [train.py:898] (0/4) Epoch 4, batch 3000, loss[loss=0.2149, simple_loss=0.2885, pruned_loss=0.0706, over 18263.00 frames. ], tot_loss[loss=0.2501, simple_loss=0.3213, pruned_loss=0.08945, over 3585021.75 frames. ], batch size: 45, lr: 2.53e-02, grad_scale: 4.0 +2023-03-08 18:10:40,650 INFO [train.py:923] (0/4) Computing validation loss +2023-03-08 18:10:52,599 INFO [train.py:932] (0/4) Epoch 4, validation: loss=0.1898, simple_loss=0.292, pruned_loss=0.04378, over 944034.00 frames. +2023-03-08 18:10:52,600 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-08 18:10:56,234 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.68 vs. limit=5.0 +2023-03-08 18:11:00,593 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.3053, 4.0409, 5.2800, 4.3662, 3.1309, 2.8919, 4.5048, 5.2929], + device='cuda:0'), covar=tensor([0.0970, 0.0980, 0.0032, 0.0237, 0.0808, 0.0955, 0.0216, 0.0026], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0144, 0.0060, 0.0125, 0.0153, 0.0156, 0.0127, 0.0065], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-08 18:11:17,314 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.366e+02 5.126e+02 6.271e+02 8.521e+02 2.590e+03, threshold=1.254e+03, percent-clipped=11.0 +2023-03-08 18:11:50,803 INFO [train.py:898] (0/4) Epoch 4, batch 3050, loss[loss=0.2396, simple_loss=0.3137, pruned_loss=0.08272, over 18308.00 frames. ], tot_loss[loss=0.2487, simple_loss=0.32, pruned_loss=0.08871, over 3589210.76 frames. ], batch size: 49, lr: 2.53e-02, grad_scale: 4.0 +2023-03-08 18:12:46,235 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-14000.pt +2023-03-08 18:12:53,755 INFO [train.py:898] (0/4) Epoch 4, batch 3100, loss[loss=0.254, simple_loss=0.3293, pruned_loss=0.08939, over 18499.00 frames. ], tot_loss[loss=0.2485, simple_loss=0.3198, pruned_loss=0.08857, over 3594170.81 frames. ], batch size: 53, lr: 2.53e-02, grad_scale: 4.0 +2023-03-08 18:13:18,982 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.897e+02 4.672e+02 5.881e+02 7.283e+02 1.291e+03, threshold=1.176e+03, percent-clipped=1.0 +2023-03-08 18:13:52,240 INFO [train.py:898] (0/4) Epoch 4, batch 3150, loss[loss=0.3222, simple_loss=0.3695, pruned_loss=0.1374, over 12311.00 frames. ], tot_loss[loss=0.2491, simple_loss=0.3204, pruned_loss=0.08889, over 3583204.56 frames. ], batch size: 129, lr: 2.52e-02, grad_scale: 4.0 +2023-03-08 18:13:54,979 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6872, 4.0374, 5.3076, 4.5493, 3.6543, 2.8549, 5.0004, 5.3066], + device='cuda:0'), covar=tensor([0.0865, 0.0934, 0.0049, 0.0196, 0.0627, 0.0938, 0.0148, 0.0042], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0144, 0.0062, 0.0126, 0.0152, 0.0159, 0.0127, 0.0066], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0003, 0.0002, 0.0001], + device='cuda:0') +2023-03-08 18:14:15,426 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14072.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:14:40,066 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14093.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 18:14:51,080 INFO [train.py:898] (0/4) Epoch 4, batch 3200, loss[loss=0.2489, simple_loss=0.3203, pruned_loss=0.08876, over 18557.00 frames. ], tot_loss[loss=0.2489, simple_loss=0.3202, pruned_loss=0.08875, over 3595962.97 frames. ], batch size: 54, lr: 2.52e-02, grad_scale: 8.0 +2023-03-08 18:14:51,371 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14103.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:15:10,605 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14120.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 18:15:14,749 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.167e+02 5.012e+02 5.863e+02 7.076e+02 1.938e+03, threshold=1.173e+03, percent-clipped=6.0 +2023-03-08 18:15:26,328 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14133.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:15:29,188 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-08 18:15:49,264 INFO [train.py:898] (0/4) Epoch 4, batch 3250, loss[loss=0.2167, simple_loss=0.2869, pruned_loss=0.07325, over 18349.00 frames. ], tot_loss[loss=0.247, simple_loss=0.3188, pruned_loss=0.0876, over 3605069.78 frames. ], batch size: 46, lr: 2.51e-02, grad_scale: 8.0 +2023-03-08 18:15:52,839 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0711, 5.4586, 3.0737, 5.2083, 5.0677, 5.4184, 5.1232, 2.5265], + device='cuda:0'), covar=tensor([0.0166, 0.0038, 0.0586, 0.0059, 0.0054, 0.0053, 0.0096, 0.1050], + device='cuda:0'), in_proj_covar=tensor([0.0061, 0.0048, 0.0080, 0.0056, 0.0055, 0.0047, 0.0060, 0.0084], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0004, 0.0003, 0.0003, 0.0002, 0.0003, 0.0004], + device='cuda:0') +2023-03-08 18:15:53,938 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2546, 4.9143, 5.2759, 5.0097, 5.1417, 5.8869, 5.5234, 5.2522], + device='cuda:0'), covar=tensor([0.0872, 0.0631, 0.0662, 0.0594, 0.1218, 0.0676, 0.0525, 0.1445], + device='cuda:0'), in_proj_covar=tensor([0.0227, 0.0171, 0.0177, 0.0170, 0.0220, 0.0253, 0.0162, 0.0242], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 18:16:01,692 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14164.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:16:22,317 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14181.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 18:16:47,517 INFO [train.py:898] (0/4) Epoch 4, batch 3300, loss[loss=0.2433, simple_loss=0.3215, pruned_loss=0.08258, over 18404.00 frames. ], tot_loss[loss=0.2478, simple_loss=0.3198, pruned_loss=0.08793, over 3612017.45 frames. ], batch size: 52, lr: 2.51e-02, grad_scale: 8.0 +2023-03-08 18:17:10,861 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.117e+02 4.913e+02 6.024e+02 7.201e+02 1.529e+03, threshold=1.205e+03, percent-clipped=5.0 +2023-03-08 18:17:14,146 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-08 18:17:45,354 INFO [train.py:898] (0/4) Epoch 4, batch 3350, loss[loss=0.2374, simple_loss=0.3163, pruned_loss=0.0793, over 18316.00 frames. ], tot_loss[loss=0.2471, simple_loss=0.319, pruned_loss=0.08761, over 3607374.87 frames. ], batch size: 54, lr: 2.51e-02, grad_scale: 8.0 +2023-03-08 18:17:52,974 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1898, 4.9828, 5.2680, 5.1943, 5.0080, 5.8323, 5.4067, 5.2207], + device='cuda:0'), covar=tensor([0.0787, 0.0594, 0.0571, 0.0528, 0.1436, 0.0698, 0.0683, 0.1407], + device='cuda:0'), in_proj_covar=tensor([0.0233, 0.0175, 0.0178, 0.0173, 0.0223, 0.0254, 0.0166, 0.0244], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 18:18:37,177 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0811, 1.9603, 3.5723, 2.7626, 3.3939, 4.7493, 4.0599, 3.9957], + device='cuda:0'), covar=tensor([0.0273, 0.0812, 0.0384, 0.0523, 0.0789, 0.0034, 0.0178, 0.0144], + device='cuda:0'), in_proj_covar=tensor([0.0137, 0.0176, 0.0121, 0.0172, 0.0256, 0.0095, 0.0146, 0.0127], + device='cuda:0'), out_proj_covar=tensor([1.0713e-04, 1.3852e-04, 1.0388e-04, 1.2358e-04, 1.9746e-04, 6.8098e-05, + 1.1397e-04, 9.8363e-05], device='cuda:0') +2023-03-08 18:18:44,673 INFO [train.py:898] (0/4) Epoch 4, batch 3400, loss[loss=0.2427, simple_loss=0.3207, pruned_loss=0.08234, over 18292.00 frames. ], tot_loss[loss=0.2474, simple_loss=0.3196, pruned_loss=0.08757, over 3603051.19 frames. ], batch size: 57, lr: 2.50e-02, grad_scale: 8.0 +2023-03-08 18:19:08,456 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.677e+02 5.094e+02 5.876e+02 7.439e+02 2.634e+03, threshold=1.175e+03, percent-clipped=7.0 +2023-03-08 18:19:42,054 INFO [train.py:898] (0/4) Epoch 4, batch 3450, loss[loss=0.2388, simple_loss=0.3071, pruned_loss=0.08527, over 18504.00 frames. ], tot_loss[loss=0.2473, simple_loss=0.3192, pruned_loss=0.08767, over 3594884.53 frames. ], batch size: 47, lr: 2.50e-02, grad_scale: 4.0 +2023-03-08 18:20:29,205 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14393.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:20:38,213 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-08 18:20:40,688 INFO [train.py:898] (0/4) Epoch 4, batch 3500, loss[loss=0.2174, simple_loss=0.2882, pruned_loss=0.07335, over 18440.00 frames. ], tot_loss[loss=0.2472, simple_loss=0.3189, pruned_loss=0.08773, over 3599960.77 frames. ], batch size: 42, lr: 2.49e-02, grad_scale: 4.0 +2023-03-08 18:20:52,777 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14413.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:21:05,562 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.205e+02 4.879e+02 5.762e+02 7.140e+02 1.904e+03, threshold=1.152e+03, percent-clipped=3.0 +2023-03-08 18:21:09,070 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14428.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:21:22,859 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=14441.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:21:25,119 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9737, 5.2070, 2.5237, 4.9445, 4.9235, 5.2793, 5.0242, 2.5530], + device='cuda:0'), covar=tensor([0.0158, 0.0053, 0.0726, 0.0076, 0.0071, 0.0062, 0.0096, 0.1010], + device='cuda:0'), in_proj_covar=tensor([0.0061, 0.0049, 0.0079, 0.0057, 0.0056, 0.0048, 0.0062, 0.0086], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0004, 0.0003, 0.0003, 0.0002, 0.0003, 0.0004], + device='cuda:0') +2023-03-08 18:21:35,660 INFO [train.py:898] (0/4) Epoch 4, batch 3550, loss[loss=0.2531, simple_loss=0.3318, pruned_loss=0.08717, over 18566.00 frames. ], tot_loss[loss=0.2461, simple_loss=0.3181, pruned_loss=0.08705, over 3601970.66 frames. ], batch size: 54, lr: 2.49e-02, grad_scale: 4.0 +2023-03-08 18:21:42,347 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14459.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:21:58,273 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4421, 3.2074, 1.9403, 4.1414, 2.7551, 4.2456, 2.1778, 3.7668], + device='cuda:0'), covar=tensor([0.0446, 0.0825, 0.1243, 0.0257, 0.0880, 0.0100, 0.0984, 0.0248], + device='cuda:0'), in_proj_covar=tensor([0.0142, 0.0180, 0.0157, 0.0141, 0.0153, 0.0101, 0.0159, 0.0146], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 18:21:58,299 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14474.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:22:00,163 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14476.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 18:22:13,083 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14487.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:22:30,849 INFO [train.py:898] (0/4) Epoch 4, batch 3600, loss[loss=0.2717, simple_loss=0.3445, pruned_loss=0.09946, over 17744.00 frames. ], tot_loss[loss=0.2471, simple_loss=0.3189, pruned_loss=0.08762, over 3600275.79 frames. ], batch size: 70, lr: 2.49e-02, grad_scale: 8.0 +2023-03-08 18:22:53,543 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.017e+02 5.164e+02 6.546e+02 8.016e+02 1.916e+03, threshold=1.309e+03, percent-clipped=5.0 +2023-03-08 18:23:05,908 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-4.pt +2023-03-08 18:23:35,182 INFO [train.py:898] (0/4) Epoch 5, batch 0, loss[loss=0.2867, simple_loss=0.364, pruned_loss=0.1047, over 18500.00 frames. ], tot_loss[loss=0.2867, simple_loss=0.364, pruned_loss=0.1047, over 18500.00 frames. ], batch size: 59, lr: 2.31e-02, grad_scale: 8.0 +2023-03-08 18:23:35,184 INFO [train.py:923] (0/4) Computing validation loss +2023-03-08 18:23:46,758 INFO [train.py:932] (0/4) Epoch 5, validation: loss=0.1908, simple_loss=0.2926, pruned_loss=0.04454, over 944034.00 frames. +2023-03-08 18:23:46,759 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-08 18:23:47,086 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5031, 5.2337, 5.1959, 5.0408, 4.8937, 5.1125, 4.4209, 4.9293], + device='cuda:0'), covar=tensor([0.0297, 0.0262, 0.0193, 0.0178, 0.0357, 0.0217, 0.1144, 0.0288], + device='cuda:0'), in_proj_covar=tensor([0.0116, 0.0152, 0.0134, 0.0124, 0.0145, 0.0149, 0.0218, 0.0136], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0004, 0.0004, 0.0004, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-08 18:23:59,379 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14548.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:24:44,283 INFO [train.py:898] (0/4) Epoch 5, batch 50, loss[loss=0.249, simple_loss=0.3237, pruned_loss=0.08712, over 17956.00 frames. ], tot_loss[loss=0.245, simple_loss=0.3176, pruned_loss=0.08617, over 814599.53 frames. ], batch size: 65, lr: 2.31e-02, grad_scale: 8.0 +2023-03-08 18:25:29,118 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.345e+02 4.923e+02 5.805e+02 7.432e+02 1.503e+03, threshold=1.161e+03, percent-clipped=2.0 +2023-03-08 18:25:43,377 INFO [train.py:898] (0/4) Epoch 5, batch 100, loss[loss=0.241, simple_loss=0.3093, pruned_loss=0.08637, over 18521.00 frames. ], tot_loss[loss=0.2425, simple_loss=0.3154, pruned_loss=0.08482, over 1427216.34 frames. ], batch size: 49, lr: 2.31e-02, grad_scale: 8.0 +2023-03-08 18:25:57,208 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7667, 2.7801, 4.2218, 4.3289, 2.4652, 4.5851, 3.7975, 2.5719], + device='cuda:0'), covar=tensor([0.0223, 0.0858, 0.0104, 0.0107, 0.1200, 0.0091, 0.0265, 0.0910], + device='cuda:0'), in_proj_covar=tensor([0.0128, 0.0160, 0.0089, 0.0101, 0.0175, 0.0119, 0.0126, 0.0163], + device='cuda:0'), out_proj_covar=tensor([1.2502e-04, 1.5552e-04, 9.3685e-05, 9.6385e-05, 1.6766e-04, 1.1305e-04, + 1.3174e-04, 1.6368e-04], device='cuda:0') +2023-03-08 18:26:14,324 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4915, 3.4139, 1.8345, 4.0408, 2.8534, 4.3801, 1.8088, 3.8414], + device='cuda:0'), covar=tensor([0.0486, 0.0756, 0.1480, 0.0383, 0.0901, 0.0082, 0.1332, 0.0298], + device='cuda:0'), in_proj_covar=tensor([0.0143, 0.0182, 0.0160, 0.0145, 0.0154, 0.0104, 0.0161, 0.0150], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 18:26:19,427 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3362, 5.3119, 4.3679, 5.1994, 5.1815, 4.8054, 5.1725, 4.6465], + device='cuda:0'), covar=tensor([0.0486, 0.0509, 0.2173, 0.0748, 0.0619, 0.0475, 0.0496, 0.0832], + device='cuda:0'), in_proj_covar=tensor([0.0283, 0.0307, 0.0474, 0.0254, 0.0234, 0.0294, 0.0311, 0.0384], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0006, 0.0004, 0.0003, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 18:26:41,845 INFO [train.py:898] (0/4) Epoch 5, batch 150, loss[loss=0.2821, simple_loss=0.346, pruned_loss=0.1091, over 17904.00 frames. ], tot_loss[loss=0.2444, simple_loss=0.317, pruned_loss=0.08595, over 1903025.49 frames. ], batch size: 65, lr: 2.30e-02, grad_scale: 8.0 +2023-03-08 18:27:26,966 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.743e+02 4.388e+02 5.420e+02 7.125e+02 1.692e+03, threshold=1.084e+03, percent-clipped=3.0 +2023-03-08 18:27:30,697 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14728.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:27:31,946 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6790, 3.7471, 2.0444, 4.4642, 3.2034, 4.7562, 2.0464, 4.3810], + device='cuda:0'), covar=tensor([0.0450, 0.0813, 0.1560, 0.0293, 0.0834, 0.0112, 0.1314, 0.0245], + device='cuda:0'), in_proj_covar=tensor([0.0149, 0.0187, 0.0163, 0.0147, 0.0157, 0.0108, 0.0165, 0.0152], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 18:27:40,875 INFO [train.py:898] (0/4) Epoch 5, batch 200, loss[loss=0.2172, simple_loss=0.3061, pruned_loss=0.06417, over 18325.00 frames. ], tot_loss[loss=0.2438, simple_loss=0.3162, pruned_loss=0.08571, over 2269894.12 frames. ], batch size: 54, lr: 2.30e-02, grad_scale: 8.0 +2023-03-08 18:27:45,218 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=14740.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:28:06,749 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14759.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:28:17,858 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14769.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:28:26,989 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=14776.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:28:27,176 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=14776.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 18:28:34,752 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.66 vs. limit=5.0 +2023-03-08 18:28:39,575 INFO [train.py:898] (0/4) Epoch 5, batch 250, loss[loss=0.2495, simple_loss=0.3167, pruned_loss=0.09114, over 18532.00 frames. ], tot_loss[loss=0.2428, simple_loss=0.3156, pruned_loss=0.08502, over 2553987.32 frames. ], batch size: 49, lr: 2.30e-02, grad_scale: 8.0 +2023-03-08 18:28:56,720 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=14801.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:29:03,174 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=14807.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:29:21,927 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=14824.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 18:29:23,285 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.926e+02 4.780e+02 5.936e+02 7.141e+02 1.719e+03, threshold=1.187e+03, percent-clipped=5.0 +2023-03-08 18:29:37,970 INFO [train.py:898] (0/4) Epoch 5, batch 300, loss[loss=0.2209, simple_loss=0.298, pruned_loss=0.07188, over 18259.00 frames. ], tot_loss[loss=0.2412, simple_loss=0.3143, pruned_loss=0.08403, over 2795532.73 frames. ], batch size: 47, lr: 2.29e-02, grad_scale: 8.0 +2023-03-08 18:29:45,044 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=14843.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:30:35,693 INFO [train.py:898] (0/4) Epoch 5, batch 350, loss[loss=0.2436, simple_loss=0.329, pruned_loss=0.0791, over 18315.00 frames. ], tot_loss[loss=0.2405, simple_loss=0.3135, pruned_loss=0.08369, over 2979681.76 frames. ], batch size: 57, lr: 2.29e-02, grad_scale: 8.0 +2023-03-08 18:30:55,400 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.80 vs. limit=2.0 +2023-03-08 18:31:20,044 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.949e+02 4.422e+02 5.424e+02 6.713e+02 1.260e+03, threshold=1.085e+03, percent-clipped=2.0 +2023-03-08 18:31:24,814 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0507, 4.5298, 4.6711, 3.7510, 3.4127, 3.5903, 2.1827, 2.0442], + device='cuda:0'), covar=tensor([0.0123, 0.0168, 0.0032, 0.0186, 0.0376, 0.0140, 0.0835, 0.0937], + device='cuda:0'), in_proj_covar=tensor([0.0036, 0.0035, 0.0028, 0.0042, 0.0058, 0.0034, 0.0059, 0.0065], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0004, 0.0002, 0.0003, 0.0004], + device='cuda:0') +2023-03-08 18:31:34,505 INFO [train.py:898] (0/4) Epoch 5, batch 400, loss[loss=0.2551, simple_loss=0.3304, pruned_loss=0.08986, over 17855.00 frames. ], tot_loss[loss=0.2401, simple_loss=0.3135, pruned_loss=0.08333, over 3110961.23 frames. ], batch size: 70, lr: 2.29e-02, grad_scale: 8.0 +2023-03-08 18:32:17,032 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9932, 5.1943, 2.6008, 4.9217, 4.8865, 5.2094, 5.0545, 2.5212], + device='cuda:0'), covar=tensor([0.0144, 0.0043, 0.0703, 0.0060, 0.0051, 0.0039, 0.0075, 0.0990], + device='cuda:0'), in_proj_covar=tensor([0.0063, 0.0049, 0.0078, 0.0059, 0.0056, 0.0047, 0.0062, 0.0086], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0002, 0.0003, 0.0004], + device='cuda:0') +2023-03-08 18:32:31,718 INFO [train.py:898] (0/4) Epoch 5, batch 450, loss[loss=0.2369, simple_loss=0.2913, pruned_loss=0.09127, over 18454.00 frames. ], tot_loss[loss=0.2407, simple_loss=0.3139, pruned_loss=0.08381, over 3225211.51 frames. ], batch size: 43, lr: 2.28e-02, grad_scale: 8.0 +2023-03-08 18:33:16,877 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.98 vs. limit=2.0 +2023-03-08 18:33:17,342 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.939e+02 4.662e+02 5.762e+02 7.845e+02 1.537e+03, threshold=1.152e+03, percent-clipped=9.0 +2023-03-08 18:33:30,766 INFO [train.py:898] (0/4) Epoch 5, batch 500, loss[loss=0.2364, simple_loss=0.3114, pruned_loss=0.08068, over 18483.00 frames. ], tot_loss[loss=0.2401, simple_loss=0.3135, pruned_loss=0.08342, over 3308333.04 frames. ], batch size: 51, lr: 2.28e-02, grad_scale: 8.0 +2023-03-08 18:34:05,249 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3715, 5.9088, 5.3102, 5.6805, 5.3700, 5.5653, 5.9663, 5.8747], + device='cuda:0'), covar=tensor([0.1144, 0.0688, 0.0468, 0.0678, 0.1588, 0.0574, 0.0503, 0.0610], + device='cuda:0'), in_proj_covar=tensor([0.0377, 0.0308, 0.0235, 0.0336, 0.0480, 0.0341, 0.0375, 0.0305], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0005, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-08 18:34:09,981 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15069.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:34:30,389 INFO [train.py:898] (0/4) Epoch 5, batch 550, loss[loss=0.238, simple_loss=0.319, pruned_loss=0.0785, over 17149.00 frames. ], tot_loss[loss=0.2397, simple_loss=0.3132, pruned_loss=0.08304, over 3377854.19 frames. ], batch size: 78, lr: 2.28e-02, grad_scale: 8.0 +2023-03-08 18:34:41,264 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=15096.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:35:06,919 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=15117.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:35:15,839 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.759e+02 4.602e+02 5.527e+02 7.375e+02 1.442e+03, threshold=1.105e+03, percent-clipped=1.0 +2023-03-08 18:35:29,476 INFO [train.py:898] (0/4) Epoch 5, batch 600, loss[loss=0.2487, simple_loss=0.3205, pruned_loss=0.08846, over 18636.00 frames. ], tot_loss[loss=0.2379, simple_loss=0.312, pruned_loss=0.08193, over 3434855.48 frames. ], batch size: 52, lr: 2.27e-02, grad_scale: 8.0 +2023-03-08 18:35:36,515 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15143.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:36:18,685 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-03-08 18:36:28,288 INFO [train.py:898] (0/4) Epoch 5, batch 650, loss[loss=0.2219, simple_loss=0.2889, pruned_loss=0.07749, over 18167.00 frames. ], tot_loss[loss=0.238, simple_loss=0.3122, pruned_loss=0.08186, over 3472475.66 frames. ], batch size: 44, lr: 2.27e-02, grad_scale: 8.0 +2023-03-08 18:36:32,826 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=15191.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:36:33,065 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5787, 3.6396, 1.7788, 4.5173, 3.0518, 4.8710, 2.5298, 3.9827], + device='cuda:0'), covar=tensor([0.0614, 0.0800, 0.1655, 0.0332, 0.0962, 0.0093, 0.1097, 0.0348], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0189, 0.0165, 0.0158, 0.0163, 0.0111, 0.0166, 0.0156], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 18:36:37,705 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=15195.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:37:13,512 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.710e+02 4.789e+02 6.054e+02 7.320e+02 2.200e+03, threshold=1.211e+03, percent-clipped=6.0 +2023-03-08 18:37:14,072 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2363, 1.8892, 3.6944, 2.9714, 3.6289, 5.1656, 4.3612, 4.5959], + device='cuda:0'), covar=tensor([0.0267, 0.0801, 0.0476, 0.0494, 0.0798, 0.0023, 0.0194, 0.0104], + device='cuda:0'), in_proj_covar=tensor([0.0139, 0.0181, 0.0137, 0.0175, 0.0265, 0.0097, 0.0153, 0.0130], + device='cuda:0'), out_proj_covar=tensor([1.0760e-04, 1.3937e-04, 1.1564e-04, 1.2378e-04, 2.0257e-04, 6.8563e-05, + 1.1806e-04, 9.9501e-05], device='cuda:0') +2023-03-08 18:37:27,062 INFO [train.py:898] (0/4) Epoch 5, batch 700, loss[loss=0.2407, simple_loss=0.3122, pruned_loss=0.08465, over 18498.00 frames. ], tot_loss[loss=0.2378, simple_loss=0.3121, pruned_loss=0.08179, over 3500655.63 frames. ], batch size: 51, lr: 2.27e-02, grad_scale: 8.0 +2023-03-08 18:37:49,575 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=15256.0, num_to_drop=1, layers_to_drop={3} +2023-03-08 18:38:26,033 INFO [train.py:898] (0/4) Epoch 5, batch 750, loss[loss=0.2534, simple_loss=0.3294, pruned_loss=0.08869, over 17810.00 frames. ], tot_loss[loss=0.2388, simple_loss=0.313, pruned_loss=0.08226, over 3514950.85 frames. ], batch size: 70, lr: 2.26e-02, grad_scale: 8.0 +2023-03-08 18:39:10,698 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.929e+02 4.559e+02 5.323e+02 7.039e+02 1.434e+03, threshold=1.065e+03, percent-clipped=2.0 +2023-03-08 18:39:24,934 INFO [train.py:898] (0/4) Epoch 5, batch 800, loss[loss=0.2377, simple_loss=0.3179, pruned_loss=0.07874, over 18463.00 frames. ], tot_loss[loss=0.2379, simple_loss=0.3122, pruned_loss=0.08184, over 3527869.08 frames. ], batch size: 59, lr: 2.26e-02, grad_scale: 8.0 +2023-03-08 18:40:24,764 INFO [train.py:898] (0/4) Epoch 5, batch 850, loss[loss=0.1948, simple_loss=0.2701, pruned_loss=0.05979, over 18148.00 frames. ], tot_loss[loss=0.2368, simple_loss=0.3111, pruned_loss=0.08122, over 3547188.87 frames. ], batch size: 44, lr: 2.26e-02, grad_scale: 8.0 +2023-03-08 18:40:35,103 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15396.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:40:44,613 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6060, 3.5135, 3.2656, 3.0949, 3.5534, 3.0362, 2.8174, 3.6878], + device='cuda:0'), covar=tensor([0.0038, 0.0066, 0.0074, 0.0110, 0.0049, 0.0133, 0.0149, 0.0065], + device='cuda:0'), in_proj_covar=tensor([0.0056, 0.0068, 0.0065, 0.0098, 0.0063, 0.0104, 0.0111, 0.0060], + device='cuda:0'), out_proj_covar=tensor([8.4663e-05, 1.0958e-04, 1.0669e-04, 1.6074e-04, 9.8239e-05, 1.6770e-04, + 1.7899e-04, 9.6004e-05], device='cuda:0') +2023-03-08 18:41:09,838 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.891e+02 4.807e+02 5.667e+02 6.596e+02 1.929e+03, threshold=1.133e+03, percent-clipped=7.0 +2023-03-08 18:41:24,203 INFO [train.py:898] (0/4) Epoch 5, batch 900, loss[loss=0.2578, simple_loss=0.3347, pruned_loss=0.09046, over 18297.00 frames. ], tot_loss[loss=0.2378, simple_loss=0.3122, pruned_loss=0.08172, over 3555696.51 frames. ], batch size: 57, lr: 2.25e-02, grad_scale: 8.0 +2023-03-08 18:41:32,501 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=15444.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:42:23,787 INFO [train.py:898] (0/4) Epoch 5, batch 950, loss[loss=0.1926, simple_loss=0.2726, pruned_loss=0.05627, over 18374.00 frames. ], tot_loss[loss=0.238, simple_loss=0.3122, pruned_loss=0.08193, over 3559340.28 frames. ], batch size: 46, lr: 2.25e-02, grad_scale: 8.0 +2023-03-08 18:42:25,260 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8440, 3.8437, 3.8854, 3.8105, 3.8558, 3.8173, 4.0502, 3.9994], + device='cuda:0'), covar=tensor([0.0076, 0.0086, 0.0080, 0.0082, 0.0063, 0.0085, 0.0069, 0.0081], + device='cuda:0'), in_proj_covar=tensor([0.0063, 0.0051, 0.0048, 0.0062, 0.0053, 0.0070, 0.0059, 0.0058], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 18:42:30,366 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-03-08 18:42:42,495 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8987, 2.4632, 3.9926, 3.9284, 2.1644, 4.3391, 3.7371, 2.1696], + device='cuda:0'), covar=tensor([0.0237, 0.1146, 0.0118, 0.0160, 0.1459, 0.0102, 0.0367, 0.1131], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0171, 0.0094, 0.0101, 0.0180, 0.0119, 0.0138, 0.0169], + device='cuda:0'), out_proj_covar=tensor([1.2990e-04, 1.6687e-04, 9.8895e-05, 9.7191e-05, 1.7222e-04, 1.1209e-04, + 1.4223e-04, 1.6946e-04], device='cuda:0') +2023-03-08 18:42:46,120 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2863, 3.1405, 2.9464, 2.6439, 2.9857, 2.4252, 2.2146, 3.1177], + device='cuda:0'), covar=tensor([0.0054, 0.0102, 0.0150, 0.0146, 0.0129, 0.0234, 0.0275, 0.0129], + device='cuda:0'), in_proj_covar=tensor([0.0054, 0.0069, 0.0065, 0.0097, 0.0064, 0.0103, 0.0112, 0.0060], + device='cuda:0'), out_proj_covar=tensor([8.1435e-05, 1.1009e-04, 1.0632e-04, 1.5861e-04, 1.0025e-04, 1.6581e-04, + 1.8024e-04, 9.6192e-05], device='cuda:0') +2023-03-08 18:42:58,131 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-08 18:43:09,483 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.487e+02 4.499e+02 5.437e+02 6.753e+02 3.373e+03, threshold=1.087e+03, percent-clipped=5.0 +2023-03-08 18:43:17,886 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0782, 2.4931, 2.2257, 2.7263, 3.0229, 3.0406, 2.6823, 2.7671], + device='cuda:0'), covar=tensor([0.0287, 0.0264, 0.0805, 0.0254, 0.0248, 0.0134, 0.0381, 0.0246], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0072, 0.0130, 0.0095, 0.0074, 0.0053, 0.0091, 0.0093], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 18:43:23,265 INFO [train.py:898] (0/4) Epoch 5, batch 1000, loss[loss=0.2538, simple_loss=0.3279, pruned_loss=0.08983, over 16150.00 frames. ], tot_loss[loss=0.24, simple_loss=0.3142, pruned_loss=0.08283, over 3559512.51 frames. ], batch size: 94, lr: 2.25e-02, grad_scale: 8.0 +2023-03-08 18:43:39,867 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=15551.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 18:43:41,154 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8539, 4.6586, 4.6248, 3.4635, 3.8210, 3.5376, 2.5197, 2.1326], + device='cuda:0'), covar=tensor([0.0181, 0.0103, 0.0052, 0.0246, 0.0232, 0.0205, 0.0686, 0.0956], + device='cuda:0'), in_proj_covar=tensor([0.0039, 0.0037, 0.0030, 0.0045, 0.0061, 0.0035, 0.0063, 0.0069], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0004, 0.0002, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 18:44:05,591 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0160, 3.8017, 5.1981, 3.4431, 4.4675, 2.8540, 2.8739, 2.1855], + device='cuda:0'), covar=tensor([0.0560, 0.0484, 0.0039, 0.0328, 0.0340, 0.1511, 0.1417, 0.1261], + device='cuda:0'), in_proj_covar=tensor([0.0155, 0.0160, 0.0079, 0.0126, 0.0176, 0.0206, 0.0177, 0.0170], + device='cuda:0'), out_proj_covar=tensor([1.5141e-04, 1.6150e-04, 8.3077e-05, 1.2598e-04, 1.7761e-04, 2.0375e-04, + 1.8598e-04, 1.7089e-04], device='cuda:0') +2023-03-08 18:44:23,076 INFO [train.py:898] (0/4) Epoch 5, batch 1050, loss[loss=0.3248, simple_loss=0.3711, pruned_loss=0.1392, over 12519.00 frames. ], tot_loss[loss=0.2394, simple_loss=0.3135, pruned_loss=0.08263, over 3562896.84 frames. ], batch size: 131, lr: 2.24e-02, grad_scale: 8.0 +2023-03-08 18:45:06,974 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1802, 4.7928, 4.7840, 4.5886, 4.4183, 4.5902, 4.0111, 4.5864], + device='cuda:0'), covar=tensor([0.0254, 0.0248, 0.0222, 0.0255, 0.0430, 0.0240, 0.1281, 0.0258], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0165, 0.0150, 0.0141, 0.0159, 0.0167, 0.0233, 0.0146], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006, 0.0004], + device='cuda:0') +2023-03-08 18:45:08,999 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.376e+02 4.268e+02 5.180e+02 6.022e+02 1.534e+03, threshold=1.036e+03, percent-clipped=2.0 +2023-03-08 18:45:23,120 INFO [train.py:898] (0/4) Epoch 5, batch 1100, loss[loss=0.2875, simple_loss=0.3488, pruned_loss=0.113, over 18570.00 frames. ], tot_loss[loss=0.2387, simple_loss=0.3128, pruned_loss=0.08232, over 3569537.71 frames. ], batch size: 54, lr: 2.24e-02, grad_scale: 8.0 +2023-03-08 18:45:54,003 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7373, 1.7092, 3.5265, 2.8784, 3.3535, 4.9279, 4.2845, 4.4494], + device='cuda:0'), covar=tensor([0.0347, 0.0945, 0.0529, 0.0543, 0.0914, 0.0028, 0.0204, 0.0118], + device='cuda:0'), in_proj_covar=tensor([0.0142, 0.0187, 0.0142, 0.0179, 0.0268, 0.0101, 0.0154, 0.0134], + device='cuda:0'), out_proj_covar=tensor([1.0868e-04, 1.4285e-04, 1.1877e-04, 1.2641e-04, 2.0357e-04, 7.0817e-05, + 1.1859e-04, 1.0250e-04], device='cuda:0') +2023-03-08 18:46:22,716 INFO [train.py:898] (0/4) Epoch 5, batch 1150, loss[loss=0.275, simple_loss=0.3532, pruned_loss=0.09839, over 18084.00 frames. ], tot_loss[loss=0.2385, simple_loss=0.3124, pruned_loss=0.08228, over 3574319.08 frames. ], batch size: 62, lr: 2.24e-02, grad_scale: 8.0 +2023-03-08 18:46:35,168 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-08 18:46:56,712 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0691, 4.0354, 5.2409, 3.1100, 4.4795, 2.9382, 3.0319, 2.3033], + device='cuda:0'), covar=tensor([0.0556, 0.0444, 0.0039, 0.0421, 0.0395, 0.1466, 0.1555, 0.1221], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0162, 0.0079, 0.0127, 0.0176, 0.0205, 0.0176, 0.0170], + device='cuda:0'), out_proj_covar=tensor([1.5083e-04, 1.6285e-04, 8.2650e-05, 1.2656e-04, 1.7692e-04, 2.0259e-04, + 1.8451e-04, 1.7068e-04], device='cuda:0') +2023-03-08 18:47:06,233 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.304e+02 5.042e+02 5.995e+02 7.754e+02 1.495e+03, threshold=1.199e+03, percent-clipped=7.0 +2023-03-08 18:47:21,467 INFO [train.py:898] (0/4) Epoch 5, batch 1200, loss[loss=0.2511, simple_loss=0.3139, pruned_loss=0.09422, over 18276.00 frames. ], tot_loss[loss=0.238, simple_loss=0.3121, pruned_loss=0.08194, over 3587746.99 frames. ], batch size: 49, lr: 2.23e-02, grad_scale: 8.0 +2023-03-08 18:47:40,239 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0528, 4.9199, 5.1222, 4.9746, 4.9372, 5.5918, 5.3155, 5.0594], + device='cuda:0'), covar=tensor([0.0755, 0.0526, 0.0591, 0.0529, 0.1209, 0.0723, 0.0434, 0.1243], + device='cuda:0'), in_proj_covar=tensor([0.0236, 0.0172, 0.0182, 0.0178, 0.0216, 0.0265, 0.0165, 0.0247], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 18:48:19,536 INFO [train.py:898] (0/4) Epoch 5, batch 1250, loss[loss=0.2472, simple_loss=0.3227, pruned_loss=0.08583, over 16192.00 frames. ], tot_loss[loss=0.2382, simple_loss=0.3124, pruned_loss=0.082, over 3589456.68 frames. ], batch size: 94, lr: 2.23e-02, grad_scale: 8.0 +2023-03-08 18:48:33,378 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9498, 3.9392, 5.1411, 3.5765, 4.2610, 2.9609, 3.3081, 2.5121], + device='cuda:0'), covar=tensor([0.0634, 0.0504, 0.0049, 0.0338, 0.0506, 0.1463, 0.1343, 0.1233], + device='cuda:0'), in_proj_covar=tensor([0.0159, 0.0166, 0.0082, 0.0129, 0.0179, 0.0209, 0.0182, 0.0174], + device='cuda:0'), out_proj_covar=tensor([1.5556e-04, 1.6707e-04, 8.5369e-05, 1.2914e-04, 1.7984e-04, 2.0702e-04, + 1.9021e-04, 1.7394e-04], device='cuda:0') +2023-03-08 18:49:03,912 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.909e+02 4.343e+02 5.272e+02 6.723e+02 1.264e+03, threshold=1.054e+03, percent-clipped=1.0 +2023-03-08 18:49:18,486 INFO [train.py:898] (0/4) Epoch 5, batch 1300, loss[loss=0.2187, simple_loss=0.29, pruned_loss=0.07364, over 18266.00 frames. ], tot_loss[loss=0.2375, simple_loss=0.312, pruned_loss=0.08156, over 3598616.50 frames. ], batch size: 47, lr: 2.23e-02, grad_scale: 4.0 +2023-03-08 18:49:35,177 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=15851.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:50:12,053 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=15883.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:50:16,767 INFO [train.py:898] (0/4) Epoch 5, batch 1350, loss[loss=0.2815, simple_loss=0.353, pruned_loss=0.105, over 18118.00 frames. ], tot_loss[loss=0.237, simple_loss=0.3114, pruned_loss=0.08132, over 3604075.01 frames. ], batch size: 62, lr: 2.22e-02, grad_scale: 4.0 +2023-03-08 18:50:29,464 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5233, 4.4666, 4.5573, 4.3217, 4.2721, 4.4439, 4.8008, 4.7561], + device='cuda:0'), covar=tensor([0.0064, 0.0076, 0.0098, 0.0094, 0.0095, 0.0099, 0.0085, 0.0099], + device='cuda:0'), in_proj_covar=tensor([0.0062, 0.0049, 0.0047, 0.0060, 0.0052, 0.0069, 0.0057, 0.0057], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 18:50:31,445 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=15899.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:50:58,023 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1105, 4.5611, 4.5753, 4.3957, 4.2556, 4.3841, 3.8616, 4.3791], + device='cuda:0'), covar=tensor([0.0241, 0.0289, 0.0274, 0.0283, 0.0342, 0.0269, 0.1192, 0.0313], + device='cuda:0'), in_proj_covar=tensor([0.0125, 0.0165, 0.0149, 0.0142, 0.0160, 0.0167, 0.0233, 0.0146], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006, 0.0004], + device='cuda:0') +2023-03-08 18:51:02,178 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.900e+02 4.519e+02 5.491e+02 6.869e+02 1.432e+03, threshold=1.098e+03, percent-clipped=4.0 +2023-03-08 18:51:15,441 INFO [train.py:898] (0/4) Epoch 5, batch 1400, loss[loss=0.2745, simple_loss=0.3359, pruned_loss=0.1066, over 12498.00 frames. ], tot_loss[loss=0.2375, simple_loss=0.3117, pruned_loss=0.08166, over 3575685.23 frames. ], batch size: 131, lr: 2.22e-02, grad_scale: 4.0 +2023-03-08 18:51:24,342 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.4551, 4.4278, 4.4684, 4.1687, 4.3461, 4.2553, 4.6583, 4.5931], + device='cuda:0'), covar=tensor([0.0073, 0.0090, 0.0102, 0.0109, 0.0082, 0.0124, 0.0088, 0.0124], + device='cuda:0'), in_proj_covar=tensor([0.0065, 0.0051, 0.0049, 0.0063, 0.0055, 0.0072, 0.0059, 0.0059], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 18:51:24,389 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=15944.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:52:13,986 INFO [train.py:898] (0/4) Epoch 5, batch 1450, loss[loss=0.2315, simple_loss=0.3133, pruned_loss=0.07481, over 18377.00 frames. ], tot_loss[loss=0.2367, simple_loss=0.3113, pruned_loss=0.08099, over 3578592.65 frames. ], batch size: 52, lr: 2.22e-02, grad_scale: 4.0 +2023-03-08 18:52:30,594 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-16000.pt +2023-03-08 18:52:36,444 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16001.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:53:04,646 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.050e+02 4.332e+02 5.250e+02 6.313e+02 1.611e+03, threshold=1.050e+03, percent-clipped=2.0 +2023-03-08 18:53:16,910 INFO [train.py:898] (0/4) Epoch 5, batch 1500, loss[loss=0.3106, simple_loss=0.3563, pruned_loss=0.1324, over 11899.00 frames. ], tot_loss[loss=0.2365, simple_loss=0.3109, pruned_loss=0.08103, over 3575428.15 frames. ], batch size: 131, lr: 2.21e-02, grad_scale: 4.0 +2023-03-08 18:53:47,526 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16062.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:54:14,619 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16085.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:54:16,407 INFO [train.py:898] (0/4) Epoch 5, batch 1550, loss[loss=0.2518, simple_loss=0.3339, pruned_loss=0.08487, over 18348.00 frames. ], tot_loss[loss=0.2357, simple_loss=0.3102, pruned_loss=0.08059, over 3578374.76 frames. ], batch size: 56, lr: 2.21e-02, grad_scale: 4.0 +2023-03-08 18:55:02,223 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.998e+02 4.560e+02 5.884e+02 7.024e+02 2.005e+03, threshold=1.177e+03, percent-clipped=3.0 +2023-03-08 18:55:07,591 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.04 vs. limit=2.0 +2023-03-08 18:55:14,721 INFO [train.py:898] (0/4) Epoch 5, batch 1600, loss[loss=0.2492, simple_loss=0.3202, pruned_loss=0.08912, over 18406.00 frames. ], tot_loss[loss=0.2354, simple_loss=0.3099, pruned_loss=0.08044, over 3582724.59 frames. ], batch size: 52, lr: 2.21e-02, grad_scale: 8.0 +2023-03-08 18:55:26,039 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16146.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:55:29,578 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16149.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:56:14,000 INFO [train.py:898] (0/4) Epoch 5, batch 1650, loss[loss=0.2446, simple_loss=0.3221, pruned_loss=0.08356, over 17784.00 frames. ], tot_loss[loss=0.2357, simple_loss=0.3102, pruned_loss=0.08056, over 3576445.11 frames. ], batch size: 70, lr: 2.20e-02, grad_scale: 8.0 +2023-03-08 18:56:42,229 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16210.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:56:47,902 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9652, 4.8285, 4.8588, 4.8154, 4.6650, 4.7952, 5.2315, 5.1499], + device='cuda:0'), covar=tensor([0.0055, 0.0090, 0.0076, 0.0078, 0.0077, 0.0093, 0.0065, 0.0074], + device='cuda:0'), in_proj_covar=tensor([0.0062, 0.0049, 0.0046, 0.0060, 0.0052, 0.0069, 0.0057, 0.0056], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 18:57:00,819 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.794e+02 4.512e+02 5.424e+02 6.724e+02 1.524e+03, threshold=1.085e+03, percent-clipped=4.0 +2023-03-08 18:57:13,386 INFO [train.py:898] (0/4) Epoch 5, batch 1700, loss[loss=0.2188, simple_loss=0.2926, pruned_loss=0.07254, over 18396.00 frames. ], tot_loss[loss=0.2342, simple_loss=0.3088, pruned_loss=0.07979, over 3583754.98 frames. ], batch size: 50, lr: 2.20e-02, grad_scale: 8.0 +2023-03-08 18:57:15,875 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16239.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:57:21,855 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8842, 2.8432, 4.1526, 4.1604, 2.5981, 4.5313, 3.9095, 2.7386], + device='cuda:0'), covar=tensor([0.0258, 0.1094, 0.0187, 0.0178, 0.1406, 0.0099, 0.0361, 0.0981], + device='cuda:0'), in_proj_covar=tensor([0.0140, 0.0177, 0.0098, 0.0104, 0.0183, 0.0128, 0.0142, 0.0174], + device='cuda:0'), out_proj_covar=tensor([1.3592e-04, 1.7208e-04, 1.0404e-04, 9.9364e-05, 1.7521e-04, 1.2117e-04, + 1.4548e-04, 1.7343e-04], device='cuda:0') +2023-03-08 18:58:13,013 INFO [train.py:898] (0/4) Epoch 5, batch 1750, loss[loss=0.2094, simple_loss=0.285, pruned_loss=0.06694, over 18417.00 frames. ], tot_loss[loss=0.2336, simple_loss=0.3083, pruned_loss=0.07946, over 3587237.72 frames. ], batch size: 48, lr: 2.20e-02, grad_scale: 8.0 +2023-03-08 18:58:58,896 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.349e+02 4.407e+02 5.169e+02 6.479e+02 1.420e+03, threshold=1.034e+03, percent-clipped=4.0 +2023-03-08 18:59:11,657 INFO [train.py:898] (0/4) Epoch 5, batch 1800, loss[loss=0.2485, simple_loss=0.3293, pruned_loss=0.08383, over 17993.00 frames. ], tot_loss[loss=0.2344, simple_loss=0.3087, pruned_loss=0.08002, over 3580100.02 frames. ], batch size: 65, lr: 2.19e-02, grad_scale: 8.0 +2023-03-08 18:59:35,065 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16357.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:59:51,717 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-08 18:59:52,225 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16371.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 18:59:52,810 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-03-08 18:59:58,203 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.52 vs. limit=5.0 +2023-03-08 19:00:10,469 INFO [train.py:898] (0/4) Epoch 5, batch 1850, loss[loss=0.2221, simple_loss=0.3029, pruned_loss=0.07061, over 18382.00 frames. ], tot_loss[loss=0.2349, simple_loss=0.309, pruned_loss=0.08034, over 3580418.54 frames. ], batch size: 46, lr: 2.19e-02, grad_scale: 8.0 +2023-03-08 19:00:55,913 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.146e+02 4.944e+02 5.980e+02 8.174e+02 1.619e+03, threshold=1.196e+03, percent-clipped=7.0 +2023-03-08 19:01:03,053 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16432.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:01:09,064 INFO [train.py:898] (0/4) Epoch 5, batch 1900, loss[loss=0.2304, simple_loss=0.3202, pruned_loss=0.07028, over 18316.00 frames. ], tot_loss[loss=0.2347, simple_loss=0.3091, pruned_loss=0.08012, over 3583045.15 frames. ], batch size: 54, lr: 2.19e-02, grad_scale: 8.0 +2023-03-08 19:01:13,689 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16441.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:01:31,568 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-08 19:01:38,058 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0682, 4.8884, 4.9964, 4.6991, 4.8301, 4.9917, 5.3952, 5.2328], + device='cuda:0'), covar=tensor([0.0055, 0.0077, 0.0061, 0.0094, 0.0072, 0.0086, 0.0089, 0.0087], + device='cuda:0'), in_proj_covar=tensor([0.0064, 0.0050, 0.0047, 0.0062, 0.0053, 0.0070, 0.0059, 0.0057], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 19:02:07,325 INFO [train.py:898] (0/4) Epoch 5, batch 1950, loss[loss=0.2052, simple_loss=0.2815, pruned_loss=0.06446, over 18506.00 frames. ], tot_loss[loss=0.2345, simple_loss=0.3089, pruned_loss=0.08004, over 3578320.67 frames. ], batch size: 47, lr: 2.19e-02, grad_scale: 8.0 +2023-03-08 19:02:28,045 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16505.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:02:31,660 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8970, 4.6043, 4.7635, 3.9048, 3.6496, 3.5782, 2.6101, 2.1726], + device='cuda:0'), covar=tensor([0.0148, 0.0113, 0.0038, 0.0143, 0.0301, 0.0163, 0.0647, 0.0863], + device='cuda:0'), in_proj_covar=tensor([0.0039, 0.0037, 0.0032, 0.0045, 0.0062, 0.0036, 0.0062, 0.0067], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0004, 0.0002, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 19:02:53,788 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.181e+02 4.721e+02 5.735e+02 6.721e+02 1.938e+03, threshold=1.147e+03, percent-clipped=5.0 +2023-03-08 19:03:06,025 INFO [train.py:898] (0/4) Epoch 5, batch 2000, loss[loss=0.225, simple_loss=0.2936, pruned_loss=0.0782, over 18494.00 frames. ], tot_loss[loss=0.2343, simple_loss=0.3086, pruned_loss=0.08001, over 3578064.95 frames. ], batch size: 47, lr: 2.18e-02, grad_scale: 8.0 +2023-03-08 19:03:09,078 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16539.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:03:41,432 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6098, 4.2743, 4.4858, 3.6672, 3.4943, 3.5616, 2.2346, 2.0175], + device='cuda:0'), covar=tensor([0.0178, 0.0165, 0.0062, 0.0195, 0.0321, 0.0146, 0.0769, 0.0957], + device='cuda:0'), in_proj_covar=tensor([0.0039, 0.0038, 0.0031, 0.0045, 0.0061, 0.0037, 0.0062, 0.0067], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0004, 0.0002, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 19:03:48,942 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6492, 2.1496, 2.7129, 2.4636, 3.0224, 3.8920, 3.5383, 3.0313], + device='cuda:0'), covar=tensor([0.0419, 0.0927, 0.0916, 0.0673, 0.1024, 0.0069, 0.0291, 0.0275], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0197, 0.0160, 0.0188, 0.0282, 0.0107, 0.0164, 0.0140], + device='cuda:0'), out_proj_covar=tensor([1.1324e-04, 1.4864e-04, 1.3013e-04, 1.3187e-04, 2.1021e-04, 7.4024e-05, + 1.2402e-04, 1.0473e-04], device='cuda:0') +2023-03-08 19:04:04,800 INFO [train.py:898] (0/4) Epoch 5, batch 2050, loss[loss=0.29, simple_loss=0.3573, pruned_loss=0.1114, over 16118.00 frames. ], tot_loss[loss=0.2354, simple_loss=0.3098, pruned_loss=0.0805, over 3567583.15 frames. ], batch size: 94, lr: 2.18e-02, grad_scale: 8.0 +2023-03-08 19:04:05,000 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=16587.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:04:51,608 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.823e+02 4.363e+02 5.469e+02 6.595e+02 1.217e+03, threshold=1.094e+03, percent-clipped=2.0 +2023-03-08 19:05:04,766 INFO [train.py:898] (0/4) Epoch 5, batch 2100, loss[loss=0.2214, simple_loss=0.2904, pruned_loss=0.07619, over 18433.00 frames. ], tot_loss[loss=0.2354, simple_loss=0.3097, pruned_loss=0.08058, over 3564816.41 frames. ], batch size: 43, lr: 2.18e-02, grad_scale: 8.0 +2023-03-08 19:05:11,049 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-08 19:05:28,360 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16657.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:06:04,032 INFO [train.py:898] (0/4) Epoch 5, batch 2150, loss[loss=0.2594, simple_loss=0.3303, pruned_loss=0.09421, over 18109.00 frames. ], tot_loss[loss=0.2352, simple_loss=0.3095, pruned_loss=0.08048, over 3567012.93 frames. ], batch size: 62, lr: 2.17e-02, grad_scale: 8.0 +2023-03-08 19:06:24,687 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=16705.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:06:49,438 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.208e+02 4.846e+02 5.504e+02 7.182e+02 1.365e+03, threshold=1.101e+03, percent-clipped=1.0 +2023-03-08 19:06:50,763 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=16727.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:06:57,657 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.07 vs. limit=2.0 +2023-03-08 19:07:02,270 INFO [train.py:898] (0/4) Epoch 5, batch 2200, loss[loss=0.2452, simple_loss=0.3187, pruned_loss=0.08579, over 18385.00 frames. ], tot_loss[loss=0.2344, simple_loss=0.309, pruned_loss=0.07988, over 3587985.79 frames. ], batch size: 50, lr: 2.17e-02, grad_scale: 8.0 +2023-03-08 19:07:07,009 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16741.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:07:44,514 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3435, 4.4515, 2.1033, 4.6669, 5.3291, 2.3719, 3.8715, 3.8231], + device='cuda:0'), covar=tensor([0.0066, 0.1026, 0.1778, 0.0415, 0.0042, 0.1517, 0.0592, 0.0846], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0162, 0.0176, 0.0171, 0.0072, 0.0164, 0.0182, 0.0178], + device='cuda:0'), out_proj_covar=tensor([1.1104e-04, 2.2983e-04, 2.3087e-04, 2.3080e-04, 9.9403e-05, 2.2364e-04, + 2.3929e-04, 2.4054e-04], device='cuda:0') +2023-03-08 19:08:01,107 INFO [train.py:898] (0/4) Epoch 5, batch 2250, loss[loss=0.2679, simple_loss=0.3378, pruned_loss=0.09899, over 18417.00 frames. ], tot_loss[loss=0.2342, simple_loss=0.309, pruned_loss=0.0797, over 3576774.44 frames. ], batch size: 52, lr: 2.17e-02, grad_scale: 8.0 +2023-03-08 19:08:03,391 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=16789.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:08:08,041 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.4084, 5.6422, 3.4767, 5.3297, 5.2431, 5.6542, 5.5195, 3.0147], + device='cuda:0'), covar=tensor([0.0124, 0.0035, 0.0456, 0.0045, 0.0049, 0.0038, 0.0058, 0.0770], + device='cuda:0'), in_proj_covar=tensor([0.0062, 0.0049, 0.0079, 0.0062, 0.0060, 0.0049, 0.0063, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0003, 0.0004], + device='cuda:0') +2023-03-08 19:08:22,225 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=16805.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:08:26,840 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6972, 4.6235, 4.7394, 4.5162, 4.5401, 4.5197, 4.9691, 5.0023], + device='cuda:0'), covar=tensor([0.0068, 0.0075, 0.0081, 0.0089, 0.0075, 0.0128, 0.0067, 0.0084], + device='cuda:0'), in_proj_covar=tensor([0.0064, 0.0051, 0.0048, 0.0063, 0.0054, 0.0070, 0.0059, 0.0059], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 19:08:46,493 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.297e+02 4.310e+02 5.501e+02 7.389e+02 1.420e+03, threshold=1.100e+03, percent-clipped=2.0 +2023-03-08 19:09:00,035 INFO [train.py:898] (0/4) Epoch 5, batch 2300, loss[loss=0.2781, simple_loss=0.3453, pruned_loss=0.1054, over 18061.00 frames. ], tot_loss[loss=0.2349, simple_loss=0.3097, pruned_loss=0.08007, over 3586152.70 frames. ], batch size: 62, lr: 2.16e-02, grad_scale: 8.0 +2023-03-08 19:09:08,125 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3142, 5.9354, 5.4143, 5.6272, 5.3703, 5.4150, 5.9447, 5.9574], + device='cuda:0'), covar=tensor([0.1041, 0.0562, 0.0394, 0.0588, 0.1398, 0.0596, 0.0496, 0.0534], + device='cuda:0'), in_proj_covar=tensor([0.0404, 0.0330, 0.0252, 0.0353, 0.0492, 0.0363, 0.0415, 0.0334], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 19:09:10,463 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.4245, 5.1474, 5.0738, 4.9185, 4.6113, 4.8723, 4.2624, 4.8883], + device='cuda:0'), covar=tensor([0.0263, 0.0258, 0.0195, 0.0230, 0.0350, 0.0253, 0.1125, 0.0267], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0168, 0.0151, 0.0145, 0.0163, 0.0170, 0.0238, 0.0151], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0004, 0.0004, 0.0005, 0.0005, 0.0006, 0.0005], + device='cuda:0') +2023-03-08 19:09:17,907 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=16853.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:09:35,519 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16868.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 19:09:57,302 INFO [train.py:898] (0/4) Epoch 5, batch 2350, loss[loss=0.2451, simple_loss=0.3238, pruned_loss=0.08321, over 18243.00 frames. ], tot_loss[loss=0.2364, simple_loss=0.3112, pruned_loss=0.0808, over 3583437.56 frames. ], batch size: 60, lr: 2.16e-02, grad_scale: 8.0 +2023-03-08 19:10:19,717 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=16906.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:10:42,829 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.402e+02 4.892e+02 6.073e+02 7.229e+02 1.276e+03, threshold=1.215e+03, percent-clipped=6.0 +2023-03-08 19:10:47,104 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16929.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 19:10:56,187 INFO [train.py:898] (0/4) Epoch 5, batch 2400, loss[loss=0.2249, simple_loss=0.3104, pruned_loss=0.06968, over 18344.00 frames. ], tot_loss[loss=0.2359, simple_loss=0.3106, pruned_loss=0.08061, over 3581857.16 frames. ], batch size: 56, lr: 2.16e-02, grad_scale: 8.0 +2023-03-08 19:11:31,359 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=16967.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 19:11:54,510 INFO [train.py:898] (0/4) Epoch 5, batch 2450, loss[loss=0.2187, simple_loss=0.3068, pruned_loss=0.06532, over 18389.00 frames. ], tot_loss[loss=0.2354, simple_loss=0.3103, pruned_loss=0.08025, over 3580744.49 frames. ], batch size: 52, lr: 2.16e-02, grad_scale: 8.0 +2023-03-08 19:12:39,639 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.725e+02 4.440e+02 5.461e+02 7.873e+02 1.788e+03, threshold=1.092e+03, percent-clipped=5.0 +2023-03-08 19:12:41,090 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17027.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:12:49,311 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-08 19:12:53,113 INFO [train.py:898] (0/4) Epoch 5, batch 2500, loss[loss=0.2351, simple_loss=0.3138, pruned_loss=0.07821, over 16231.00 frames. ], tot_loss[loss=0.2352, simple_loss=0.3103, pruned_loss=0.08007, over 3574832.29 frames. ], batch size: 94, lr: 2.15e-02, grad_scale: 8.0 +2023-03-08 19:13:37,525 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=17075.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:13:50,996 INFO [train.py:898] (0/4) Epoch 5, batch 2550, loss[loss=0.1991, simple_loss=0.278, pruned_loss=0.06015, over 18410.00 frames. ], tot_loss[loss=0.2342, simple_loss=0.3094, pruned_loss=0.07952, over 3591754.11 frames. ], batch size: 48, lr: 2.15e-02, grad_scale: 8.0 +2023-03-08 19:14:06,771 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5400, 5.6559, 3.0603, 5.3957, 5.3059, 5.6606, 5.5618, 3.1492], + device='cuda:0'), covar=tensor([0.0111, 0.0035, 0.0556, 0.0041, 0.0046, 0.0036, 0.0053, 0.0738], + device='cuda:0'), in_proj_covar=tensor([0.0062, 0.0049, 0.0080, 0.0063, 0.0060, 0.0049, 0.0063, 0.0086], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0003, 0.0004], + device='cuda:0') +2023-03-08 19:14:36,785 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.128e+02 4.504e+02 5.454e+02 6.580e+02 1.151e+03, threshold=1.091e+03, percent-clipped=2.0 +2023-03-08 19:14:49,069 INFO [train.py:898] (0/4) Epoch 5, batch 2600, loss[loss=0.2514, simple_loss=0.3265, pruned_loss=0.08815, over 18030.00 frames. ], tot_loss[loss=0.2351, simple_loss=0.3101, pruned_loss=0.08008, over 3572073.54 frames. ], batch size: 62, lr: 2.15e-02, grad_scale: 8.0 +2023-03-08 19:14:55,610 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0363, 5.6350, 5.0539, 5.4055, 5.1309, 5.2471, 5.7049, 5.5940], + device='cuda:0'), covar=tensor([0.1009, 0.0503, 0.0577, 0.0615, 0.1317, 0.0532, 0.0390, 0.0539], + device='cuda:0'), in_proj_covar=tensor([0.0398, 0.0327, 0.0255, 0.0355, 0.0496, 0.0355, 0.0401, 0.0333], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0005, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-08 19:15:34,236 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0051, 4.0077, 5.2212, 3.5484, 4.5537, 2.8982, 3.1528, 2.3175], + device='cuda:0'), covar=tensor([0.0634, 0.0500, 0.0042, 0.0361, 0.0395, 0.1601, 0.1652, 0.1423], + device='cuda:0'), in_proj_covar=tensor([0.0163, 0.0172, 0.0084, 0.0137, 0.0187, 0.0224, 0.0197, 0.0181], + device='cuda:0'), out_proj_covar=tensor([1.5794e-04, 1.7198e-04, 8.7346e-05, 1.3413e-04, 1.8645e-04, 2.1806e-04, + 2.0109e-04, 1.8117e-04], device='cuda:0') +2023-03-08 19:15:47,078 INFO [train.py:898] (0/4) Epoch 5, batch 2650, loss[loss=0.2419, simple_loss=0.3166, pruned_loss=0.08363, over 18399.00 frames. ], tot_loss[loss=0.2354, simple_loss=0.3102, pruned_loss=0.08029, over 3577917.29 frames. ], batch size: 52, lr: 2.14e-02, grad_scale: 8.0 +2023-03-08 19:15:49,463 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17189.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:16:25,345 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5364, 3.3436, 3.0251, 2.8558, 3.0114, 2.5716, 2.6437, 3.3394], + device='cuda:0'), covar=tensor([0.0041, 0.0087, 0.0116, 0.0148, 0.0115, 0.0198, 0.0182, 0.0104], + device='cuda:0'), in_proj_covar=tensor([0.0054, 0.0070, 0.0064, 0.0101, 0.0064, 0.0105, 0.0113, 0.0061], + device='cuda:0'), out_proj_covar=tensor([7.9179e-05, 1.1065e-04, 1.0345e-04, 1.6398e-04, 9.9436e-05, 1.6717e-04, + 1.7846e-04, 9.5226e-05], device='cuda:0') +2023-03-08 19:16:31,329 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17224.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 19:16:33,374 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.843e+02 4.713e+02 5.523e+02 7.197e+02 1.239e+03, threshold=1.105e+03, percent-clipped=3.0 +2023-03-08 19:16:39,640 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8139, 3.9299, 5.0036, 3.0116, 4.3230, 2.7765, 2.9289, 2.0686], + device='cuda:0'), covar=tensor([0.0627, 0.0484, 0.0047, 0.0465, 0.0398, 0.1493, 0.1488, 0.1403], + device='cuda:0'), in_proj_covar=tensor([0.0163, 0.0171, 0.0083, 0.0136, 0.0187, 0.0220, 0.0195, 0.0180], + device='cuda:0'), out_proj_covar=tensor([1.5738e-04, 1.7096e-04, 8.6648e-05, 1.3340e-04, 1.8548e-04, 2.1498e-04, + 1.9936e-04, 1.7927e-04], device='cuda:0') +2023-03-08 19:16:45,587 INFO [train.py:898] (0/4) Epoch 5, batch 2700, loss[loss=0.2234, simple_loss=0.2882, pruned_loss=0.07928, over 18397.00 frames. ], tot_loss[loss=0.2364, simple_loss=0.3111, pruned_loss=0.08088, over 3556912.07 frames. ], batch size: 42, lr: 2.14e-02, grad_scale: 8.0 +2023-03-08 19:16:56,462 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-03-08 19:17:01,343 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17250.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:17:15,314 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17262.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 19:17:43,514 INFO [train.py:898] (0/4) Epoch 5, batch 2750, loss[loss=0.2392, simple_loss=0.3127, pruned_loss=0.08288, over 18402.00 frames. ], tot_loss[loss=0.2349, simple_loss=0.3098, pruned_loss=0.07998, over 3574054.91 frames. ], batch size: 52, lr: 2.14e-02, grad_scale: 8.0 +2023-03-08 19:18:29,772 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.562e+02 4.313e+02 5.357e+02 6.708e+02 1.255e+03, threshold=1.071e+03, percent-clipped=3.0 +2023-03-08 19:18:42,600 INFO [train.py:898] (0/4) Epoch 5, batch 2800, loss[loss=0.2399, simple_loss=0.3144, pruned_loss=0.08266, over 16067.00 frames. ], tot_loss[loss=0.2343, simple_loss=0.3093, pruned_loss=0.07968, over 3573114.33 frames. ], batch size: 94, lr: 2.14e-02, grad_scale: 8.0 +2023-03-08 19:19:12,122 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17362.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:19:36,924 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17383.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 19:19:41,194 INFO [train.py:898] (0/4) Epoch 5, batch 2850, loss[loss=0.2579, simple_loss=0.3366, pruned_loss=0.08958, over 18497.00 frames. ], tot_loss[loss=0.2345, simple_loss=0.3096, pruned_loss=0.07968, over 3577362.26 frames. ], batch size: 53, lr: 2.13e-02, grad_scale: 8.0 +2023-03-08 19:19:59,514 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-08 19:20:01,578 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5956, 4.5131, 4.5594, 4.3016, 4.3044, 4.4461, 4.8981, 4.8701], + device='cuda:0'), covar=tensor([0.0066, 0.0092, 0.0086, 0.0101, 0.0093, 0.0112, 0.0076, 0.0092], + device='cuda:0'), in_proj_covar=tensor([0.0062, 0.0050, 0.0048, 0.0062, 0.0053, 0.0070, 0.0058, 0.0059], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 19:20:24,411 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17423.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:20:27,501 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.643e+02 4.399e+02 5.133e+02 6.375e+02 2.228e+03, threshold=1.027e+03, percent-clipped=3.0 +2023-03-08 19:20:40,826 INFO [train.py:898] (0/4) Epoch 5, batch 2900, loss[loss=0.2292, simple_loss=0.3152, pruned_loss=0.07166, over 18494.00 frames. ], tot_loss[loss=0.2347, simple_loss=0.31, pruned_loss=0.07967, over 3586799.47 frames. ], batch size: 51, lr: 2.13e-02, grad_scale: 8.0 +2023-03-08 19:20:49,188 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17444.0, num_to_drop=1, layers_to_drop={3} +2023-03-08 19:20:59,664 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-08 19:21:02,789 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-08 19:21:38,382 INFO [train.py:898] (0/4) Epoch 5, batch 2950, loss[loss=0.2429, simple_loss=0.3256, pruned_loss=0.08012, over 18376.00 frames. ], tot_loss[loss=0.2349, simple_loss=0.3101, pruned_loss=0.07989, over 3570766.03 frames. ], batch size: 56, lr: 2.13e-02, grad_scale: 8.0 +2023-03-08 19:22:22,645 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17524.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 19:22:24,448 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.921e+02 4.474e+02 5.609e+02 7.240e+02 1.382e+03, threshold=1.122e+03, percent-clipped=5.0 +2023-03-08 19:22:29,235 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5034, 2.6761, 2.6997, 2.9748, 3.6456, 3.4253, 3.0663, 3.1152], + device='cuda:0'), covar=tensor([0.0292, 0.0378, 0.0676, 0.0257, 0.0171, 0.0182, 0.0409, 0.0314], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0078, 0.0133, 0.0102, 0.0076, 0.0058, 0.0096, 0.0103], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 19:22:36,711 INFO [train.py:898] (0/4) Epoch 5, batch 3000, loss[loss=0.2242, simple_loss=0.2894, pruned_loss=0.07952, over 18254.00 frames. ], tot_loss[loss=0.2339, simple_loss=0.3093, pruned_loss=0.0793, over 3583415.50 frames. ], batch size: 45, lr: 2.12e-02, grad_scale: 8.0 +2023-03-08 19:22:36,714 INFO [train.py:923] (0/4) Computing validation loss +2023-03-08 19:22:47,928 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2769, 4.6928, 4.7496, 4.5877, 4.2866, 4.3352, 4.8228, 4.6973], + device='cuda:0'), covar=tensor([0.1216, 0.0699, 0.0324, 0.0690, 0.1596, 0.0756, 0.0548, 0.0676], + device='cuda:0'), in_proj_covar=tensor([0.0393, 0.0336, 0.0262, 0.0369, 0.0492, 0.0362, 0.0414, 0.0341], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 19:22:48,701 INFO [train.py:932] (0/4) Epoch 5, validation: loss=0.1806, simple_loss=0.2829, pruned_loss=0.03918, over 944034.00 frames. +2023-03-08 19:22:48,701 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-08 19:22:58,833 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17545.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:23:09,828 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6090, 5.2229, 5.2977, 5.0228, 4.8069, 5.1036, 4.4073, 5.0231], + device='cuda:0'), covar=tensor([0.0241, 0.0314, 0.0185, 0.0231, 0.0375, 0.0202, 0.1165, 0.0291], + device='cuda:0'), in_proj_covar=tensor([0.0127, 0.0173, 0.0155, 0.0150, 0.0164, 0.0170, 0.0240, 0.0153], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0006, 0.0005], + device='cuda:0') +2023-03-08 19:23:19,290 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17562.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 19:23:30,589 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=17572.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 19:23:45,613 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5936, 5.5715, 5.1859, 5.4075, 4.6025, 5.3984, 5.6250, 5.3702], + device='cuda:0'), covar=tensor([0.2497, 0.0879, 0.0680, 0.0998, 0.2584, 0.0982, 0.0824, 0.1000], + device='cuda:0'), in_proj_covar=tensor([0.0391, 0.0330, 0.0255, 0.0364, 0.0488, 0.0356, 0.0409, 0.0338], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0005, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-08 19:23:47,552 INFO [train.py:898] (0/4) Epoch 5, batch 3050, loss[loss=0.2262, simple_loss=0.3128, pruned_loss=0.06975, over 18635.00 frames. ], tot_loss[loss=0.2343, simple_loss=0.3094, pruned_loss=0.07959, over 3584100.27 frames. ], batch size: 52, lr: 2.12e-02, grad_scale: 8.0 +2023-03-08 19:24:15,343 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=17610.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:24:33,684 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.113e+02 4.467e+02 5.746e+02 6.739e+02 1.721e+03, threshold=1.149e+03, percent-clipped=3.0 +2023-03-08 19:24:46,749 INFO [train.py:898] (0/4) Epoch 5, batch 3100, loss[loss=0.24, simple_loss=0.3089, pruned_loss=0.08557, over 18506.00 frames. ], tot_loss[loss=0.2351, simple_loss=0.3103, pruned_loss=0.07996, over 3573057.15 frames. ], batch size: 47, lr: 2.12e-02, grad_scale: 8.0 +2023-03-08 19:25:09,405 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1800, 5.1315, 4.5962, 5.1690, 5.0598, 4.5036, 5.0501, 4.6847], + device='cuda:0'), covar=tensor([0.0373, 0.0358, 0.1454, 0.0472, 0.0445, 0.0391, 0.0326, 0.0648], + device='cuda:0'), in_proj_covar=tensor([0.0292, 0.0331, 0.0479, 0.0260, 0.0250, 0.0311, 0.0319, 0.0412], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 19:25:12,375 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0882, 4.2516, 2.1426, 4.1224, 4.9946, 2.3085, 3.5393, 3.5140], + device='cuda:0'), covar=tensor([0.0115, 0.0902, 0.1787, 0.0499, 0.0064, 0.1528, 0.0739, 0.0969], + device='cuda:0'), in_proj_covar=tensor([0.0082, 0.0163, 0.0177, 0.0168, 0.0072, 0.0165, 0.0181, 0.0182], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 19:25:45,573 INFO [train.py:898] (0/4) Epoch 5, batch 3150, loss[loss=0.2646, simple_loss=0.3304, pruned_loss=0.09939, over 17056.00 frames. ], tot_loss[loss=0.2357, simple_loss=0.3109, pruned_loss=0.0803, over 3562596.11 frames. ], batch size: 78, lr: 2.12e-02, grad_scale: 8.0 +2023-03-08 19:25:50,468 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7660, 4.3773, 4.5344, 3.5914, 3.4714, 3.6272, 2.4270, 1.8304], + device='cuda:0'), covar=tensor([0.0163, 0.0148, 0.0056, 0.0234, 0.0347, 0.0139, 0.0697, 0.0991], + device='cuda:0'), in_proj_covar=tensor([0.0040, 0.0037, 0.0032, 0.0047, 0.0064, 0.0040, 0.0062, 0.0069], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 19:26:00,182 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17700.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:26:12,136 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1357, 4.8480, 5.2867, 5.1626, 4.9602, 5.7397, 5.4231, 5.1282], + device='cuda:0'), covar=tensor([0.0708, 0.0606, 0.0602, 0.0417, 0.1177, 0.0641, 0.0470, 0.1356], + device='cuda:0'), in_proj_covar=tensor([0.0252, 0.0189, 0.0195, 0.0182, 0.0231, 0.0279, 0.0181, 0.0265], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 19:26:22,120 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17718.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:26:31,143 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.328e+02 4.460e+02 5.383e+02 6.414e+02 1.196e+03, threshold=1.077e+03, percent-clipped=2.0 +2023-03-08 19:26:35,975 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17730.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:26:44,014 INFO [train.py:898] (0/4) Epoch 5, batch 3200, loss[loss=0.2456, simple_loss=0.3311, pruned_loss=0.08011, over 18561.00 frames. ], tot_loss[loss=0.2358, simple_loss=0.3107, pruned_loss=0.08041, over 3554560.00 frames. ], batch size: 54, lr: 2.11e-02, grad_scale: 8.0 +2023-03-08 19:26:46,686 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=17739.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 19:27:12,034 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17761.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:27:19,347 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2810, 2.5112, 2.3979, 2.7137, 3.3668, 3.3483, 2.6778, 2.9427], + device='cuda:0'), covar=tensor([0.0367, 0.0326, 0.0635, 0.0383, 0.0232, 0.0128, 0.0456, 0.0304], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0074, 0.0130, 0.0101, 0.0076, 0.0056, 0.0095, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 19:27:42,483 INFO [train.py:898] (0/4) Epoch 5, batch 3250, loss[loss=0.2424, simple_loss=0.3193, pruned_loss=0.08272, over 18488.00 frames. ], tot_loss[loss=0.2358, simple_loss=0.3104, pruned_loss=0.08058, over 3557685.55 frames. ], batch size: 53, lr: 2.11e-02, grad_scale: 8.0 +2023-03-08 19:27:47,312 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=17791.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:27:58,075 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2500, 2.5808, 2.4135, 2.8540, 3.3549, 3.2148, 2.8067, 3.0234], + device='cuda:0'), covar=tensor([0.0315, 0.0345, 0.0800, 0.0338, 0.0216, 0.0195, 0.0447, 0.0268], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0075, 0.0132, 0.0100, 0.0076, 0.0057, 0.0095, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 19:28:03,610 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3044, 5.9653, 5.4079, 5.7907, 5.4286, 5.5164, 6.0537, 5.9629], + device='cuda:0'), covar=tensor([0.1161, 0.0577, 0.0416, 0.0621, 0.1395, 0.0574, 0.0442, 0.0585], + device='cuda:0'), in_proj_covar=tensor([0.0403, 0.0340, 0.0259, 0.0370, 0.0505, 0.0368, 0.0420, 0.0341], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 19:28:28,459 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.849e+02 4.466e+02 5.558e+02 6.730e+02 1.712e+03, threshold=1.112e+03, percent-clipped=5.0 +2023-03-08 19:28:40,498 INFO [train.py:898] (0/4) Epoch 5, batch 3300, loss[loss=0.2215, simple_loss=0.2962, pruned_loss=0.07339, over 18547.00 frames. ], tot_loss[loss=0.2334, simple_loss=0.308, pruned_loss=0.07938, over 3575090.06 frames. ], batch size: 49, lr: 2.11e-02, grad_scale: 16.0 +2023-03-08 19:28:50,298 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=17845.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:29:39,232 INFO [train.py:898] (0/4) Epoch 5, batch 3350, loss[loss=0.2069, simple_loss=0.2911, pruned_loss=0.06128, over 18393.00 frames. ], tot_loss[loss=0.2321, simple_loss=0.3071, pruned_loss=0.07856, over 3589677.41 frames. ], batch size: 48, lr: 2.11e-02, grad_scale: 16.0 +2023-03-08 19:29:46,128 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=17893.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:30:25,122 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.611e+02 4.192e+02 5.439e+02 6.827e+02 1.413e+03, threshold=1.088e+03, percent-clipped=2.0 +2023-03-08 19:30:38,004 INFO [train.py:898] (0/4) Epoch 5, batch 3400, loss[loss=0.2244, simple_loss=0.3092, pruned_loss=0.06979, over 18478.00 frames. ], tot_loss[loss=0.2319, simple_loss=0.3072, pruned_loss=0.07826, over 3604120.97 frames. ], batch size: 51, lr: 2.10e-02, grad_scale: 16.0 +2023-03-08 19:30:50,737 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3501, 4.8667, 4.8618, 4.8270, 4.4960, 4.6669, 4.0561, 4.7164], + device='cuda:0'), covar=tensor([0.0224, 0.0266, 0.0237, 0.0213, 0.0358, 0.0303, 0.1268, 0.0256], + device='cuda:0'), in_proj_covar=tensor([0.0125, 0.0171, 0.0158, 0.0154, 0.0164, 0.0172, 0.0241, 0.0153], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0006, 0.0005], + device='cuda:0') +2023-03-08 19:30:55,940 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3629, 4.3930, 2.7125, 4.5699, 5.3773, 2.5665, 4.0239, 4.1359], + device='cuda:0'), covar=tensor([0.0042, 0.0761, 0.1249, 0.0360, 0.0035, 0.1205, 0.0491, 0.0611], + device='cuda:0'), in_proj_covar=tensor([0.0083, 0.0164, 0.0174, 0.0168, 0.0072, 0.0161, 0.0178, 0.0182], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 19:31:04,955 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=17960.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:31:36,926 INFO [train.py:898] (0/4) Epoch 5, batch 3450, loss[loss=0.2505, simple_loss=0.3286, pruned_loss=0.08619, over 17791.00 frames. ], tot_loss[loss=0.232, simple_loss=0.3075, pruned_loss=0.07828, over 3606436.90 frames. ], batch size: 70, lr: 2.10e-02, grad_scale: 16.0 +2023-03-08 19:31:51,170 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-18000.pt +2023-03-08 19:31:58,623 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.4756, 5.5552, 3.2103, 5.3652, 5.1368, 5.6480, 5.4045, 2.7996], + device='cuda:0'), covar=tensor([0.0101, 0.0037, 0.0566, 0.0053, 0.0059, 0.0039, 0.0060, 0.0956], + device='cuda:0'), in_proj_covar=tensor([0.0061, 0.0050, 0.0080, 0.0064, 0.0061, 0.0051, 0.0065, 0.0086], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0003, 0.0004], + device='cuda:0') +2023-03-08 19:32:17,347 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18018.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:32:20,899 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18021.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:32:26,802 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.873e+02 4.562e+02 5.412e+02 6.904e+02 1.746e+03, threshold=1.082e+03, percent-clipped=5.0 +2023-03-08 19:32:39,599 INFO [train.py:898] (0/4) Epoch 5, batch 3500, loss[loss=0.2073, simple_loss=0.2926, pruned_loss=0.06101, over 18362.00 frames. ], tot_loss[loss=0.2326, simple_loss=0.3077, pruned_loss=0.07875, over 3598945.67 frames. ], batch size: 46, lr: 2.10e-02, grad_scale: 16.0 +2023-03-08 19:32:42,107 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18039.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 19:33:01,673 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18056.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:33:12,381 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=18066.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:33:31,000 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-03-08 19:33:33,528 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18086.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:33:34,498 INFO [train.py:898] (0/4) Epoch 5, batch 3550, loss[loss=0.2178, simple_loss=0.294, pruned_loss=0.07082, over 18505.00 frames. ], tot_loss[loss=0.2318, simple_loss=0.3071, pruned_loss=0.07826, over 3594647.92 frames. ], batch size: 47, lr: 2.09e-02, grad_scale: 16.0 +2023-03-08 19:33:34,627 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=18087.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 19:34:17,502 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.924e+02 4.145e+02 5.139e+02 6.642e+02 1.282e+03, threshold=1.028e+03, percent-clipped=2.0 +2023-03-08 19:34:29,569 INFO [train.py:898] (0/4) Epoch 5, batch 3600, loss[loss=0.2345, simple_loss=0.3144, pruned_loss=0.07735, over 18621.00 frames. ], tot_loss[loss=0.233, simple_loss=0.308, pruned_loss=0.07902, over 3580729.95 frames. ], batch size: 52, lr: 2.09e-02, grad_scale: 16.0 +2023-03-08 19:34:30,868 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5765, 2.4493, 2.4181, 2.8589, 3.3002, 3.2223, 3.0006, 3.1116], + device='cuda:0'), covar=tensor([0.0287, 0.0359, 0.0835, 0.0312, 0.0280, 0.0204, 0.0435, 0.0325], + device='cuda:0'), in_proj_covar=tensor([0.0097, 0.0075, 0.0127, 0.0098, 0.0074, 0.0055, 0.0095, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 19:35:00,679 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5544, 3.3250, 4.0415, 3.0681, 3.5294, 2.7612, 2.5035, 2.4225], + device='cuda:0'), covar=tensor([0.0579, 0.0471, 0.0063, 0.0274, 0.0388, 0.1234, 0.1425, 0.0943], + device='cuda:0'), in_proj_covar=tensor([0.0163, 0.0170, 0.0083, 0.0132, 0.0184, 0.0219, 0.0196, 0.0179], + device='cuda:0'), out_proj_covar=tensor([1.5546e-04, 1.6897e-04, 8.4985e-05, 1.2996e-04, 1.8214e-04, 2.1178e-04, + 1.9884e-04, 1.7818e-04], device='cuda:0') +2023-03-08 19:35:05,414 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-5.pt +2023-03-08 19:35:34,900 INFO [train.py:898] (0/4) Epoch 6, batch 0, loss[loss=0.2343, simple_loss=0.3075, pruned_loss=0.08058, over 18536.00 frames. ], tot_loss[loss=0.2343, simple_loss=0.3075, pruned_loss=0.08058, over 18536.00 frames. ], batch size: 49, lr: 1.95e-02, grad_scale: 16.0 +2023-03-08 19:35:34,902 INFO [train.py:923] (0/4) Computing validation loss +2023-03-08 19:35:45,329 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.6807, 5.5172, 5.6180, 5.4229, 5.4948, 5.4793, 5.7528, 5.7640], + device='cuda:0'), covar=tensor([0.0039, 0.0067, 0.0059, 0.0063, 0.0046, 0.0070, 0.0057, 0.0068], + device='cuda:0'), in_proj_covar=tensor([0.0061, 0.0046, 0.0047, 0.0058, 0.0052, 0.0068, 0.0057, 0.0056], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 19:35:46,592 INFO [train.py:932] (0/4) Epoch 6, validation: loss=0.1816, simple_loss=0.2843, pruned_loss=0.0395, over 944034.00 frames. +2023-03-08 19:35:46,592 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-08 19:35:50,262 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18174.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:35:51,404 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18175.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:36:44,909 INFO [train.py:898] (0/4) Epoch 6, batch 50, loss[loss=0.2402, simple_loss=0.3248, pruned_loss=0.07783, over 18346.00 frames. ], tot_loss[loss=0.2325, simple_loss=0.309, pruned_loss=0.07803, over 818989.51 frames. ], batch size: 55, lr: 1.95e-02, grad_scale: 8.0 +2023-03-08 19:36:50,386 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9200, 3.9626, 4.6945, 3.1118, 4.0560, 2.5396, 2.9170, 2.1097], + device='cuda:0'), covar=tensor([0.0651, 0.0476, 0.0050, 0.0415, 0.0508, 0.1790, 0.1665, 0.1411], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0174, 0.0084, 0.0134, 0.0189, 0.0222, 0.0201, 0.0182], + device='cuda:0'), out_proj_covar=tensor([1.5783e-04, 1.7274e-04, 8.6460e-05, 1.3212e-04, 1.8624e-04, 2.1568e-04, + 2.0339e-04, 1.8052e-04], device='cuda:0') +2023-03-08 19:36:52,252 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.724e+02 4.768e+02 5.691e+02 6.790e+02 1.877e+03, threshold=1.138e+03, percent-clipped=9.0 +2023-03-08 19:37:01,471 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18235.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:37:02,637 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18236.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:37:32,460 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9816, 3.9841, 4.9359, 3.2740, 4.2102, 2.7144, 3.2722, 2.1666], + device='cuda:0'), covar=tensor([0.0644, 0.0498, 0.0044, 0.0365, 0.0454, 0.1609, 0.1433, 0.1344], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0176, 0.0085, 0.0137, 0.0192, 0.0224, 0.0203, 0.0184], + device='cuda:0'), out_proj_covar=tensor([1.5957e-04, 1.7492e-04, 8.7140e-05, 1.3413e-04, 1.8873e-04, 2.1776e-04, + 2.0580e-04, 1.8241e-04], device='cuda:0') +2023-03-08 19:37:43,155 INFO [train.py:898] (0/4) Epoch 6, batch 100, loss[loss=0.2137, simple_loss=0.2972, pruned_loss=0.0651, over 18571.00 frames. ], tot_loss[loss=0.2303, simple_loss=0.3074, pruned_loss=0.07657, over 1440773.52 frames. ], batch size: 54, lr: 1.95e-02, grad_scale: 8.0 +2023-03-08 19:38:36,365 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18316.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:38:41,806 INFO [train.py:898] (0/4) Epoch 6, batch 150, loss[loss=0.1924, simple_loss=0.2702, pruned_loss=0.0573, over 18508.00 frames. ], tot_loss[loss=0.23, simple_loss=0.3069, pruned_loss=0.07653, over 1909311.30 frames. ], batch size: 44, lr: 1.94e-02, grad_scale: 8.0 +2023-03-08 19:38:48,541 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.997e+02 4.126e+02 4.935e+02 6.131e+02 1.362e+03, threshold=9.869e+02, percent-clipped=2.0 +2023-03-08 19:39:22,925 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18356.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:39:37,045 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2192, 5.2561, 4.2481, 4.9918, 5.2316, 4.7260, 5.0335, 4.7027], + device='cuda:0'), covar=tensor([0.0514, 0.0540, 0.2490, 0.1004, 0.0527, 0.0468, 0.0662, 0.1130], + device='cuda:0'), in_proj_covar=tensor([0.0307, 0.0348, 0.0502, 0.0278, 0.0252, 0.0329, 0.0337, 0.0436], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-08 19:39:40,016 INFO [train.py:898] (0/4) Epoch 6, batch 200, loss[loss=0.2396, simple_loss=0.322, pruned_loss=0.07854, over 18251.00 frames. ], tot_loss[loss=0.2305, simple_loss=0.307, pruned_loss=0.07705, over 2293949.64 frames. ], batch size: 57, lr: 1.94e-02, grad_scale: 8.0 +2023-03-08 19:39:57,821 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18386.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:40:19,045 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=18404.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:40:39,255 INFO [train.py:898] (0/4) Epoch 6, batch 250, loss[loss=0.1978, simple_loss=0.2765, pruned_loss=0.05956, over 18474.00 frames. ], tot_loss[loss=0.231, simple_loss=0.3076, pruned_loss=0.07718, over 2578637.43 frames. ], batch size: 47, lr: 1.94e-02, grad_scale: 8.0 +2023-03-08 19:40:46,043 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.187e+02 4.523e+02 5.806e+02 6.957e+02 1.437e+03, threshold=1.161e+03, percent-clipped=4.0 +2023-03-08 19:40:47,054 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-08 19:40:54,327 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=18434.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:41:38,853 INFO [train.py:898] (0/4) Epoch 6, batch 300, loss[loss=0.3103, simple_loss=0.3648, pruned_loss=0.1279, over 12869.00 frames. ], tot_loss[loss=0.2304, simple_loss=0.3068, pruned_loss=0.07704, over 2792053.58 frames. ], batch size: 131, lr: 1.94e-02, grad_scale: 8.0 +2023-03-08 19:42:01,388 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18490.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:42:21,820 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18508.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 19:42:29,201 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1808, 5.3536, 3.0788, 5.1361, 5.0411, 5.4091, 5.1742, 2.5870], + device='cuda:0'), covar=tensor([0.0144, 0.0047, 0.0672, 0.0071, 0.0060, 0.0055, 0.0089, 0.1051], + device='cuda:0'), in_proj_covar=tensor([0.0063, 0.0051, 0.0084, 0.0066, 0.0062, 0.0053, 0.0067, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 19:42:36,902 INFO [train.py:898] (0/4) Epoch 6, batch 350, loss[loss=0.194, simple_loss=0.2773, pruned_loss=0.05532, over 18511.00 frames. ], tot_loss[loss=0.2294, simple_loss=0.3057, pruned_loss=0.07654, over 2970908.76 frames. ], batch size: 47, lr: 1.93e-02, grad_scale: 8.0 +2023-03-08 19:42:44,090 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.993e+02 4.095e+02 4.916e+02 6.743e+02 1.094e+03, threshold=9.831e+02, percent-clipped=0.0 +2023-03-08 19:42:47,459 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18530.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:42:48,562 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18531.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:43:11,102 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18551.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:43:11,551 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-03-08 19:43:31,788 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18569.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 19:43:33,751 INFO [train.py:898] (0/4) Epoch 6, batch 400, loss[loss=0.2382, simple_loss=0.3203, pruned_loss=0.07806, over 18368.00 frames. ], tot_loss[loss=0.2287, simple_loss=0.3048, pruned_loss=0.07631, over 3103282.37 frames. ], batch size: 55, lr: 1.93e-02, grad_scale: 8.0 +2023-03-08 19:44:26,299 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18616.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:44:31,603 INFO [train.py:898] (0/4) Epoch 6, batch 450, loss[loss=0.1971, simple_loss=0.2738, pruned_loss=0.06023, over 18338.00 frames. ], tot_loss[loss=0.2298, simple_loss=0.3058, pruned_loss=0.07689, over 3199393.89 frames. ], batch size: 46, lr: 1.93e-02, grad_scale: 8.0 +2023-03-08 19:44:38,640 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.446e+02 4.323e+02 5.172e+02 6.972e+02 1.405e+03, threshold=1.034e+03, percent-clipped=7.0 +2023-03-08 19:44:53,238 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4451, 3.1636, 1.8347, 4.0866, 2.8565, 4.2577, 2.1634, 3.7734], + device='cuda:0'), covar=tensor([0.0456, 0.0757, 0.1259, 0.0337, 0.0790, 0.0176, 0.1113, 0.0244], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0189, 0.0160, 0.0171, 0.0165, 0.0132, 0.0171, 0.0156], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 19:44:59,166 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.99 vs. limit=2.0 +2023-03-08 19:45:20,684 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=18664.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:45:29,552 INFO [train.py:898] (0/4) Epoch 6, batch 500, loss[loss=0.2118, simple_loss=0.2959, pruned_loss=0.06388, over 18282.00 frames. ], tot_loss[loss=0.2285, simple_loss=0.305, pruned_loss=0.07598, over 3296090.41 frames. ], batch size: 49, lr: 1.93e-02, grad_scale: 8.0 +2023-03-08 19:45:33,081 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18674.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:45:58,168 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.74 vs. limit=2.0 +2023-03-08 19:46:28,533 INFO [train.py:898] (0/4) Epoch 6, batch 550, loss[loss=0.1862, simple_loss=0.2575, pruned_loss=0.05746, over 18426.00 frames. ], tot_loss[loss=0.2295, simple_loss=0.3059, pruned_loss=0.07661, over 3355310.61 frames. ], batch size: 43, lr: 1.92e-02, grad_scale: 8.0 +2023-03-08 19:46:35,265 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.775e+02 4.304e+02 5.425e+02 6.689e+02 1.717e+03, threshold=1.085e+03, percent-clipped=5.0 +2023-03-08 19:46:45,322 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18735.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:47:20,488 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2276, 5.1534, 4.6235, 5.1181, 5.0655, 4.5409, 5.1001, 4.7528], + device='cuda:0'), covar=tensor([0.0400, 0.0351, 0.1480, 0.0710, 0.0457, 0.0416, 0.0346, 0.0821], + device='cuda:0'), in_proj_covar=tensor([0.0325, 0.0365, 0.0510, 0.0288, 0.0263, 0.0335, 0.0343, 0.0448], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-08 19:47:26,403 INFO [train.py:898] (0/4) Epoch 6, batch 600, loss[loss=0.243, simple_loss=0.3086, pruned_loss=0.08867, over 18371.00 frames. ], tot_loss[loss=0.2292, simple_loss=0.3053, pruned_loss=0.07654, over 3402161.55 frames. ], batch size: 46, lr: 1.92e-02, grad_scale: 8.0 +2023-03-08 19:47:29,593 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7795, 2.8438, 4.2336, 3.9070, 2.4389, 4.5472, 3.9661, 2.6252], + device='cuda:0'), covar=tensor([0.0292, 0.1004, 0.0106, 0.0202, 0.1385, 0.0112, 0.0258, 0.0889], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0190, 0.0106, 0.0116, 0.0191, 0.0147, 0.0154, 0.0177], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0001, 0.0001, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 19:47:49,622 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2099, 5.8804, 5.1831, 5.6536, 5.2953, 5.4524, 5.8659, 5.8075], + device='cuda:0'), covar=tensor([0.1139, 0.0520, 0.0500, 0.0574, 0.1356, 0.0606, 0.0466, 0.0530], + device='cuda:0'), in_proj_covar=tensor([0.0415, 0.0342, 0.0266, 0.0374, 0.0518, 0.0377, 0.0433, 0.0351], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 19:48:13,663 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5791, 2.9416, 4.0907, 3.8110, 2.4238, 4.3775, 3.8579, 2.7840], + device='cuda:0'), covar=tensor([0.0394, 0.0955, 0.0131, 0.0214, 0.1501, 0.0129, 0.0290, 0.0842], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0192, 0.0107, 0.0118, 0.0195, 0.0150, 0.0157, 0.0180], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 19:48:25,632 INFO [train.py:898] (0/4) Epoch 6, batch 650, loss[loss=0.2138, simple_loss=0.2973, pruned_loss=0.06513, over 18267.00 frames. ], tot_loss[loss=0.2287, simple_loss=0.3051, pruned_loss=0.0762, over 3440635.60 frames. ], batch size: 47, lr: 1.92e-02, grad_scale: 8.0 +2023-03-08 19:48:33,820 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.489e+02 4.300e+02 5.176e+02 5.986e+02 1.905e+03, threshold=1.035e+03, percent-clipped=4.0 +2023-03-08 19:48:37,415 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18830.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:48:38,658 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=18831.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:48:56,014 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18846.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:49:01,907 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1910, 4.1347, 2.4673, 4.3316, 5.1798, 2.2420, 3.9816, 4.0182], + device='cuda:0'), covar=tensor([0.0058, 0.0737, 0.1342, 0.0388, 0.0034, 0.1357, 0.0544, 0.0561], + device='cuda:0'), in_proj_covar=tensor([0.0086, 0.0167, 0.0173, 0.0171, 0.0074, 0.0162, 0.0183, 0.0181], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 19:49:17,267 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=18864.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 19:49:25,150 INFO [train.py:898] (0/4) Epoch 6, batch 700, loss[loss=0.2315, simple_loss=0.311, pruned_loss=0.07602, over 18275.00 frames. ], tot_loss[loss=0.2274, simple_loss=0.3038, pruned_loss=0.07549, over 3469192.61 frames. ], batch size: 49, lr: 1.92e-02, grad_scale: 8.0 +2023-03-08 19:49:27,776 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.1915, 3.5226, 4.8637, 3.9726, 2.8922, 2.5955, 4.0323, 4.8147], + device='cuda:0'), covar=tensor([0.0921, 0.1228, 0.0057, 0.0300, 0.0861, 0.1069, 0.0341, 0.0132], + device='cuda:0'), in_proj_covar=tensor([0.0128, 0.0179, 0.0069, 0.0137, 0.0160, 0.0159, 0.0138, 0.0086], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0001, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 19:49:33,614 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=18878.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:49:34,637 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=18879.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:50:23,144 INFO [train.py:898] (0/4) Epoch 6, batch 750, loss[loss=0.2081, simple_loss=0.2899, pruned_loss=0.06321, over 18271.00 frames. ], tot_loss[loss=0.2272, simple_loss=0.3038, pruned_loss=0.07532, over 3501380.22 frames. ], batch size: 49, lr: 1.91e-02, grad_scale: 8.0 +2023-03-08 19:50:29,059 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5535, 6.1902, 5.4821, 6.0418, 5.6384, 5.8055, 6.2275, 6.2014], + device='cuda:0'), covar=tensor([0.1202, 0.0505, 0.0375, 0.0519, 0.1371, 0.0602, 0.0441, 0.0498], + device='cuda:0'), in_proj_covar=tensor([0.0419, 0.0345, 0.0267, 0.0374, 0.0519, 0.0379, 0.0436, 0.0356], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 19:50:29,879 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.737e+02 4.561e+02 5.480e+02 6.846e+02 1.883e+03, threshold=1.096e+03, percent-clipped=6.0 +2023-03-08 19:50:33,947 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=18930.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:50:41,273 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2710, 5.3791, 3.2384, 5.1953, 5.0272, 5.4772, 5.1744, 2.8400], + device='cuda:0'), covar=tensor([0.0121, 0.0043, 0.0565, 0.0050, 0.0064, 0.0038, 0.0082, 0.0855], + device='cuda:0'), in_proj_covar=tensor([0.0063, 0.0051, 0.0082, 0.0064, 0.0061, 0.0051, 0.0066, 0.0087], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0003, 0.0004], + device='cuda:0') +2023-03-08 19:51:21,250 INFO [train.py:898] (0/4) Epoch 6, batch 800, loss[loss=0.1896, simple_loss=0.265, pruned_loss=0.05714, over 18402.00 frames. ], tot_loss[loss=0.2253, simple_loss=0.3018, pruned_loss=0.07435, over 3523385.27 frames. ], batch size: 42, lr: 1.91e-02, grad_scale: 8.0 +2023-03-08 19:51:46,386 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=18991.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:52:20,899 INFO [train.py:898] (0/4) Epoch 6, batch 850, loss[loss=0.2364, simple_loss=0.3201, pruned_loss=0.07634, over 18344.00 frames. ], tot_loss[loss=0.2246, simple_loss=0.3009, pruned_loss=0.07415, over 3543097.04 frames. ], batch size: 55, lr: 1.91e-02, grad_scale: 8.0 +2023-03-08 19:52:28,231 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.682e+02 3.950e+02 4.754e+02 5.949e+02 2.076e+03, threshold=9.508e+02, percent-clipped=3.0 +2023-03-08 19:52:31,927 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19030.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:53:19,106 INFO [train.py:898] (0/4) Epoch 6, batch 900, loss[loss=0.2226, simple_loss=0.3035, pruned_loss=0.07083, over 17952.00 frames. ], tot_loss[loss=0.2252, simple_loss=0.3016, pruned_loss=0.07442, over 3556532.37 frames. ], batch size: 65, lr: 1.91e-02, grad_scale: 8.0 +2023-03-08 19:53:41,488 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.98 vs. limit=2.0 +2023-03-08 19:54:17,490 INFO [train.py:898] (0/4) Epoch 6, batch 950, loss[loss=0.26, simple_loss=0.3375, pruned_loss=0.09126, over 18190.00 frames. ], tot_loss[loss=0.2265, simple_loss=0.3025, pruned_loss=0.07518, over 3556516.75 frames. ], batch size: 62, lr: 1.90e-02, grad_scale: 8.0 +2023-03-08 19:54:24,265 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.702e+02 4.325e+02 5.195e+02 6.173e+02 1.123e+03, threshold=1.039e+03, percent-clipped=4.0 +2023-03-08 19:54:26,812 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=19129.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 19:54:47,340 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19146.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:54:53,386 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.66 vs. limit=2.0 +2023-03-08 19:55:08,095 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19164.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 19:55:15,683 INFO [train.py:898] (0/4) Epoch 6, batch 1000, loss[loss=0.2219, simple_loss=0.3026, pruned_loss=0.07058, over 18475.00 frames. ], tot_loss[loss=0.2266, simple_loss=0.3028, pruned_loss=0.07514, over 3565627.40 frames. ], batch size: 51, lr: 1.90e-02, grad_scale: 8.0 +2023-03-08 19:55:38,150 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=19190.0, num_to_drop=1, layers_to_drop={3} +2023-03-08 19:55:42,510 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=19194.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:55:49,569 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.2288, 3.9495, 5.2846, 4.0964, 2.3787, 2.5326, 3.9936, 5.3066], + device='cuda:0'), covar=tensor([0.0993, 0.1217, 0.0078, 0.0383, 0.1168, 0.1312, 0.0443, 0.0071], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0182, 0.0070, 0.0137, 0.0159, 0.0160, 0.0139, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0001, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 19:56:05,319 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=19212.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 19:56:15,293 INFO [train.py:898] (0/4) Epoch 6, batch 1050, loss[loss=0.2141, simple_loss=0.3022, pruned_loss=0.063, over 18625.00 frames. ], tot_loss[loss=0.2259, simple_loss=0.3028, pruned_loss=0.07455, over 3571782.32 frames. ], batch size: 52, lr: 1.90e-02, grad_scale: 8.0 +2023-03-08 19:56:21,989 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.826e+02 4.167e+02 5.198e+02 6.760e+02 1.282e+03, threshold=1.040e+03, percent-clipped=3.0 +2023-03-08 19:57:14,159 INFO [train.py:898] (0/4) Epoch 6, batch 1100, loss[loss=0.2485, simple_loss=0.319, pruned_loss=0.08905, over 18290.00 frames. ], tot_loss[loss=0.2258, simple_loss=0.3031, pruned_loss=0.07427, over 3576922.39 frames. ], batch size: 57, lr: 1.90e-02, grad_scale: 8.0 +2023-03-08 19:57:20,056 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8625, 4.5729, 4.7485, 3.3714, 3.7965, 3.7764, 2.7741, 2.0519], + device='cuda:0'), covar=tensor([0.0180, 0.0184, 0.0049, 0.0241, 0.0281, 0.0157, 0.0637, 0.0898], + device='cuda:0'), in_proj_covar=tensor([0.0042, 0.0039, 0.0034, 0.0048, 0.0065, 0.0042, 0.0064, 0.0069], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 19:57:31,192 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19286.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:58:13,169 INFO [train.py:898] (0/4) Epoch 6, batch 1150, loss[loss=0.212, simple_loss=0.2889, pruned_loss=0.06757, over 18495.00 frames. ], tot_loss[loss=0.2248, simple_loss=0.3019, pruned_loss=0.07386, over 3574687.26 frames. ], batch size: 47, lr: 1.90e-02, grad_scale: 8.0 +2023-03-08 19:58:19,718 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.58 vs. limit=2.0 +2023-03-08 19:58:20,058 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.317e+02 3.874e+02 4.905e+02 6.066e+02 2.100e+03, threshold=9.811e+02, percent-clipped=2.0 +2023-03-08 19:58:23,655 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19330.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:59:11,796 INFO [train.py:898] (0/4) Epoch 6, batch 1200, loss[loss=0.264, simple_loss=0.3295, pruned_loss=0.09921, over 12862.00 frames. ], tot_loss[loss=0.2238, simple_loss=0.3007, pruned_loss=0.07349, over 3573504.65 frames. ], batch size: 131, lr: 1.89e-02, grad_scale: 8.0 +2023-03-08 19:59:19,973 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=19378.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:59:37,893 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=19394.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 19:59:44,441 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-08 19:59:50,517 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-08 19:59:51,720 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-03-08 20:00:10,013 INFO [train.py:898] (0/4) Epoch 6, batch 1250, loss[loss=0.1764, simple_loss=0.2513, pruned_loss=0.05075, over 18435.00 frames. ], tot_loss[loss=0.2241, simple_loss=0.3008, pruned_loss=0.07368, over 3576248.25 frames. ], batch size: 43, lr: 1.89e-02, grad_scale: 8.0 +2023-03-08 20:00:16,821 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.768e+02 4.335e+02 5.425e+02 6.798e+02 1.467e+03, threshold=1.085e+03, percent-clipped=6.0 +2023-03-08 20:00:29,291 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-08 20:00:49,560 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=19455.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:01:08,435 INFO [train.py:898] (0/4) Epoch 6, batch 1300, loss[loss=0.2157, simple_loss=0.3036, pruned_loss=0.06392, over 18023.00 frames. ], tot_loss[loss=0.2243, simple_loss=0.301, pruned_loss=0.07375, over 3569031.19 frames. ], batch size: 65, lr: 1.89e-02, grad_scale: 8.0 +2023-03-08 20:01:25,086 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19485.0, num_to_drop=1, layers_to_drop={3} +2023-03-08 20:01:54,018 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.71 vs. limit=5.0 +2023-03-08 20:02:07,610 INFO [train.py:898] (0/4) Epoch 6, batch 1350, loss[loss=0.2155, simple_loss=0.3042, pruned_loss=0.06341, over 18331.00 frames. ], tot_loss[loss=0.2245, simple_loss=0.3015, pruned_loss=0.07372, over 3576096.93 frames. ], batch size: 56, lr: 1.89e-02, grad_scale: 8.0 +2023-03-08 20:02:15,078 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.530e+02 3.979e+02 4.823e+02 6.493e+02 1.049e+03, threshold=9.645e+02, percent-clipped=0.0 +2023-03-08 20:02:30,160 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-08 20:02:59,302 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8487, 4.6982, 4.9417, 4.6619, 4.6916, 4.7727, 5.1435, 5.0728], + device='cuda:0'), covar=tensor([0.0057, 0.0077, 0.0070, 0.0083, 0.0067, 0.0079, 0.0088, 0.0102], + device='cuda:0'), in_proj_covar=tensor([0.0066, 0.0051, 0.0049, 0.0064, 0.0054, 0.0072, 0.0062, 0.0060], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 20:03:05,912 INFO [train.py:898] (0/4) Epoch 6, batch 1400, loss[loss=0.238, simple_loss=0.3221, pruned_loss=0.07698, over 18608.00 frames. ], tot_loss[loss=0.2257, simple_loss=0.3028, pruned_loss=0.07435, over 3568626.03 frames. ], batch size: 52, lr: 1.88e-02, grad_scale: 8.0 +2023-03-08 20:03:06,172 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4380, 5.8898, 5.4120, 5.7050, 5.4379, 5.4364, 5.9579, 5.8439], + device='cuda:0'), covar=tensor([0.0926, 0.0573, 0.0471, 0.0622, 0.1293, 0.0579, 0.0453, 0.0577], + device='cuda:0'), in_proj_covar=tensor([0.0421, 0.0343, 0.0267, 0.0375, 0.0520, 0.0373, 0.0442, 0.0346], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-08 20:03:08,890 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.94 vs. limit=2.0 +2023-03-08 20:03:23,663 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19586.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:03:54,380 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-08 20:04:01,552 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=19618.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:04:04,685 INFO [train.py:898] (0/4) Epoch 6, batch 1450, loss[loss=0.2265, simple_loss=0.3106, pruned_loss=0.07124, over 18378.00 frames. ], tot_loss[loss=0.2257, simple_loss=0.3029, pruned_loss=0.07422, over 3580768.70 frames. ], batch size: 55, lr: 1.88e-02, grad_scale: 8.0 +2023-03-08 20:04:06,548 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.98 vs. limit=2.0 +2023-03-08 20:04:11,634 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.837e+02 4.168e+02 4.993e+02 6.128e+02 1.216e+03, threshold=9.987e+02, percent-clipped=6.0 +2023-03-08 20:04:20,291 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=19634.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:04:23,121 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.01 vs. limit=5.0 +2023-03-08 20:04:47,357 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7970, 4.6702, 4.8406, 4.5509, 4.5335, 4.6885, 4.9812, 5.0240], + device='cuda:0'), covar=tensor([0.0062, 0.0087, 0.0074, 0.0104, 0.0077, 0.0084, 0.0086, 0.0105], + device='cuda:0'), in_proj_covar=tensor([0.0068, 0.0052, 0.0050, 0.0066, 0.0055, 0.0074, 0.0065, 0.0062], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 20:05:03,112 INFO [train.py:898] (0/4) Epoch 6, batch 1500, loss[loss=0.2262, simple_loss=0.3062, pruned_loss=0.07303, over 18380.00 frames. ], tot_loss[loss=0.2239, simple_loss=0.3012, pruned_loss=0.07331, over 3591742.81 frames. ], batch size: 52, lr: 1.88e-02, grad_scale: 8.0 +2023-03-08 20:05:12,447 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=19679.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:05:20,792 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9388, 1.8432, 3.1350, 2.7385, 3.7340, 5.2239, 4.2938, 4.5248], + device='cuda:0'), covar=tensor([0.0473, 0.1188, 0.0977, 0.0793, 0.1024, 0.0034, 0.0270, 0.0160], + device='cuda:0'), in_proj_covar=tensor([0.0166, 0.0218, 0.0201, 0.0204, 0.0299, 0.0122, 0.0186, 0.0153], + device='cuda:0'), out_proj_covar=tensor([1.1766e-04, 1.5633e-04, 1.5277e-04, 1.3553e-04, 2.1461e-04, 8.1897e-05, + 1.3135e-04, 1.0880e-04], device='cuda:0') +2023-03-08 20:05:22,225 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-03-08 20:05:22,528 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.13 vs. limit=2.0 +2023-03-08 20:06:01,661 INFO [train.py:898] (0/4) Epoch 6, batch 1550, loss[loss=0.2734, simple_loss=0.3291, pruned_loss=0.1089, over 12673.00 frames. ], tot_loss[loss=0.2242, simple_loss=0.3014, pruned_loss=0.07355, over 3566367.53 frames. ], batch size: 130, lr: 1.88e-02, grad_scale: 8.0 +2023-03-08 20:06:09,096 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.820e+02 4.112e+02 5.168e+02 6.335e+02 1.578e+03, threshold=1.034e+03, percent-clipped=4.0 +2023-03-08 20:06:18,114 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.79 vs. limit=5.0 +2023-03-08 20:06:36,314 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19750.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:07:01,072 INFO [train.py:898] (0/4) Epoch 6, batch 1600, loss[loss=0.2312, simple_loss=0.3119, pruned_loss=0.07526, over 18634.00 frames. ], tot_loss[loss=0.2243, simple_loss=0.3014, pruned_loss=0.0736, over 3567795.47 frames. ], batch size: 52, lr: 1.87e-02, grad_scale: 8.0 +2023-03-08 20:07:17,637 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=19785.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 20:07:58,560 INFO [train.py:898] (0/4) Epoch 6, batch 1650, loss[loss=0.2132, simple_loss=0.2958, pruned_loss=0.06529, over 18265.00 frames. ], tot_loss[loss=0.225, simple_loss=0.3023, pruned_loss=0.07382, over 3579199.57 frames. ], batch size: 47, lr: 1.87e-02, grad_scale: 8.0 +2023-03-08 20:08:06,542 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.131e+02 4.738e+02 5.632e+02 7.460e+02 1.782e+03, threshold=1.126e+03, percent-clipped=8.0 +2023-03-08 20:08:13,431 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=19833.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 20:08:26,204 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7277, 2.3507, 2.1344, 2.4145, 2.9752, 2.8520, 2.5156, 2.5012], + device='cuda:0'), covar=tensor([0.0205, 0.0228, 0.0675, 0.0340, 0.0196, 0.0116, 0.0352, 0.0344], + device='cuda:0'), in_proj_covar=tensor([0.0102, 0.0074, 0.0133, 0.0107, 0.0078, 0.0060, 0.0100, 0.0105], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 20:08:29,482 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8165, 4.3453, 4.4561, 3.5059, 3.5210, 3.5528, 2.0219, 1.9561], + device='cuda:0'), covar=tensor([0.0167, 0.0156, 0.0050, 0.0272, 0.0378, 0.0208, 0.0905, 0.1118], + device='cuda:0'), in_proj_covar=tensor([0.0044, 0.0042, 0.0036, 0.0050, 0.0070, 0.0045, 0.0067, 0.0073], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004], + device='cuda:0') +2023-03-08 20:08:56,802 INFO [train.py:898] (0/4) Epoch 6, batch 1700, loss[loss=0.2478, simple_loss=0.3252, pruned_loss=0.08523, over 16959.00 frames. ], tot_loss[loss=0.2236, simple_loss=0.301, pruned_loss=0.07315, over 3587513.42 frames. ], batch size: 78, lr: 1.87e-02, grad_scale: 8.0 +2023-03-08 20:09:55,326 INFO [train.py:898] (0/4) Epoch 6, batch 1750, loss[loss=0.2106, simple_loss=0.2847, pruned_loss=0.06826, over 18250.00 frames. ], tot_loss[loss=0.2229, simple_loss=0.3004, pruned_loss=0.07269, over 3595870.23 frames. ], batch size: 45, lr: 1.87e-02, grad_scale: 8.0 +2023-03-08 20:10:02,903 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.466e+02 3.770e+02 4.789e+02 6.100e+02 1.481e+03, threshold=9.579e+02, percent-clipped=1.0 +2023-03-08 20:10:05,891 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.35 vs. limit=5.0 +2023-03-08 20:10:21,896 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3293, 4.0363, 5.0624, 3.0952, 4.3845, 2.7356, 2.9957, 2.1632], + device='cuda:0'), covar=tensor([0.0618, 0.0552, 0.0063, 0.0483, 0.0483, 0.1813, 0.1946, 0.1508], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0181, 0.0088, 0.0142, 0.0192, 0.0229, 0.0215, 0.0186], + device='cuda:0'), out_proj_covar=tensor([1.6113e-04, 1.7701e-04, 8.8400e-05, 1.3863e-04, 1.8645e-04, 2.1994e-04, + 2.1392e-04, 1.8218e-04], device='cuda:0') +2023-03-08 20:10:38,166 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7810, 3.8164, 3.8446, 3.7031, 3.7376, 3.6752, 4.0199, 3.9799], + device='cuda:0'), covar=tensor([0.0086, 0.0092, 0.0082, 0.0105, 0.0094, 0.0129, 0.0084, 0.0110], + device='cuda:0'), in_proj_covar=tensor([0.0067, 0.0051, 0.0050, 0.0065, 0.0055, 0.0073, 0.0063, 0.0061], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 20:10:45,110 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7191, 3.4286, 3.3321, 3.0673, 3.2592, 2.6928, 3.0507, 3.6565], + device='cuda:0'), covar=tensor([0.0025, 0.0087, 0.0070, 0.0127, 0.0077, 0.0166, 0.0133, 0.0047], + device='cuda:0'), in_proj_covar=tensor([0.0060, 0.0079, 0.0072, 0.0113, 0.0071, 0.0116, 0.0120, 0.0065], + device='cuda:0'), out_proj_covar=tensor([8.6279e-05, 1.2307e-04, 1.1150e-04, 1.8181e-04, 1.0829e-04, 1.8385e-04, + 1.8652e-04, 9.8090e-05], device='cuda:0') +2023-03-08 20:10:53,715 INFO [train.py:898] (0/4) Epoch 6, batch 1800, loss[loss=0.2324, simple_loss=0.3115, pruned_loss=0.07666, over 18325.00 frames. ], tot_loss[loss=0.2233, simple_loss=0.3008, pruned_loss=0.0729, over 3593224.97 frames. ], batch size: 54, lr: 1.87e-02, grad_scale: 8.0 +2023-03-08 20:10:57,367 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=19974.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:11:28,578 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-20000.pt +2023-03-08 20:11:55,782 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20020.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:11:56,565 INFO [train.py:898] (0/4) Epoch 6, batch 1850, loss[loss=0.2073, simple_loss=0.2874, pruned_loss=0.06359, over 18277.00 frames. ], tot_loss[loss=0.224, simple_loss=0.301, pruned_loss=0.07343, over 3588533.20 frames. ], batch size: 49, lr: 1.86e-02, grad_scale: 8.0 +2023-03-08 20:11:58,143 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-08 20:12:03,174 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.737e+02 4.293e+02 5.407e+02 6.655e+02 1.124e+03, threshold=1.081e+03, percent-clipped=3.0 +2023-03-08 20:12:31,175 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20050.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:12:55,162 INFO [train.py:898] (0/4) Epoch 6, batch 1900, loss[loss=0.2301, simple_loss=0.3075, pruned_loss=0.07638, over 18275.00 frames. ], tot_loss[loss=0.2244, simple_loss=0.3013, pruned_loss=0.07369, over 3591196.46 frames. ], batch size: 57, lr: 1.86e-02, grad_scale: 8.0 +2023-03-08 20:13:07,078 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20081.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:13:17,152 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20089.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:13:27,666 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=20098.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:13:29,170 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.01 vs. limit=5.0 +2023-03-08 20:13:44,259 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20112.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:13:54,337 INFO [train.py:898] (0/4) Epoch 6, batch 1950, loss[loss=0.2358, simple_loss=0.3125, pruned_loss=0.0795, over 18208.00 frames. ], tot_loss[loss=0.2229, simple_loss=0.3003, pruned_loss=0.0728, over 3590213.29 frames. ], batch size: 60, lr: 1.86e-02, grad_scale: 8.0 +2023-03-08 20:14:01,197 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.918e+02 4.095e+02 5.447e+02 6.778e+02 1.959e+03, threshold=1.089e+03, percent-clipped=6.0 +2023-03-08 20:14:28,669 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20150.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:14:52,815 INFO [train.py:898] (0/4) Epoch 6, batch 2000, loss[loss=0.251, simple_loss=0.3263, pruned_loss=0.08785, over 17001.00 frames. ], tot_loss[loss=0.2234, simple_loss=0.3006, pruned_loss=0.0731, over 3577962.57 frames. ], batch size: 78, lr: 1.86e-02, grad_scale: 8.0 +2023-03-08 20:14:55,585 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20173.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:15:32,050 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20204.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:15:47,645 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-08 20:15:51,543 INFO [train.py:898] (0/4) Epoch 6, batch 2050, loss[loss=0.2549, simple_loss=0.326, pruned_loss=0.09193, over 16090.00 frames. ], tot_loss[loss=0.2226, simple_loss=0.2999, pruned_loss=0.07264, over 3586153.21 frames. ], batch size: 94, lr: 1.86e-02, grad_scale: 4.0 +2023-03-08 20:15:59,260 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.572e+02 4.223e+02 5.010e+02 6.268e+02 1.616e+03, threshold=1.002e+03, percent-clipped=2.0 +2023-03-08 20:16:43,364 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20265.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:16:49,846 INFO [train.py:898] (0/4) Epoch 6, batch 2100, loss[loss=0.232, simple_loss=0.3057, pruned_loss=0.07917, over 18293.00 frames. ], tot_loss[loss=0.2218, simple_loss=0.2991, pruned_loss=0.07226, over 3596580.27 frames. ], batch size: 49, lr: 1.85e-02, grad_scale: 4.0 +2023-03-08 20:16:53,534 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20274.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:16:55,307 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-03-08 20:17:49,004 INFO [train.py:898] (0/4) Epoch 6, batch 2150, loss[loss=0.2082, simple_loss=0.2871, pruned_loss=0.06463, over 18400.00 frames. ], tot_loss[loss=0.2214, simple_loss=0.2987, pruned_loss=0.07203, over 3610893.98 frames. ], batch size: 52, lr: 1.85e-02, grad_scale: 4.0 +2023-03-08 20:17:50,291 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=20322.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:17:56,866 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.681e+02 4.510e+02 5.215e+02 6.724e+02 1.670e+03, threshold=1.043e+03, percent-clipped=8.0 +2023-03-08 20:18:47,256 INFO [train.py:898] (0/4) Epoch 6, batch 2200, loss[loss=0.2572, simple_loss=0.3375, pruned_loss=0.08851, over 18305.00 frames. ], tot_loss[loss=0.2217, simple_loss=0.299, pruned_loss=0.07225, over 3607732.90 frames. ], batch size: 57, lr: 1.85e-02, grad_scale: 4.0 +2023-03-08 20:18:53,150 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20376.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:19:06,974 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20388.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:19:42,884 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20418.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:19:46,525 INFO [train.py:898] (0/4) Epoch 6, batch 2250, loss[loss=0.2164, simple_loss=0.3042, pruned_loss=0.06425, over 18301.00 frames. ], tot_loss[loss=0.2224, simple_loss=0.2999, pruned_loss=0.07241, over 3606844.22 frames. ], batch size: 54, lr: 1.85e-02, grad_scale: 4.0 +2023-03-08 20:19:54,690 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.728e+02 4.214e+02 4.964e+02 6.088e+02 1.302e+03, threshold=9.929e+02, percent-clipped=3.0 +2023-03-08 20:20:13,847 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20445.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:20:18,780 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20449.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:20:29,467 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-03-08 20:20:41,596 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20468.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:20:42,049 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.60 vs. limit=2.0 +2023-03-08 20:20:44,843 INFO [train.py:898] (0/4) Epoch 6, batch 2300, loss[loss=0.1844, simple_loss=0.2588, pruned_loss=0.05503, over 18413.00 frames. ], tot_loss[loss=0.2215, simple_loss=0.299, pruned_loss=0.07203, over 3596828.96 frames. ], batch size: 43, lr: 1.84e-02, grad_scale: 4.0 +2023-03-08 20:20:54,713 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20479.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:21:18,268 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9191, 2.8526, 1.9552, 3.3424, 2.4569, 3.3137, 2.0447, 2.8561], + device='cuda:0'), covar=tensor([0.0430, 0.0728, 0.1114, 0.0434, 0.0763, 0.0258, 0.1037, 0.0357], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0195, 0.0167, 0.0181, 0.0170, 0.0152, 0.0175, 0.0165], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 20:21:43,767 INFO [train.py:898] (0/4) Epoch 6, batch 2350, loss[loss=0.25, simple_loss=0.3303, pruned_loss=0.08485, over 17809.00 frames. ], tot_loss[loss=0.2215, simple_loss=0.299, pruned_loss=0.07195, over 3603168.97 frames. ], batch size: 70, lr: 1.84e-02, grad_scale: 4.0 +2023-03-08 20:21:48,374 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2215, 5.1955, 2.5358, 5.0423, 4.8980, 5.3233, 5.1311, 2.4286], + device='cuda:0'), covar=tensor([0.0131, 0.0063, 0.0826, 0.0070, 0.0074, 0.0049, 0.0085, 0.1179], + device='cuda:0'), in_proj_covar=tensor([0.0064, 0.0054, 0.0082, 0.0066, 0.0063, 0.0053, 0.0068, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 20:21:52,077 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.562e+02 3.728e+02 4.764e+02 5.894e+02 1.500e+03, threshold=9.528e+02, percent-clipped=4.0 +2023-03-08 20:22:28,614 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20560.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:22:42,382 INFO [train.py:898] (0/4) Epoch 6, batch 2400, loss[loss=0.2536, simple_loss=0.3221, pruned_loss=0.09258, over 17107.00 frames. ], tot_loss[loss=0.221, simple_loss=0.2985, pruned_loss=0.07174, over 3598392.58 frames. ], batch size: 78, lr: 1.84e-02, grad_scale: 8.0 +2023-03-08 20:23:02,872 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=20588.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:23:08,411 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8595, 3.3894, 3.5029, 2.9930, 3.3927, 2.4993, 2.3101, 3.7278], + device='cuda:0'), covar=tensor([0.0021, 0.0084, 0.0051, 0.0098, 0.0072, 0.0182, 0.0221, 0.0049], + device='cuda:0'), in_proj_covar=tensor([0.0060, 0.0080, 0.0073, 0.0114, 0.0073, 0.0118, 0.0122, 0.0065], + device='cuda:0'), out_proj_covar=tensor([8.4097e-05, 1.2380e-04, 1.1297e-04, 1.8236e-04, 1.0987e-04, 1.8572e-04, + 1.9016e-04, 9.6470e-05], device='cuda:0') +2023-03-08 20:23:41,479 INFO [train.py:898] (0/4) Epoch 6, batch 2450, loss[loss=0.1986, simple_loss=0.2825, pruned_loss=0.0574, over 18391.00 frames. ], tot_loss[loss=0.2198, simple_loss=0.298, pruned_loss=0.07082, over 3605385.51 frames. ], batch size: 50, lr: 1.84e-02, grad_scale: 8.0 +2023-03-08 20:23:49,390 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.965e+02 4.013e+02 4.826e+02 5.800e+02 1.376e+03, threshold=9.653e+02, percent-clipped=2.0 +2023-03-08 20:24:13,855 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=20649.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:24:38,628 INFO [train.py:898] (0/4) Epoch 6, batch 2500, loss[loss=0.1899, simple_loss=0.2562, pruned_loss=0.06176, over 17738.00 frames. ], tot_loss[loss=0.2214, simple_loss=0.2991, pruned_loss=0.07182, over 3584674.99 frames. ], batch size: 39, lr: 1.84e-02, grad_scale: 8.0 +2023-03-08 20:24:44,503 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20676.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:24:56,496 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4201, 5.9856, 5.3992, 5.8189, 5.4929, 5.5388, 6.0271, 5.9465], + device='cuda:0'), covar=tensor([0.1017, 0.0554, 0.0391, 0.0550, 0.1421, 0.0653, 0.0532, 0.0570], + device='cuda:0'), in_proj_covar=tensor([0.0428, 0.0347, 0.0276, 0.0381, 0.0533, 0.0386, 0.0455, 0.0358], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-08 20:25:35,729 INFO [train.py:898] (0/4) Epoch 6, batch 2550, loss[loss=0.216, simple_loss=0.2977, pruned_loss=0.06716, over 17169.00 frames. ], tot_loss[loss=0.223, simple_loss=0.3007, pruned_loss=0.07269, over 3585042.11 frames. ], batch size: 78, lr: 1.83e-02, grad_scale: 4.0 +2023-03-08 20:25:36,108 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6056, 3.3268, 1.9382, 4.2370, 2.8869, 4.4104, 2.0899, 3.9192], + device='cuda:0'), covar=tensor([0.0432, 0.0732, 0.1345, 0.0331, 0.0868, 0.0181, 0.1227, 0.0281], + device='cuda:0'), in_proj_covar=tensor([0.0161, 0.0194, 0.0163, 0.0180, 0.0167, 0.0155, 0.0176, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 20:25:40,319 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=20724.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:25:45,642 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.233e+02 4.696e+02 5.623e+02 7.673e+02 1.890e+03, threshold=1.125e+03, percent-clipped=13.0 +2023-03-08 20:26:03,242 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20744.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:26:04,465 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20745.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:26:22,938 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-08 20:26:30,188 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20768.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:26:34,036 INFO [train.py:898] (0/4) Epoch 6, batch 2600, loss[loss=0.1886, simple_loss=0.2631, pruned_loss=0.05706, over 18163.00 frames. ], tot_loss[loss=0.2217, simple_loss=0.2993, pruned_loss=0.07203, over 3582402.87 frames. ], batch size: 44, lr: 1.83e-02, grad_scale: 4.0 +2023-03-08 20:26:38,084 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20774.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:26:52,600 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.83 vs. limit=2.0 +2023-03-08 20:27:00,480 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=20793.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:27:27,114 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=20816.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:27:32,622 INFO [train.py:898] (0/4) Epoch 6, batch 2650, loss[loss=0.2193, simple_loss=0.3028, pruned_loss=0.06792, over 18290.00 frames. ], tot_loss[loss=0.2213, simple_loss=0.2993, pruned_loss=0.07161, over 3587500.77 frames. ], batch size: 57, lr: 1.83e-02, grad_scale: 4.0 +2023-03-08 20:27:33,428 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-08 20:27:43,390 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.822e+02 3.946e+02 4.764e+02 5.553e+02 1.236e+03, threshold=9.528e+02, percent-clipped=1.0 +2023-03-08 20:28:16,096 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.62 vs. limit=2.0 +2023-03-08 20:28:19,213 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=20860.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:28:31,289 INFO [train.py:898] (0/4) Epoch 6, batch 2700, loss[loss=0.2625, simple_loss=0.3309, pruned_loss=0.09703, over 15973.00 frames. ], tot_loss[loss=0.2218, simple_loss=0.2999, pruned_loss=0.07184, over 3591513.49 frames. ], batch size: 94, lr: 1.83e-02, grad_scale: 4.0 +2023-03-08 20:29:14,535 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=20908.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:29:28,982 INFO [train.py:898] (0/4) Epoch 6, batch 2750, loss[loss=0.2172, simple_loss=0.2958, pruned_loss=0.06929, over 18378.00 frames. ], tot_loss[loss=0.2229, simple_loss=0.3009, pruned_loss=0.07248, over 3594395.46 frames. ], batch size: 50, lr: 1.83e-02, grad_scale: 4.0 +2023-03-08 20:29:38,695 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.441e+02 4.086e+02 5.349e+02 6.220e+02 9.692e+02, threshold=1.070e+03, percent-clipped=2.0 +2023-03-08 20:29:57,005 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=20944.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:30:17,853 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6621, 3.6590, 5.0967, 4.0625, 2.5784, 2.8081, 4.0031, 5.2539], + device='cuda:0'), covar=tensor([0.0827, 0.1390, 0.0053, 0.0379, 0.1183, 0.1170, 0.0428, 0.0069], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0190, 0.0072, 0.0144, 0.0161, 0.0164, 0.0146, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 20:30:21,384 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.93 vs. limit=2.0 +2023-03-08 20:30:27,627 INFO [train.py:898] (0/4) Epoch 6, batch 2800, loss[loss=0.1984, simple_loss=0.2765, pruned_loss=0.0601, over 18267.00 frames. ], tot_loss[loss=0.221, simple_loss=0.299, pruned_loss=0.07148, over 3602924.16 frames. ], batch size: 47, lr: 1.82e-02, grad_scale: 8.0 +2023-03-08 20:31:26,858 INFO [train.py:898] (0/4) Epoch 6, batch 2850, loss[loss=0.2033, simple_loss=0.2927, pruned_loss=0.05695, over 18397.00 frames. ], tot_loss[loss=0.2206, simple_loss=0.2986, pruned_loss=0.07133, over 3593874.28 frames. ], batch size: 52, lr: 1.82e-02, grad_scale: 4.0 +2023-03-08 20:31:37,651 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.172e+02 4.192e+02 4.961e+02 6.331e+02 1.118e+03, threshold=9.922e+02, percent-clipped=2.0 +2023-03-08 20:31:54,497 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21044.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:32:03,863 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21052.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:32:24,740 INFO [train.py:898] (0/4) Epoch 6, batch 2900, loss[loss=0.2166, simple_loss=0.2853, pruned_loss=0.07395, over 18520.00 frames. ], tot_loss[loss=0.2229, simple_loss=0.3006, pruned_loss=0.07261, over 3593662.71 frames. ], batch size: 47, lr: 1.82e-02, grad_scale: 4.0 +2023-03-08 20:32:28,173 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21074.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:32:49,269 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=21092.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:33:14,688 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21113.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:33:20,448 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21118.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:33:21,572 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2963, 5.4139, 3.1289, 5.0208, 5.0600, 5.3831, 5.1640, 2.7047], + device='cuda:0'), covar=tensor([0.0117, 0.0037, 0.0613, 0.0069, 0.0056, 0.0047, 0.0076, 0.0945], + device='cuda:0'), in_proj_covar=tensor([0.0064, 0.0054, 0.0083, 0.0067, 0.0064, 0.0053, 0.0068, 0.0087], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 20:33:23,466 INFO [train.py:898] (0/4) Epoch 6, batch 2950, loss[loss=0.232, simple_loss=0.3172, pruned_loss=0.07335, over 18044.00 frames. ], tot_loss[loss=0.2231, simple_loss=0.3009, pruned_loss=0.07263, over 3594769.72 frames. ], batch size: 65, lr: 1.82e-02, grad_scale: 4.0 +2023-03-08 20:33:24,833 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=21122.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:33:33,497 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.673e+02 4.312e+02 5.615e+02 7.522e+02 2.010e+03, threshold=1.123e+03, percent-clipped=9.0 +2023-03-08 20:33:42,769 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9719, 3.7891, 5.1036, 3.2584, 4.3679, 2.7024, 2.9997, 2.2906], + device='cuda:0'), covar=tensor([0.0681, 0.0605, 0.0051, 0.0461, 0.0447, 0.1803, 0.1902, 0.1395], + device='cuda:0'), in_proj_covar=tensor([0.0166, 0.0183, 0.0088, 0.0141, 0.0192, 0.0221, 0.0216, 0.0181], + device='cuda:0'), out_proj_covar=tensor([1.5492e-04, 1.7790e-04, 8.6715e-05, 1.3634e-04, 1.8457e-04, 2.1279e-04, + 2.1267e-04, 1.7662e-04], device='cuda:0') +2023-03-08 20:34:09,321 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.3810, 3.5513, 5.2640, 4.4041, 3.0494, 2.9312, 4.5359, 5.3028], + device='cuda:0'), covar=tensor([0.0923, 0.1617, 0.0045, 0.0267, 0.0951, 0.1070, 0.0257, 0.0102], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0196, 0.0074, 0.0147, 0.0167, 0.0169, 0.0151, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 20:34:22,616 INFO [train.py:898] (0/4) Epoch 6, batch 3000, loss[loss=0.281, simple_loss=0.3435, pruned_loss=0.1093, over 12535.00 frames. ], tot_loss[loss=0.2224, simple_loss=0.2998, pruned_loss=0.07255, over 3587760.25 frames. ], batch size: 130, lr: 1.82e-02, grad_scale: 4.0 +2023-03-08 20:34:22,618 INFO [train.py:923] (0/4) Computing validation loss +2023-03-08 20:34:34,666 INFO [train.py:932] (0/4) Epoch 6, validation: loss=0.1727, simple_loss=0.276, pruned_loss=0.03476, over 944034.00 frames. +2023-03-08 20:34:34,667 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-08 20:34:45,173 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21179.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:34:53,991 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-08 20:34:55,925 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0643, 3.1008, 4.7219, 4.4169, 2.8546, 4.8528, 4.1440, 3.0759], + device='cuda:0'), covar=tensor([0.0274, 0.1032, 0.0068, 0.0161, 0.1204, 0.0126, 0.0233, 0.0809], + device='cuda:0'), in_proj_covar=tensor([0.0155, 0.0196, 0.0106, 0.0117, 0.0193, 0.0154, 0.0160, 0.0176], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0001, 0.0001, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 20:35:22,925 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0686, 2.9363, 1.6889, 3.7828, 2.5231, 3.7636, 1.8719, 3.2616], + device='cuda:0'), covar=tensor([0.0543, 0.0876, 0.1475, 0.0348, 0.0888, 0.0196, 0.1364, 0.0387], + device='cuda:0'), in_proj_covar=tensor([0.0164, 0.0194, 0.0167, 0.0180, 0.0168, 0.0159, 0.0175, 0.0166], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 20:35:33,721 INFO [train.py:898] (0/4) Epoch 6, batch 3050, loss[loss=0.2583, simple_loss=0.3279, pruned_loss=0.09432, over 12592.00 frames. ], tot_loss[loss=0.2222, simple_loss=0.2992, pruned_loss=0.07258, over 3581237.51 frames. ], batch size: 129, lr: 1.81e-02, grad_scale: 4.0 +2023-03-08 20:35:45,055 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.742e+02 3.885e+02 4.643e+02 5.808e+02 1.137e+03, threshold=9.287e+02, percent-clipped=1.0 +2023-03-08 20:36:02,291 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21244.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:36:12,619 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.22 vs. limit=5.0 +2023-03-08 20:36:32,475 INFO [train.py:898] (0/4) Epoch 6, batch 3100, loss[loss=0.2398, simple_loss=0.3129, pruned_loss=0.0834, over 16393.00 frames. ], tot_loss[loss=0.2217, simple_loss=0.2988, pruned_loss=0.07227, over 3579907.17 frames. ], batch size: 94, lr: 1.81e-02, grad_scale: 4.0 +2023-03-08 20:36:44,099 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-03-08 20:36:58,420 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=21292.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:37:21,431 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21312.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:37:31,448 INFO [train.py:898] (0/4) Epoch 6, batch 3150, loss[loss=0.2148, simple_loss=0.2929, pruned_loss=0.06831, over 18483.00 frames. ], tot_loss[loss=0.2211, simple_loss=0.2983, pruned_loss=0.07188, over 3583654.70 frames. ], batch size: 51, lr: 1.81e-02, grad_scale: 4.0 +2023-03-08 20:37:39,512 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21328.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:37:41,465 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.840e+02 4.055e+02 4.775e+02 6.175e+02 1.308e+03, threshold=9.551e+02, percent-clipped=5.0 +2023-03-08 20:38:29,900 INFO [train.py:898] (0/4) Epoch 6, batch 3200, loss[loss=0.1975, simple_loss=0.2828, pruned_loss=0.05606, over 18392.00 frames. ], tot_loss[loss=0.2209, simple_loss=0.2983, pruned_loss=0.07172, over 3578505.06 frames. ], batch size: 52, lr: 1.81e-02, grad_scale: 8.0 +2023-03-08 20:38:32,545 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21373.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:38:52,005 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21389.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:39:14,432 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21408.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:39:28,806 INFO [train.py:898] (0/4) Epoch 6, batch 3250, loss[loss=0.2287, simple_loss=0.3103, pruned_loss=0.07358, over 18407.00 frames. ], tot_loss[loss=0.2223, simple_loss=0.2999, pruned_loss=0.07235, over 3582106.13 frames. ], batch size: 52, lr: 1.81e-02, grad_scale: 8.0 +2023-03-08 20:39:39,004 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.706e+02 4.124e+02 5.141e+02 6.520e+02 1.245e+03, threshold=1.028e+03, percent-clipped=2.0 +2023-03-08 20:40:03,113 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-03-08 20:40:28,083 INFO [train.py:898] (0/4) Epoch 6, batch 3300, loss[loss=0.2244, simple_loss=0.3058, pruned_loss=0.07146, over 18267.00 frames. ], tot_loss[loss=0.2228, simple_loss=0.3003, pruned_loss=0.07264, over 3569401.22 frames. ], batch size: 49, lr: 1.80e-02, grad_scale: 8.0 +2023-03-08 20:40:31,928 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21474.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:40:57,953 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6178, 5.6204, 5.2777, 5.4972, 4.6377, 5.4300, 5.7003, 5.5691], + device='cuda:0'), covar=tensor([0.2659, 0.1010, 0.0824, 0.1182, 0.3022, 0.1120, 0.1026, 0.0983], + device='cuda:0'), in_proj_covar=tensor([0.0444, 0.0360, 0.0280, 0.0394, 0.0538, 0.0393, 0.0477, 0.0364], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-08 20:41:27,331 INFO [train.py:898] (0/4) Epoch 6, batch 3350, loss[loss=0.2288, simple_loss=0.3018, pruned_loss=0.07793, over 18090.00 frames. ], tot_loss[loss=0.2223, simple_loss=0.2997, pruned_loss=0.07241, over 3564766.95 frames. ], batch size: 62, lr: 1.80e-02, grad_scale: 8.0 +2023-03-08 20:41:37,406 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.459e+02 4.317e+02 5.194e+02 7.020e+02 1.247e+03, threshold=1.039e+03, percent-clipped=7.0 +2023-03-08 20:42:25,863 INFO [train.py:898] (0/4) Epoch 6, batch 3400, loss[loss=0.2177, simple_loss=0.2843, pruned_loss=0.07551, over 18465.00 frames. ], tot_loss[loss=0.2221, simple_loss=0.2999, pruned_loss=0.07217, over 3580220.69 frames. ], batch size: 44, lr: 1.80e-02, grad_scale: 4.0 +2023-03-08 20:42:53,846 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21595.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:43:15,630 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21613.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:43:24,529 INFO [train.py:898] (0/4) Epoch 6, batch 3450, loss[loss=0.2197, simple_loss=0.3013, pruned_loss=0.069, over 18539.00 frames. ], tot_loss[loss=0.2212, simple_loss=0.2992, pruned_loss=0.07155, over 3587020.24 frames. ], batch size: 49, lr: 1.80e-02, grad_scale: 4.0 +2023-03-08 20:43:26,093 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8627, 3.5963, 2.1211, 4.4663, 3.1868, 4.7177, 2.7099, 3.9627], + device='cuda:0'), covar=tensor([0.0389, 0.0733, 0.1443, 0.0292, 0.0839, 0.0158, 0.1040, 0.0362], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0198, 0.0170, 0.0185, 0.0172, 0.0165, 0.0177, 0.0168], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 20:43:35,795 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.934e+02 4.048e+02 5.156e+02 6.271e+02 2.369e+03, threshold=1.031e+03, percent-clipped=5.0 +2023-03-08 20:44:05,761 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21656.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:44:20,554 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21668.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:44:23,903 INFO [train.py:898] (0/4) Epoch 6, batch 3500, loss[loss=0.1844, simple_loss=0.2644, pruned_loss=0.05217, over 18397.00 frames. ], tot_loss[loss=0.2208, simple_loss=0.299, pruned_loss=0.07128, over 3579208.30 frames. ], batch size: 48, lr: 1.80e-02, grad_scale: 4.0 +2023-03-08 20:44:27,463 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21674.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:44:30,879 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9486, 4.1150, 5.2011, 3.2791, 4.2917, 3.0134, 3.0252, 2.0989], + device='cuda:0'), covar=tensor([0.0706, 0.0476, 0.0049, 0.0438, 0.0505, 0.1540, 0.1886, 0.1508], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0186, 0.0090, 0.0143, 0.0199, 0.0228, 0.0222, 0.0186], + device='cuda:0'), out_proj_covar=tensor([1.5834e-04, 1.7831e-04, 8.9136e-05, 1.3626e-04, 1.9086e-04, 2.1855e-04, + 2.1684e-04, 1.8038e-04], device='cuda:0') +2023-03-08 20:44:38,483 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21684.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:45:04,808 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21708.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:45:18,375 INFO [train.py:898] (0/4) Epoch 6, batch 3550, loss[loss=0.2038, simple_loss=0.2932, pruned_loss=0.05716, over 18237.00 frames. ], tot_loss[loss=0.2207, simple_loss=0.299, pruned_loss=0.07119, over 3583491.35 frames. ], batch size: 60, lr: 1.79e-02, grad_scale: 4.0 +2023-03-08 20:45:25,039 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-08 20:45:28,808 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.680e+02 4.052e+02 4.672e+02 6.032e+02 1.745e+03, threshold=9.344e+02, percent-clipped=2.0 +2023-03-08 20:45:55,820 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6327, 5.2245, 5.3059, 5.2409, 4.8746, 5.1319, 4.4585, 5.1524], + device='cuda:0'), covar=tensor([0.0238, 0.0252, 0.0160, 0.0236, 0.0309, 0.0208, 0.1068, 0.0236], + device='cuda:0'), in_proj_covar=tensor([0.0140, 0.0182, 0.0167, 0.0173, 0.0172, 0.0183, 0.0251, 0.0165], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0006, 0.0005], + device='cuda:0') +2023-03-08 20:45:56,742 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=21756.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:45:57,066 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6270, 1.8099, 2.7643, 2.5591, 3.4514, 4.8165, 4.2479, 4.0254], + device='cuda:0'), covar=tensor([0.0615, 0.1353, 0.1285, 0.0910, 0.1150, 0.0047, 0.0282, 0.0232], + device='cuda:0'), in_proj_covar=tensor([0.0181, 0.0234, 0.0222, 0.0215, 0.0314, 0.0134, 0.0206, 0.0162], + device='cuda:0'), out_proj_covar=tensor([1.2485e-04, 1.6260e-04, 1.6277e-04, 1.3813e-04, 2.1939e-04, 8.9934e-05, + 1.3905e-04, 1.1275e-04], device='cuda:0') +2023-03-08 20:46:04,482 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21763.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:46:12,830 INFO [train.py:898] (0/4) Epoch 6, batch 3600, loss[loss=0.2025, simple_loss=0.2877, pruned_loss=0.05866, over 18554.00 frames. ], tot_loss[loss=0.2207, simple_loss=0.2991, pruned_loss=0.0711, over 3589666.93 frames. ], batch size: 54, lr: 1.79e-02, grad_scale: 8.0 +2023-03-08 20:46:16,480 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21774.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:46:30,655 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-08 20:46:49,331 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-6.pt +2023-03-08 20:47:18,366 INFO [train.py:898] (0/4) Epoch 7, batch 0, loss[loss=0.1931, simple_loss=0.2684, pruned_loss=0.05893, over 18162.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2684, pruned_loss=0.05893, over 18162.00 frames. ], batch size: 44, lr: 1.68e-02, grad_scale: 8.0 +2023-03-08 20:47:18,368 INFO [train.py:923] (0/4) Computing validation loss +2023-03-08 20:47:30,266 INFO [train.py:932] (0/4) Epoch 7, validation: loss=0.175, simple_loss=0.2779, pruned_loss=0.0361, over 944034.00 frames. +2023-03-08 20:47:30,267 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-08 20:47:50,700 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=21822.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:47:53,165 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21824.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:48:01,604 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.548e+02 4.151e+02 4.797e+02 6.024e+02 1.150e+03, threshold=9.595e+02, percent-clipped=4.0 +2023-03-08 20:48:10,515 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-08 20:48:23,573 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21850.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:48:28,870 INFO [train.py:898] (0/4) Epoch 7, batch 50, loss[loss=0.2143, simple_loss=0.3007, pruned_loss=0.06394, over 18514.00 frames. ], tot_loss[loss=0.2159, simple_loss=0.296, pruned_loss=0.06786, over 823287.27 frames. ], batch size: 53, lr: 1.68e-02, grad_scale: 8.0 +2023-03-08 20:49:05,851 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.62 vs. limit=2.0 +2023-03-08 20:49:13,614 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0357, 4.6100, 4.8003, 3.4643, 3.7497, 3.6376, 2.3260, 2.1639], + device='cuda:0'), covar=tensor([0.0153, 0.0149, 0.0037, 0.0238, 0.0342, 0.0202, 0.0788, 0.0956], + device='cuda:0'), in_proj_covar=tensor([0.0045, 0.0042, 0.0037, 0.0051, 0.0072, 0.0047, 0.0068, 0.0073], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0005, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 20:49:27,610 INFO [train.py:898] (0/4) Epoch 7, batch 100, loss[loss=0.2286, simple_loss=0.3148, pruned_loss=0.0712, over 18233.00 frames. ], tot_loss[loss=0.2153, simple_loss=0.2948, pruned_loss=0.06792, over 1442128.38 frames. ], batch size: 60, lr: 1.67e-02, grad_scale: 8.0 +2023-03-08 20:49:34,981 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6113, 3.4801, 1.8060, 4.4374, 2.9300, 4.5890, 2.3442, 3.7718], + device='cuda:0'), covar=tensor([0.0434, 0.0774, 0.1408, 0.0334, 0.0777, 0.0176, 0.1033, 0.0395], + device='cuda:0'), in_proj_covar=tensor([0.0163, 0.0193, 0.0167, 0.0185, 0.0169, 0.0164, 0.0176, 0.0165], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 20:49:35,004 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=21911.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:49:58,913 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.330e+02 3.869e+02 4.841e+02 5.741e+02 1.343e+03, threshold=9.682e+02, percent-clipped=1.0 +2023-03-08 20:50:12,637 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.51 vs. limit=2.0 +2023-03-08 20:50:22,189 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21951.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:50:26,552 INFO [train.py:898] (0/4) Epoch 7, batch 150, loss[loss=0.2167, simple_loss=0.3062, pruned_loss=0.06366, over 18342.00 frames. ], tot_loss[loss=0.2165, simple_loss=0.2958, pruned_loss=0.06856, over 1908774.83 frames. ], batch size: 55, lr: 1.67e-02, grad_scale: 8.0 +2023-03-08 20:50:41,094 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21968.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:50:42,098 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=21969.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:50:43,444 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=21970.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:51:01,846 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=21984.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:51:08,212 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.83 vs. limit=5.0 +2023-03-08 20:51:20,408 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-22000.pt +2023-03-08 20:51:30,385 INFO [train.py:898] (0/4) Epoch 7, batch 200, loss[loss=0.2423, simple_loss=0.3158, pruned_loss=0.08442, over 18355.00 frames. ], tot_loss[loss=0.2167, simple_loss=0.2953, pruned_loss=0.06904, over 2268999.54 frames. ], batch size: 56, lr: 1.67e-02, grad_scale: 8.0 +2023-03-08 20:51:31,078 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.04 vs. limit=5.0 +2023-03-08 20:51:39,783 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6273, 4.5965, 4.6691, 4.4034, 4.5237, 4.5422, 4.9451, 4.8354], + device='cuda:0'), covar=tensor([0.0071, 0.0075, 0.0081, 0.0115, 0.0070, 0.0104, 0.0109, 0.0107], + device='cuda:0'), in_proj_covar=tensor([0.0068, 0.0050, 0.0050, 0.0065, 0.0054, 0.0076, 0.0064, 0.0063], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 20:51:43,120 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=22016.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:51:59,830 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.533e+02 3.799e+02 4.724e+02 5.611e+02 1.174e+03, threshold=9.448e+02, percent-clipped=1.0 +2023-03-08 20:52:00,222 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22031.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:52:01,877 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=22032.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:52:28,976 INFO [train.py:898] (0/4) Epoch 7, batch 250, loss[loss=0.2341, simple_loss=0.3097, pruned_loss=0.07932, over 16190.00 frames. ], tot_loss[loss=0.2169, simple_loss=0.2954, pruned_loss=0.06918, over 2556183.89 frames. ], batch size: 94, lr: 1.67e-02, grad_scale: 8.0 +2023-03-08 20:53:11,469 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8627, 4.2270, 4.4394, 3.3789, 3.5587, 3.4152, 2.4780, 2.0133], + device='cuda:0'), covar=tensor([0.0176, 0.0214, 0.0086, 0.0243, 0.0333, 0.0189, 0.0817, 0.1069], + device='cuda:0'), in_proj_covar=tensor([0.0046, 0.0043, 0.0037, 0.0052, 0.0072, 0.0047, 0.0069, 0.0075], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0005, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 20:53:23,958 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4371, 2.5937, 3.9560, 3.9190, 2.7384, 4.2859, 3.6932, 2.6306], + device='cuda:0'), covar=tensor([0.0356, 0.1241, 0.0166, 0.0159, 0.1249, 0.0118, 0.0313, 0.1053], + device='cuda:0'), in_proj_covar=tensor([0.0157, 0.0205, 0.0113, 0.0119, 0.0197, 0.0153, 0.0164, 0.0183], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 20:53:28,160 INFO [train.py:898] (0/4) Epoch 7, batch 300, loss[loss=0.1949, simple_loss=0.2599, pruned_loss=0.06495, over 18397.00 frames. ], tot_loss[loss=0.2169, simple_loss=0.2959, pruned_loss=0.06894, over 2781946.71 frames. ], batch size: 42, lr: 1.67e-02, grad_scale: 8.0 +2023-03-08 20:53:29,649 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5521, 3.2539, 3.3174, 2.8767, 3.1352, 2.5145, 2.5879, 3.4907], + device='cuda:0'), covar=tensor([0.0029, 0.0067, 0.0057, 0.0098, 0.0069, 0.0149, 0.0157, 0.0046], + device='cuda:0'), in_proj_covar=tensor([0.0063, 0.0084, 0.0077, 0.0121, 0.0079, 0.0120, 0.0127, 0.0069], + device='cuda:0'), out_proj_covar=tensor([8.9123e-05, 1.2929e-04, 1.1731e-04, 1.9143e-04, 1.2023e-04, 1.8691e-04, + 1.9778e-04, 1.0085e-04], device='cuda:0') +2023-03-08 20:53:29,879 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-08 20:53:44,076 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22119.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:53:56,284 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22130.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:53:56,970 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.650e+02 3.954e+02 4.582e+02 6.003e+02 1.434e+03, threshold=9.164e+02, percent-clipped=5.0 +2023-03-08 20:54:04,179 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-08 20:54:26,849 INFO [train.py:898] (0/4) Epoch 7, batch 350, loss[loss=0.2115, simple_loss=0.3005, pruned_loss=0.06123, over 18570.00 frames. ], tot_loss[loss=0.2164, simple_loss=0.2957, pruned_loss=0.06849, over 2970872.43 frames. ], batch size: 54, lr: 1.67e-02, grad_scale: 8.0 +2023-03-08 20:54:49,532 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5277, 5.5168, 4.9821, 5.5107, 5.4828, 4.8495, 5.3945, 5.1206], + device='cuda:0'), covar=tensor([0.0339, 0.0258, 0.1465, 0.0601, 0.0379, 0.0380, 0.0336, 0.0796], + device='cuda:0'), in_proj_covar=tensor([0.0334, 0.0383, 0.0533, 0.0308, 0.0280, 0.0359, 0.0371, 0.0477], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-08 20:54:50,758 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22176.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:55:08,583 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22191.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:55:25,953 INFO [train.py:898] (0/4) Epoch 7, batch 400, loss[loss=0.2401, simple_loss=0.3152, pruned_loss=0.0825, over 16081.00 frames. ], tot_loss[loss=0.2183, simple_loss=0.297, pruned_loss=0.06976, over 3105050.13 frames. ], batch size: 95, lr: 1.66e-02, grad_scale: 8.0 +2023-03-08 20:55:27,290 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22206.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:55:37,473 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22215.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:55:45,758 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-08 20:55:55,646 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.341e+02 3.724e+02 4.646e+02 6.354e+02 1.977e+03, threshold=9.292e+02, percent-clipped=7.0 +2023-03-08 20:56:02,872 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22237.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:56:02,980 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22237.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:56:20,920 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22251.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:56:25,013 INFO [train.py:898] (0/4) Epoch 7, batch 450, loss[loss=0.2307, simple_loss=0.3108, pruned_loss=0.07536, over 18633.00 frames. ], tot_loss[loss=0.218, simple_loss=0.2967, pruned_loss=0.06961, over 3206126.59 frames. ], batch size: 52, lr: 1.66e-02, grad_scale: 8.0 +2023-03-08 20:56:40,106 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2824, 5.0113, 5.2685, 5.3012, 5.0752, 5.8785, 5.5048, 5.2313], + device='cuda:0'), covar=tensor([0.0788, 0.0629, 0.0604, 0.0591, 0.1364, 0.0724, 0.0591, 0.1444], + device='cuda:0'), in_proj_covar=tensor([0.0265, 0.0204, 0.0209, 0.0205, 0.0248, 0.0293, 0.0196, 0.0293], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 20:56:41,317 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22269.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:56:49,424 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22276.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:57:15,005 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22298.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:57:15,862 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=22299.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:57:23,805 INFO [train.py:898] (0/4) Epoch 7, batch 500, loss[loss=0.2111, simple_loss=0.2848, pruned_loss=0.06873, over 18280.00 frames. ], tot_loss[loss=0.2179, simple_loss=0.2965, pruned_loss=0.06967, over 3287184.12 frames. ], batch size: 49, lr: 1.66e-02, grad_scale: 8.0 +2023-03-08 20:57:37,414 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=22317.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:57:47,842 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22326.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:57:53,239 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.810e+02 3.764e+02 4.516e+02 5.475e+02 8.626e+02, threshold=9.031e+02, percent-clipped=0.0 +2023-03-08 20:58:23,007 INFO [train.py:898] (0/4) Epoch 7, batch 550, loss[loss=0.2305, simple_loss=0.3168, pruned_loss=0.07209, over 18448.00 frames. ], tot_loss[loss=0.2178, simple_loss=0.2965, pruned_loss=0.06952, over 3349767.88 frames. ], batch size: 59, lr: 1.66e-02, grad_scale: 8.0 +2023-03-08 20:58:36,137 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.79 vs. limit=2.0 +2023-03-08 20:58:55,265 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4742, 5.4670, 4.9781, 5.4161, 5.4447, 4.7708, 5.3766, 5.0545], + device='cuda:0'), covar=tensor([0.0380, 0.0315, 0.1300, 0.0676, 0.0470, 0.0402, 0.0327, 0.0766], + device='cuda:0'), in_proj_covar=tensor([0.0340, 0.0381, 0.0530, 0.0308, 0.0279, 0.0361, 0.0374, 0.0485], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-08 20:58:57,660 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22385.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:59:16,561 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-08 20:59:21,869 INFO [train.py:898] (0/4) Epoch 7, batch 600, loss[loss=0.206, simple_loss=0.2863, pruned_loss=0.06284, over 18493.00 frames. ], tot_loss[loss=0.2187, simple_loss=0.2971, pruned_loss=0.07013, over 3392523.59 frames. ], batch size: 51, lr: 1.66e-02, grad_scale: 8.0 +2023-03-08 20:59:38,725 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22419.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 20:59:51,503 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-08 20:59:52,060 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.568e+02 3.797e+02 4.706e+02 5.910e+02 1.335e+03, threshold=9.411e+02, percent-clipped=3.0 +2023-03-08 21:00:04,809 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1716, 5.0415, 5.3093, 5.1647, 5.0508, 5.8648, 5.3594, 5.3312], + device='cuda:0'), covar=tensor([0.0859, 0.0676, 0.0615, 0.0610, 0.1362, 0.0660, 0.0582, 0.1632], + device='cuda:0'), in_proj_covar=tensor([0.0266, 0.0199, 0.0209, 0.0205, 0.0245, 0.0299, 0.0194, 0.0292], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 21:00:09,532 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22446.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:00:19,240 INFO [train.py:898] (0/4) Epoch 7, batch 650, loss[loss=0.1882, simple_loss=0.2576, pruned_loss=0.05938, over 18417.00 frames. ], tot_loss[loss=0.2185, simple_loss=0.2969, pruned_loss=0.07009, over 3429040.25 frames. ], batch size: 43, lr: 1.65e-02, grad_scale: 8.0 +2023-03-08 21:00:35,126 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=22467.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:00:56,992 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22486.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:01:00,358 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3056, 5.3305, 4.8118, 5.3205, 5.3202, 4.7103, 5.2157, 4.8614], + device='cuda:0'), covar=tensor([0.0373, 0.0361, 0.1334, 0.0587, 0.0457, 0.0371, 0.0357, 0.0938], + device='cuda:0'), in_proj_covar=tensor([0.0338, 0.0380, 0.0530, 0.0304, 0.0276, 0.0359, 0.0373, 0.0481], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-08 21:01:18,254 INFO [train.py:898] (0/4) Epoch 7, batch 700, loss[loss=0.2173, simple_loss=0.2998, pruned_loss=0.06743, over 18025.00 frames. ], tot_loss[loss=0.2192, simple_loss=0.2974, pruned_loss=0.07047, over 3462303.10 frames. ], batch size: 65, lr: 1.65e-02, grad_scale: 8.0 +2023-03-08 21:01:19,551 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9474, 5.6070, 5.1630, 5.2857, 5.0485, 5.0670, 5.6705, 5.6452], + device='cuda:0'), covar=tensor([0.1249, 0.0630, 0.0678, 0.0643, 0.1509, 0.0656, 0.0538, 0.0517], + device='cuda:0'), in_proj_covar=tensor([0.0447, 0.0367, 0.0285, 0.0399, 0.0544, 0.0398, 0.0485, 0.0370], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-08 21:01:19,589 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22506.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:01:49,203 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.755e+02 3.999e+02 4.789e+02 5.664e+02 1.125e+03, threshold=9.578e+02, percent-clipped=2.0 +2023-03-08 21:01:50,630 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22532.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:02:14,545 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.4319, 4.6109, 4.6076, 4.3434, 4.4111, 4.3437, 4.7413, 4.6658], + device='cuda:0'), covar=tensor([0.0073, 0.0059, 0.0064, 0.0095, 0.0064, 0.0100, 0.0072, 0.0078], + device='cuda:0'), in_proj_covar=tensor([0.0067, 0.0050, 0.0050, 0.0065, 0.0055, 0.0074, 0.0063, 0.0063], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 21:02:15,568 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=22554.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:02:16,502 INFO [train.py:898] (0/4) Epoch 7, batch 750, loss[loss=0.2331, simple_loss=0.3106, pruned_loss=0.07786, over 18093.00 frames. ], tot_loss[loss=0.2201, simple_loss=0.2981, pruned_loss=0.07112, over 3479492.64 frames. ], batch size: 62, lr: 1.65e-02, grad_scale: 8.0 +2023-03-08 21:02:37,339 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22571.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:03:01,765 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22593.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:03:15,339 INFO [train.py:898] (0/4) Epoch 7, batch 800, loss[loss=0.2739, simple_loss=0.3309, pruned_loss=0.1084, over 11891.00 frames. ], tot_loss[loss=0.2198, simple_loss=0.2979, pruned_loss=0.07089, over 3500650.22 frames. ], batch size: 130, lr: 1.65e-02, grad_scale: 8.0 +2023-03-08 21:03:41,306 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22626.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:03:46,429 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.314e+02 4.023e+02 4.944e+02 6.162e+02 1.524e+03, threshold=9.887e+02, percent-clipped=2.0 +2023-03-08 21:03:55,290 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22639.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:04:04,530 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8757, 4.5236, 4.6771, 3.2156, 3.5243, 3.7155, 2.4170, 2.0762], + device='cuda:0'), covar=tensor([0.0144, 0.0118, 0.0070, 0.0294, 0.0331, 0.0141, 0.0766, 0.0954], + device='cuda:0'), in_proj_covar=tensor([0.0045, 0.0041, 0.0038, 0.0051, 0.0071, 0.0046, 0.0068, 0.0073], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0005, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 21:04:13,325 INFO [train.py:898] (0/4) Epoch 7, batch 850, loss[loss=0.2576, simple_loss=0.3233, pruned_loss=0.09596, over 18471.00 frames. ], tot_loss[loss=0.2192, simple_loss=0.2973, pruned_loss=0.0705, over 3515391.66 frames. ], batch size: 59, lr: 1.65e-02, grad_scale: 8.0 +2023-03-08 21:04:31,953 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.4433, 1.7502, 2.8175, 2.7873, 3.4549, 4.9558, 4.3378, 4.3158], + device='cuda:0'), covar=tensor([0.0738, 0.1508, 0.1508, 0.0935, 0.1318, 0.0046, 0.0308, 0.0208], + device='cuda:0'), in_proj_covar=tensor([0.0188, 0.0242, 0.0235, 0.0218, 0.0323, 0.0138, 0.0214, 0.0168], + device='cuda:0'), out_proj_covar=tensor([1.2831e-04, 1.6657e-04, 1.6941e-04, 1.3888e-04, 2.2249e-04, 9.1274e-05, + 1.4268e-04, 1.1548e-04], device='cuda:0') +2023-03-08 21:04:36,640 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=22674.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:05:07,166 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22700.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 21:05:12,417 INFO [train.py:898] (0/4) Epoch 7, batch 900, loss[loss=0.215, simple_loss=0.3019, pruned_loss=0.06402, over 17835.00 frames. ], tot_loss[loss=0.2174, simple_loss=0.2961, pruned_loss=0.06935, over 3544997.51 frames. ], batch size: 70, lr: 1.65e-02, grad_scale: 8.0 +2023-03-08 21:05:17,251 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.4058, 5.5657, 2.8088, 5.2534, 5.2643, 5.6159, 5.3661, 2.6617], + device='cuda:0'), covar=tensor([0.0122, 0.0039, 0.0673, 0.0051, 0.0058, 0.0032, 0.0071, 0.0914], + device='cuda:0'), in_proj_covar=tensor([0.0067, 0.0055, 0.0085, 0.0070, 0.0066, 0.0055, 0.0069, 0.0090], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 21:05:19,417 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3645, 6.0763, 5.5022, 5.7586, 5.4755, 5.5644, 6.0745, 6.0202], + device='cuda:0'), covar=tensor([0.1162, 0.0510, 0.0425, 0.0667, 0.1409, 0.0592, 0.0452, 0.0502], + device='cuda:0'), in_proj_covar=tensor([0.0444, 0.0369, 0.0283, 0.0401, 0.0540, 0.0398, 0.0481, 0.0371], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-08 21:05:21,770 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2973, 5.2604, 4.7307, 5.2175, 5.2191, 4.5337, 5.1605, 4.7909], + device='cuda:0'), covar=tensor([0.0414, 0.0428, 0.1534, 0.0773, 0.0431, 0.0472, 0.0399, 0.0905], + device='cuda:0'), in_proj_covar=tensor([0.0344, 0.0387, 0.0538, 0.0314, 0.0280, 0.0367, 0.0381, 0.0488], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-08 21:05:44,109 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.332e+02 3.928e+02 4.679e+02 5.681e+02 1.388e+03, threshold=9.358e+02, percent-clipped=3.0 +2023-03-08 21:05:55,293 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22741.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:06:05,003 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.66 vs. limit=5.0 +2023-03-08 21:06:09,497 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.00 vs. limit=2.0 +2023-03-08 21:06:11,231 INFO [train.py:898] (0/4) Epoch 7, batch 950, loss[loss=0.2291, simple_loss=0.3084, pruned_loss=0.07492, over 18479.00 frames. ], tot_loss[loss=0.2166, simple_loss=0.2956, pruned_loss=0.06885, over 3563131.89 frames. ], batch size: 51, lr: 1.64e-02, grad_scale: 8.0 +2023-03-08 21:06:49,144 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22786.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:06:58,142 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22794.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:07:10,655 INFO [train.py:898] (0/4) Epoch 7, batch 1000, loss[loss=0.2287, simple_loss=0.3119, pruned_loss=0.07272, over 18042.00 frames. ], tot_loss[loss=0.217, simple_loss=0.296, pruned_loss=0.06899, over 3573826.38 frames. ], batch size: 65, lr: 1.64e-02, grad_scale: 8.0 +2023-03-08 21:07:33,517 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=22825.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:07:41,080 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.326e+02 3.793e+02 5.006e+02 6.046e+02 1.509e+03, threshold=1.001e+03, percent-clipped=4.0 +2023-03-08 21:07:43,035 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22832.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:07:45,159 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=22834.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:07:50,684 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.2657, 3.8180, 5.0629, 3.8582, 3.2866, 2.9102, 4.2695, 5.1389], + device='cuda:0'), covar=tensor([0.0835, 0.1107, 0.0049, 0.0303, 0.0687, 0.0940, 0.0293, 0.0080], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0203, 0.0076, 0.0143, 0.0164, 0.0168, 0.0151, 0.0104], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 21:07:59,892 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2554, 5.3055, 2.9013, 5.0823, 5.0015, 5.2898, 4.9955, 2.4011], + device='cuda:0'), covar=tensor([0.0132, 0.0052, 0.0698, 0.0073, 0.0065, 0.0066, 0.0102, 0.1165], + device='cuda:0'), in_proj_covar=tensor([0.0068, 0.0055, 0.0085, 0.0071, 0.0067, 0.0056, 0.0070, 0.0090], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 21:08:09,923 INFO [train.py:898] (0/4) Epoch 7, batch 1050, loss[loss=0.175, simple_loss=0.259, pruned_loss=0.04551, over 18412.00 frames. ], tot_loss[loss=0.2148, simple_loss=0.2943, pruned_loss=0.06765, over 3583027.81 frames. ], batch size: 48, lr: 1.64e-02, grad_scale: 8.0 +2023-03-08 21:08:10,376 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22855.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:08:28,500 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22871.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:08:38,404 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=22880.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:08:47,206 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=22886.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:08:55,752 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=22893.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:09:09,172 INFO [train.py:898] (0/4) Epoch 7, batch 1100, loss[loss=0.1986, simple_loss=0.2744, pruned_loss=0.06141, over 17684.00 frames. ], tot_loss[loss=0.2145, simple_loss=0.2943, pruned_loss=0.06731, over 3591014.36 frames. ], batch size: 39, lr: 1.64e-02, grad_scale: 8.0 +2023-03-08 21:09:25,017 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=22919.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:09:38,077 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.302e+02 3.634e+02 4.460e+02 5.357e+02 1.645e+03, threshold=8.921e+02, percent-clipped=3.0 +2023-03-08 21:09:48,559 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-08 21:09:51,087 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=22941.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:09:51,607 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-08 21:10:07,845 INFO [train.py:898] (0/4) Epoch 7, batch 1150, loss[loss=0.1863, simple_loss=0.2632, pruned_loss=0.05469, over 18440.00 frames. ], tot_loss[loss=0.2148, simple_loss=0.2944, pruned_loss=0.06758, over 3598171.54 frames. ], batch size: 43, lr: 1.64e-02, grad_scale: 8.0 +2023-03-08 21:10:55,034 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=22995.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 21:11:06,251 INFO [train.py:898] (0/4) Epoch 7, batch 1200, loss[loss=0.2353, simple_loss=0.3156, pruned_loss=0.07753, over 16332.00 frames. ], tot_loss[loss=0.216, simple_loss=0.2955, pruned_loss=0.06824, over 3592131.86 frames. ], batch size: 94, lr: 1.64e-02, grad_scale: 8.0 +2023-03-08 21:11:35,898 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.453e+02 4.180e+02 4.856e+02 6.104e+02 1.411e+03, threshold=9.713e+02, percent-clipped=4.0 +2023-03-08 21:11:49,407 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23041.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:12:05,443 INFO [train.py:898] (0/4) Epoch 7, batch 1250, loss[loss=0.2365, simple_loss=0.3105, pruned_loss=0.08128, over 17794.00 frames. ], tot_loss[loss=0.2151, simple_loss=0.2947, pruned_loss=0.06772, over 3602773.04 frames. ], batch size: 70, lr: 1.63e-02, grad_scale: 8.0 +2023-03-08 21:12:44,785 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=23089.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:13:04,079 INFO [train.py:898] (0/4) Epoch 7, batch 1300, loss[loss=0.1896, simple_loss=0.2786, pruned_loss=0.05028, over 18501.00 frames. ], tot_loss[loss=0.2146, simple_loss=0.2941, pruned_loss=0.06759, over 3604011.15 frames. ], batch size: 53, lr: 1.63e-02, grad_scale: 8.0 +2023-03-08 21:13:07,934 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23108.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:13:31,033 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-08 21:13:33,527 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.650e+02 3.767e+02 4.521e+02 6.073e+02 1.537e+03, threshold=9.042e+02, percent-clipped=6.0 +2023-03-08 21:13:41,232 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-08 21:13:57,457 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23150.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:14:03,016 INFO [train.py:898] (0/4) Epoch 7, batch 1350, loss[loss=0.2075, simple_loss=0.2818, pruned_loss=0.06662, over 18247.00 frames. ], tot_loss[loss=0.2147, simple_loss=0.2941, pruned_loss=0.06762, over 3602993.53 frames. ], batch size: 47, lr: 1.63e-02, grad_scale: 8.0 +2023-03-08 21:14:18,443 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.04 vs. limit=2.0 +2023-03-08 21:14:19,232 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23169.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:14:26,480 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-08 21:14:32,479 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23181.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:15:02,204 INFO [train.py:898] (0/4) Epoch 7, batch 1400, loss[loss=0.2174, simple_loss=0.3047, pruned_loss=0.06504, over 17696.00 frames. ], tot_loss[loss=0.2152, simple_loss=0.2945, pruned_loss=0.06792, over 3584758.75 frames. ], batch size: 70, lr: 1.63e-02, grad_scale: 8.0 +2023-03-08 21:15:15,916 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0114, 4.9253, 5.1463, 5.0253, 4.9654, 5.6965, 5.3412, 5.2304], + device='cuda:0'), covar=tensor([0.0826, 0.0637, 0.0654, 0.0565, 0.1162, 0.0649, 0.0521, 0.1465], + device='cuda:0'), in_proj_covar=tensor([0.0271, 0.0200, 0.0209, 0.0202, 0.0243, 0.0297, 0.0196, 0.0291], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 21:15:16,080 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23217.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:15:29,734 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23229.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:15:31,618 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.272e+02 3.620e+02 4.462e+02 5.791e+02 9.661e+02, threshold=8.924e+02, percent-clipped=4.0 +2023-03-08 21:15:53,170 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9749, 4.9979, 5.0364, 4.8345, 4.8417, 4.8476, 5.1927, 5.2509], + device='cuda:0'), covar=tensor([0.0050, 0.0053, 0.0051, 0.0064, 0.0053, 0.0080, 0.0102, 0.0082], + device='cuda:0'), in_proj_covar=tensor([0.0066, 0.0049, 0.0049, 0.0064, 0.0053, 0.0073, 0.0062, 0.0061], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 21:16:00,105 INFO [train.py:898] (0/4) Epoch 7, batch 1450, loss[loss=0.2624, simple_loss=0.3359, pruned_loss=0.09448, over 16404.00 frames. ], tot_loss[loss=0.2159, simple_loss=0.2949, pruned_loss=0.06839, over 3581463.85 frames. ], batch size: 94, lr: 1.63e-02, grad_scale: 8.0 +2023-03-08 21:16:03,580 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-08 21:16:12,889 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-08 21:16:26,940 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23278.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:16:37,084 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9191, 4.8861, 4.9467, 4.6447, 4.6188, 4.6602, 5.0936, 5.1138], + device='cuda:0'), covar=tensor([0.0059, 0.0061, 0.0055, 0.0077, 0.0065, 0.0087, 0.0060, 0.0090], + device='cuda:0'), in_proj_covar=tensor([0.0067, 0.0049, 0.0050, 0.0064, 0.0054, 0.0074, 0.0063, 0.0062], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 21:16:39,742 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.28 vs. limit=5.0 +2023-03-08 21:16:40,547 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23290.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:16:46,258 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23295.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:16:57,597 INFO [train.py:898] (0/4) Epoch 7, batch 1500, loss[loss=0.17, simple_loss=0.2467, pruned_loss=0.04668, over 18059.00 frames. ], tot_loss[loss=0.2165, simple_loss=0.2956, pruned_loss=0.06865, over 3587591.30 frames. ], batch size: 40, lr: 1.63e-02, grad_scale: 8.0 +2023-03-08 21:16:58,342 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-03-08 21:17:01,226 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-08 21:17:27,942 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.706e+02 4.019e+02 4.944e+02 6.041e+02 1.007e+03, threshold=9.887e+02, percent-clipped=2.0 +2023-03-08 21:17:41,547 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=23343.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:17:55,132 INFO [train.py:898] (0/4) Epoch 7, batch 1550, loss[loss=0.203, simple_loss=0.2887, pruned_loss=0.05862, over 18392.00 frames. ], tot_loss[loss=0.2153, simple_loss=0.2944, pruned_loss=0.06813, over 3582602.55 frames. ], batch size: 48, lr: 1.62e-02, grad_scale: 8.0 +2023-03-08 21:18:24,833 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2706, 2.6659, 2.4876, 2.7255, 3.3460, 3.2241, 2.8401, 2.6663], + device='cuda:0'), covar=tensor([0.0187, 0.0256, 0.0590, 0.0393, 0.0144, 0.0143, 0.0366, 0.0336], + device='cuda:0'), in_proj_covar=tensor([0.0105, 0.0083, 0.0136, 0.0114, 0.0080, 0.0066, 0.0106, 0.0107], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 21:18:52,855 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4871, 5.4343, 5.0210, 5.4921, 5.4001, 4.8929, 5.4094, 5.1088], + device='cuda:0'), covar=tensor([0.0375, 0.0383, 0.1339, 0.0560, 0.0508, 0.0387, 0.0311, 0.0827], + device='cuda:0'), in_proj_covar=tensor([0.0345, 0.0391, 0.0541, 0.0313, 0.0286, 0.0362, 0.0378, 0.0478], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-08 21:18:53,705 INFO [train.py:898] (0/4) Epoch 7, batch 1600, loss[loss=0.1792, simple_loss=0.2626, pruned_loss=0.04792, over 18251.00 frames. ], tot_loss[loss=0.216, simple_loss=0.2949, pruned_loss=0.06857, over 3575852.78 frames. ], batch size: 45, lr: 1.62e-02, grad_scale: 8.0 +2023-03-08 21:19:05,448 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-03-08 21:19:25,463 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 3.006e+02 3.819e+02 4.598e+02 5.727e+02 1.079e+03, threshold=9.196e+02, percent-clipped=2.0 +2023-03-08 21:19:47,098 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23450.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:19:52,541 INFO [train.py:898] (0/4) Epoch 7, batch 1650, loss[loss=0.2175, simple_loss=0.2998, pruned_loss=0.0676, over 18569.00 frames. ], tot_loss[loss=0.2163, simple_loss=0.2952, pruned_loss=0.06876, over 3567558.71 frames. ], batch size: 54, lr: 1.62e-02, grad_scale: 8.0 +2023-03-08 21:20:04,685 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23464.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:20:24,360 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23481.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:20:32,458 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=23488.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:20:43,629 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=23498.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:20:51,443 INFO [train.py:898] (0/4) Epoch 7, batch 1700, loss[loss=0.1842, simple_loss=0.2607, pruned_loss=0.05381, over 18416.00 frames. ], tot_loss[loss=0.2154, simple_loss=0.2944, pruned_loss=0.0682, over 3575272.36 frames. ], batch size: 43, lr: 1.62e-02, grad_scale: 8.0 +2023-03-08 21:21:17,883 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7303, 4.2381, 2.6748, 3.9477, 3.9999, 4.2244, 4.0437, 2.7302], + device='cuda:0'), covar=tensor([0.0130, 0.0040, 0.0562, 0.0155, 0.0060, 0.0047, 0.0085, 0.0776], + device='cuda:0'), in_proj_covar=tensor([0.0067, 0.0053, 0.0082, 0.0068, 0.0065, 0.0055, 0.0069, 0.0086], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 21:21:21,160 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=23529.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:21:23,330 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.409e+02 3.809e+02 4.583e+02 5.657e+02 1.396e+03, threshold=9.165e+02, percent-clipped=6.0 +2023-03-08 21:21:44,290 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=23549.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 21:21:44,398 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.4225, 3.4696, 4.9315, 3.8388, 2.9754, 2.8226, 4.2153, 4.9015], + device='cuda:0'), covar=tensor([0.0779, 0.1161, 0.0071, 0.0330, 0.0848, 0.0971, 0.0289, 0.0130], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0203, 0.0076, 0.0146, 0.0163, 0.0165, 0.0151, 0.0106], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 21:21:50,751 INFO [train.py:898] (0/4) Epoch 7, batch 1750, loss[loss=0.2286, simple_loss=0.3081, pruned_loss=0.07456, over 18238.00 frames. ], tot_loss[loss=0.2164, simple_loss=0.2951, pruned_loss=0.06883, over 3558264.67 frames. ], batch size: 60, lr: 1.62e-02, grad_scale: 16.0 +2023-03-08 21:22:07,768 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0137, 4.9836, 4.5039, 4.9890, 4.9559, 4.2645, 4.8585, 4.5295], + device='cuda:0'), covar=tensor([0.0392, 0.0440, 0.1512, 0.0660, 0.0459, 0.0486, 0.0409, 0.0939], + device='cuda:0'), in_proj_covar=tensor([0.0345, 0.0392, 0.0545, 0.0316, 0.0285, 0.0367, 0.0386, 0.0490], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-08 21:22:13,759 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23573.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:22:27,628 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23585.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:22:50,509 INFO [train.py:898] (0/4) Epoch 7, batch 1800, loss[loss=0.1911, simple_loss=0.2656, pruned_loss=0.05829, over 18467.00 frames. ], tot_loss[loss=0.2152, simple_loss=0.2939, pruned_loss=0.06822, over 3569745.85 frames. ], batch size: 44, lr: 1.62e-02, grad_scale: 16.0 +2023-03-08 21:23:21,011 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.500e+02 3.624e+02 4.625e+02 5.651e+02 1.017e+03, threshold=9.251e+02, percent-clipped=3.0 +2023-03-08 21:23:47,600 INFO [train.py:898] (0/4) Epoch 7, batch 1850, loss[loss=0.2235, simple_loss=0.3107, pruned_loss=0.06816, over 18132.00 frames. ], tot_loss[loss=0.216, simple_loss=0.2951, pruned_loss=0.06844, over 3587442.00 frames. ], batch size: 62, lr: 1.61e-02, grad_scale: 16.0 +2023-03-08 21:24:45,517 INFO [train.py:898] (0/4) Epoch 7, batch 1900, loss[loss=0.1862, simple_loss=0.276, pruned_loss=0.04821, over 18402.00 frames. ], tot_loss[loss=0.2149, simple_loss=0.2942, pruned_loss=0.06781, over 3599392.98 frames. ], batch size: 48, lr: 1.61e-02, grad_scale: 16.0 +2023-03-08 21:25:00,512 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5713, 4.4536, 4.4879, 3.4373, 3.6872, 3.6858, 2.5137, 1.9178], + device='cuda:0'), covar=tensor([0.0233, 0.0124, 0.0065, 0.0232, 0.0310, 0.0181, 0.0760, 0.1040], + device='cuda:0'), in_proj_covar=tensor([0.0048, 0.0043, 0.0041, 0.0051, 0.0073, 0.0049, 0.0069, 0.0074], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0005, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 21:25:17,161 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.556e+02 4.080e+02 4.948e+02 6.191e+02 1.850e+03, threshold=9.895e+02, percent-clipped=8.0 +2023-03-08 21:25:43,793 INFO [train.py:898] (0/4) Epoch 7, batch 1950, loss[loss=0.1951, simple_loss=0.2725, pruned_loss=0.05886, over 18431.00 frames. ], tot_loss[loss=0.2146, simple_loss=0.2942, pruned_loss=0.06755, over 3610982.01 frames. ], batch size: 43, lr: 1.61e-02, grad_scale: 16.0 +2023-03-08 21:25:54,219 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23764.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:25:58,537 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9981, 5.4906, 5.0878, 5.2081, 5.0459, 5.0417, 5.5659, 5.5169], + device='cuda:0'), covar=tensor([0.1075, 0.0651, 0.0538, 0.0669, 0.1371, 0.0649, 0.0452, 0.0565], + device='cuda:0'), in_proj_covar=tensor([0.0454, 0.0367, 0.0285, 0.0409, 0.0552, 0.0405, 0.0487, 0.0385], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-08 21:26:42,025 INFO [train.py:898] (0/4) Epoch 7, batch 2000, loss[loss=0.2121, simple_loss=0.2912, pruned_loss=0.06651, over 18349.00 frames. ], tot_loss[loss=0.2132, simple_loss=0.2929, pruned_loss=0.06672, over 3611302.68 frames. ], batch size: 46, lr: 1.61e-02, grad_scale: 8.0 +2023-03-08 21:26:50,293 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=23812.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:27:13,551 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.566e+02 3.909e+02 4.831e+02 5.895e+02 1.179e+03, threshold=9.662e+02, percent-clipped=2.0 +2023-03-08 21:27:29,157 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=23844.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 21:27:40,874 INFO [train.py:898] (0/4) Epoch 7, batch 2050, loss[loss=0.2173, simple_loss=0.2978, pruned_loss=0.06836, over 18558.00 frames. ], tot_loss[loss=0.2134, simple_loss=0.2932, pruned_loss=0.0668, over 3608363.11 frames. ], batch size: 54, lr: 1.61e-02, grad_scale: 8.0 +2023-03-08 21:27:58,298 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.3467, 3.6455, 5.0676, 4.2317, 3.1529, 3.0526, 4.5881, 5.2384], + device='cuda:0'), covar=tensor([0.0899, 0.1496, 0.0108, 0.0287, 0.0870, 0.1011, 0.0257, 0.0163], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0211, 0.0079, 0.0149, 0.0166, 0.0168, 0.0155, 0.0107], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 21:28:01,679 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23873.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:28:04,402 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-08 21:28:16,276 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=23885.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:28:18,538 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7762, 3.9481, 2.3701, 3.9681, 4.8185, 2.3944, 3.5406, 3.6513], + device='cuda:0'), covar=tensor([0.0067, 0.0873, 0.1489, 0.0482, 0.0055, 0.1321, 0.0704, 0.0695], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0190, 0.0178, 0.0175, 0.0078, 0.0165, 0.0190, 0.0183], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0003, 0.0002, 0.0002, 0.0001, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 21:28:38,838 INFO [train.py:898] (0/4) Epoch 7, batch 2100, loss[loss=0.2145, simple_loss=0.2967, pruned_loss=0.06617, over 18397.00 frames. ], tot_loss[loss=0.2134, simple_loss=0.2931, pruned_loss=0.06688, over 3605300.52 frames. ], batch size: 52, lr: 1.61e-02, grad_scale: 8.0 +2023-03-08 21:28:57,349 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=23921.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:29:09,534 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.380e+02 3.638e+02 4.339e+02 5.858e+02 1.130e+03, threshold=8.677e+02, percent-clipped=1.0 +2023-03-08 21:29:10,876 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=23933.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:29:14,503 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8317, 3.5235, 1.8170, 4.4502, 2.9267, 4.3622, 1.8974, 3.6450], + device='cuda:0'), covar=tensor([0.0358, 0.0525, 0.1263, 0.0291, 0.0741, 0.0263, 0.1301, 0.0393], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0195, 0.0166, 0.0194, 0.0170, 0.0177, 0.0177, 0.0164], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 21:29:37,936 INFO [train.py:898] (0/4) Epoch 7, batch 2150, loss[loss=0.2194, simple_loss=0.3047, pruned_loss=0.06709, over 18575.00 frames. ], tot_loss[loss=0.2126, simple_loss=0.2922, pruned_loss=0.06653, over 3604779.57 frames. ], batch size: 54, lr: 1.60e-02, grad_scale: 8.0 +2023-03-08 21:30:20,603 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-08 21:30:31,224 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-24000.pt +2023-03-08 21:30:41,408 INFO [train.py:898] (0/4) Epoch 7, batch 2200, loss[loss=0.21, simple_loss=0.2992, pruned_loss=0.06038, over 18469.00 frames. ], tot_loss[loss=0.2126, simple_loss=0.2923, pruned_loss=0.06645, over 3591559.97 frames. ], batch size: 59, lr: 1.60e-02, grad_scale: 8.0 +2023-03-08 21:30:46,090 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7672, 4.5281, 4.8195, 4.4858, 4.5982, 4.6657, 4.9597, 4.8580], + device='cuda:0'), covar=tensor([0.0067, 0.0125, 0.0097, 0.0119, 0.0079, 0.0101, 0.0071, 0.0098], + device='cuda:0'), in_proj_covar=tensor([0.0069, 0.0050, 0.0051, 0.0065, 0.0055, 0.0074, 0.0064, 0.0064], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 21:31:11,876 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.180e+02 3.947e+02 4.705e+02 5.545e+02 1.194e+03, threshold=9.409e+02, percent-clipped=3.0 +2023-03-08 21:31:17,839 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24037.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:31:35,198 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1707, 4.3991, 2.3077, 4.2892, 5.1832, 2.1322, 3.7504, 4.1326], + device='cuda:0'), covar=tensor([0.0051, 0.0685, 0.1409, 0.0473, 0.0044, 0.1371, 0.0649, 0.0533], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0191, 0.0178, 0.0178, 0.0076, 0.0165, 0.0191, 0.0184], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0003, 0.0002, 0.0002, 0.0001, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 21:31:40,420 INFO [train.py:898] (0/4) Epoch 7, batch 2250, loss[loss=0.2075, simple_loss=0.2937, pruned_loss=0.06067, over 17768.00 frames. ], tot_loss[loss=0.2117, simple_loss=0.2916, pruned_loss=0.06588, over 3597576.10 frames. ], batch size: 70, lr: 1.60e-02, grad_scale: 8.0 +2023-03-08 21:32:30,411 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24098.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:32:38,537 INFO [train.py:898] (0/4) Epoch 7, batch 2300, loss[loss=0.235, simple_loss=0.3168, pruned_loss=0.07659, over 18566.00 frames. ], tot_loss[loss=0.2121, simple_loss=0.2917, pruned_loss=0.0663, over 3602443.61 frames. ], batch size: 54, lr: 1.60e-02, grad_scale: 8.0 +2023-03-08 21:32:59,605 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-08 21:33:08,948 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.576e+02 4.201e+02 5.158e+02 6.206e+02 1.861e+03, threshold=1.032e+03, percent-clipped=10.0 +2023-03-08 21:33:23,098 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=24144.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 21:33:36,540 INFO [train.py:898] (0/4) Epoch 7, batch 2350, loss[loss=0.1841, simple_loss=0.2579, pruned_loss=0.05518, over 17736.00 frames. ], tot_loss[loss=0.2117, simple_loss=0.2914, pruned_loss=0.06606, over 3612120.19 frames. ], batch size: 39, lr: 1.60e-02, grad_scale: 8.0 +2023-03-08 21:34:11,579 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4874, 2.8889, 2.5148, 2.6693, 3.5394, 3.2048, 2.9858, 2.7749], + device='cuda:0'), covar=tensor([0.0278, 0.0295, 0.0621, 0.0440, 0.0207, 0.0255, 0.0347, 0.0375], + device='cuda:0'), in_proj_covar=tensor([0.0108, 0.0086, 0.0142, 0.0118, 0.0083, 0.0070, 0.0110, 0.0113], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 21:34:19,529 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=24192.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:34:36,079 INFO [train.py:898] (0/4) Epoch 7, batch 2400, loss[loss=0.2139, simple_loss=0.2972, pruned_loss=0.06526, over 18471.00 frames. ], tot_loss[loss=0.2123, simple_loss=0.2921, pruned_loss=0.06629, over 3603886.90 frames. ], batch size: 59, lr: 1.60e-02, grad_scale: 8.0 +2023-03-08 21:34:47,364 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8588, 5.4566, 5.1724, 5.2329, 4.9251, 5.0138, 5.5697, 5.4642], + device='cuda:0'), covar=tensor([0.1174, 0.0725, 0.0495, 0.0683, 0.1492, 0.0684, 0.0518, 0.0622], + device='cuda:0'), in_proj_covar=tensor([0.0460, 0.0379, 0.0289, 0.0412, 0.0560, 0.0410, 0.0496, 0.0386], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-08 21:35:09,122 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.669e+02 3.723e+02 4.227e+02 5.514e+02 9.877e+02, threshold=8.454e+02, percent-clipped=0.0 +2023-03-08 21:35:20,058 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.02 vs. limit=5.0 +2023-03-08 21:35:35,292 INFO [train.py:898] (0/4) Epoch 7, batch 2450, loss[loss=0.2032, simple_loss=0.2881, pruned_loss=0.05915, over 18338.00 frames. ], tot_loss[loss=0.2121, simple_loss=0.2919, pruned_loss=0.06613, over 3603368.11 frames. ], batch size: 55, lr: 1.59e-02, grad_scale: 8.0 +2023-03-08 21:36:33,631 INFO [train.py:898] (0/4) Epoch 7, batch 2500, loss[loss=0.183, simple_loss=0.2518, pruned_loss=0.05707, over 18380.00 frames. ], tot_loss[loss=0.211, simple_loss=0.2909, pruned_loss=0.06551, over 3610457.77 frames. ], batch size: 42, lr: 1.59e-02, grad_scale: 8.0 +2023-03-08 21:37:06,514 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.422e+02 3.679e+02 4.544e+02 5.517e+02 9.659e+02, threshold=9.088e+02, percent-clipped=5.0 +2023-03-08 21:37:31,457 INFO [train.py:898] (0/4) Epoch 7, batch 2550, loss[loss=0.2045, simple_loss=0.2856, pruned_loss=0.06165, over 18372.00 frames. ], tot_loss[loss=0.2122, simple_loss=0.2924, pruned_loss=0.06602, over 3616491.46 frames. ], batch size: 50, lr: 1.59e-02, grad_scale: 8.0 +2023-03-08 21:37:37,223 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-08 21:38:05,834 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-03-08 21:38:16,207 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=24393.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:38:29,985 INFO [train.py:898] (0/4) Epoch 7, batch 2600, loss[loss=0.2177, simple_loss=0.2886, pruned_loss=0.07337, over 18245.00 frames. ], tot_loss[loss=0.2128, simple_loss=0.2927, pruned_loss=0.0665, over 3604875.21 frames. ], batch size: 45, lr: 1.59e-02, grad_scale: 4.0 +2023-03-08 21:39:05,100 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.252e+02 3.780e+02 4.661e+02 5.457e+02 1.160e+03, threshold=9.322e+02, percent-clipped=5.0 +2023-03-08 21:39:22,756 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1516, 4.2769, 2.2383, 4.2083, 5.2254, 2.4839, 3.7135, 3.8170], + device='cuda:0'), covar=tensor([0.0073, 0.0881, 0.1583, 0.0483, 0.0042, 0.1234, 0.0609, 0.0762], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0190, 0.0177, 0.0174, 0.0075, 0.0163, 0.0187, 0.0182], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0003, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 21:39:29,153 INFO [train.py:898] (0/4) Epoch 7, batch 2650, loss[loss=0.2158, simple_loss=0.3031, pruned_loss=0.06421, over 18307.00 frames. ], tot_loss[loss=0.2123, simple_loss=0.2923, pruned_loss=0.06619, over 3604967.22 frames. ], batch size: 54, lr: 1.59e-02, grad_scale: 4.0 +2023-03-08 21:39:36,226 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6892, 3.5723, 2.0056, 4.4873, 3.1700, 4.8169, 2.2112, 3.9074], + device='cuda:0'), covar=tensor([0.0417, 0.0656, 0.1247, 0.0319, 0.0741, 0.0144, 0.1093, 0.0312], + device='cuda:0'), in_proj_covar=tensor([0.0165, 0.0193, 0.0166, 0.0194, 0.0167, 0.0177, 0.0176, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 21:40:27,804 INFO [train.py:898] (0/4) Epoch 7, batch 2700, loss[loss=0.2288, simple_loss=0.3104, pruned_loss=0.07353, over 17165.00 frames. ], tot_loss[loss=0.212, simple_loss=0.2918, pruned_loss=0.06611, over 3603051.27 frames. ], batch size: 78, lr: 1.59e-02, grad_scale: 4.0 +2023-03-08 21:41:02,502 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.355e+02 3.494e+02 4.469e+02 5.662e+02 1.849e+03, threshold=8.938e+02, percent-clipped=8.0 +2023-03-08 21:41:17,683 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24547.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:41:26,599 INFO [train.py:898] (0/4) Epoch 7, batch 2750, loss[loss=0.2057, simple_loss=0.2921, pruned_loss=0.05964, over 18342.00 frames. ], tot_loss[loss=0.2107, simple_loss=0.2908, pruned_loss=0.06531, over 3611762.69 frames. ], batch size: 56, lr: 1.59e-02, grad_scale: 4.0 +2023-03-08 21:41:43,515 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24569.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:42:06,553 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8573, 3.6144, 5.2723, 3.1171, 4.3354, 2.4535, 2.8440, 2.1069], + device='cuda:0'), covar=tensor([0.0807, 0.0684, 0.0052, 0.0511, 0.0591, 0.2169, 0.2278, 0.1516], + device='cuda:0'), in_proj_covar=tensor([0.0181, 0.0196, 0.0094, 0.0151, 0.0207, 0.0235, 0.0243, 0.0193], + device='cuda:0'), out_proj_covar=tensor([1.6542e-04, 1.8406e-04, 9.1008e-05, 1.4079e-04, 1.9525e-04, 2.2137e-04, + 2.3081e-04, 1.8454e-04], device='cuda:0') +2023-03-08 21:42:25,725 INFO [train.py:898] (0/4) Epoch 7, batch 2800, loss[loss=0.2522, simple_loss=0.3292, pruned_loss=0.08764, over 18471.00 frames. ], tot_loss[loss=0.2117, simple_loss=0.2921, pruned_loss=0.06564, over 3613295.79 frames. ], batch size: 59, lr: 1.58e-02, grad_scale: 8.0 +2023-03-08 21:42:27,283 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1975, 4.2968, 2.2978, 4.4264, 5.2171, 2.4391, 3.7193, 3.8305], + device='cuda:0'), covar=tensor([0.0059, 0.0871, 0.1541, 0.0401, 0.0037, 0.1211, 0.0617, 0.0709], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0194, 0.0180, 0.0177, 0.0077, 0.0165, 0.0190, 0.0187], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0003, 0.0002, 0.0002, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 21:42:29,380 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24608.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:42:55,519 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24630.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:43:01,137 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.569e+02 3.885e+02 4.580e+02 5.331e+02 1.147e+03, threshold=9.161e+02, percent-clipped=3.0 +2023-03-08 21:43:19,535 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9453, 4.9633, 5.0236, 4.8056, 4.7682, 4.8223, 5.1923, 5.0964], + device='cuda:0'), covar=tensor([0.0064, 0.0061, 0.0065, 0.0077, 0.0060, 0.0093, 0.0059, 0.0082], + device='cuda:0'), in_proj_covar=tensor([0.0069, 0.0050, 0.0052, 0.0065, 0.0055, 0.0075, 0.0065, 0.0063], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 21:43:23,573 INFO [train.py:898] (0/4) Epoch 7, batch 2850, loss[loss=0.2218, simple_loss=0.3005, pruned_loss=0.07156, over 18119.00 frames. ], tot_loss[loss=0.2126, simple_loss=0.293, pruned_loss=0.06608, over 3602490.73 frames. ], batch size: 62, lr: 1.58e-02, grad_scale: 4.0 +2023-03-08 21:44:09,068 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=24693.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:44:22,196 INFO [train.py:898] (0/4) Epoch 7, batch 2900, loss[loss=0.2248, simple_loss=0.3132, pruned_loss=0.06825, over 18294.00 frames. ], tot_loss[loss=0.2129, simple_loss=0.2933, pruned_loss=0.0663, over 3600840.83 frames. ], batch size: 57, lr: 1.58e-02, grad_scale: 4.0 +2023-03-08 21:44:57,698 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.094e+02 3.800e+02 4.687e+02 5.853e+02 1.844e+03, threshold=9.374e+02, percent-clipped=5.0 +2023-03-08 21:45:05,116 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=24741.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:45:20,856 INFO [train.py:898] (0/4) Epoch 7, batch 2950, loss[loss=0.2109, simple_loss=0.2976, pruned_loss=0.0621, over 18265.00 frames. ], tot_loss[loss=0.2129, simple_loss=0.2934, pruned_loss=0.06616, over 3598777.93 frames. ], batch size: 57, lr: 1.58e-02, grad_scale: 4.0 +2023-03-08 21:45:32,618 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24765.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:46:20,381 INFO [train.py:898] (0/4) Epoch 7, batch 3000, loss[loss=0.1803, simple_loss=0.2582, pruned_loss=0.05121, over 18364.00 frames. ], tot_loss[loss=0.2115, simple_loss=0.2922, pruned_loss=0.06537, over 3601081.83 frames. ], batch size: 46, lr: 1.58e-02, grad_scale: 4.0 +2023-03-08 21:46:20,384 INFO [train.py:923] (0/4) Computing validation loss +2023-03-08 21:46:29,975 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.3903, 1.8687, 2.2164, 2.3904, 2.8321, 3.7180, 3.5044, 2.9846], + device='cuda:0'), covar=tensor([0.0775, 0.1449, 0.1878, 0.1053, 0.1244, 0.0140, 0.0372, 0.0404], + device='cuda:0'), in_proj_covar=tensor([0.0196, 0.0249, 0.0247, 0.0223, 0.0329, 0.0146, 0.0220, 0.0170], + device='cuda:0'), out_proj_covar=tensor([1.3079e-04, 1.6739e-04, 1.7281e-04, 1.3848e-04, 2.2163e-04, 9.4843e-05, + 1.4186e-04, 1.1387e-04], device='cuda:0') +2023-03-08 21:46:32,366 INFO [train.py:932] (0/4) Epoch 7, validation: loss=0.1689, simple_loss=0.2715, pruned_loss=0.03314, over 944034.00 frames. +2023-03-08 21:46:32,367 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-08 21:46:57,849 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24826.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:47:08,248 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.797e+02 4.000e+02 4.650e+02 5.894e+02 1.091e+03, threshold=9.301e+02, percent-clipped=1.0 +2023-03-08 21:47:23,287 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24848.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:47:30,758 INFO [train.py:898] (0/4) Epoch 7, batch 3050, loss[loss=0.223, simple_loss=0.3016, pruned_loss=0.07216, over 16064.00 frames. ], tot_loss[loss=0.2112, simple_loss=0.2918, pruned_loss=0.06524, over 3600029.04 frames. ], batch size: 94, lr: 1.58e-02, grad_scale: 4.0 +2023-03-08 21:47:44,792 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.4995, 2.0584, 2.7853, 2.8168, 3.6203, 5.2583, 4.5708, 4.4350], + device='cuda:0'), covar=tensor([0.0809, 0.1436, 0.1726, 0.0971, 0.1315, 0.0046, 0.0328, 0.0259], + device='cuda:0'), in_proj_covar=tensor([0.0196, 0.0249, 0.0248, 0.0224, 0.0332, 0.0146, 0.0223, 0.0171], + device='cuda:0'), out_proj_covar=tensor([1.3055e-04, 1.6713e-04, 1.7358e-04, 1.3940e-04, 2.2336e-04, 9.5283e-05, + 1.4339e-04, 1.1450e-04], device='cuda:0') +2023-03-08 21:48:26,665 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=24903.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:48:28,716 INFO [train.py:898] (0/4) Epoch 7, batch 3100, loss[loss=0.1932, simple_loss=0.2782, pruned_loss=0.0541, over 18500.00 frames. ], tot_loss[loss=0.2111, simple_loss=0.291, pruned_loss=0.06556, over 3597241.26 frames. ], batch size: 47, lr: 1.57e-02, grad_scale: 2.0 +2023-03-08 21:48:33,530 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=24909.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:48:52,685 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=24925.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:49:05,434 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.662e+02 4.170e+02 4.880e+02 6.294e+02 1.409e+03, threshold=9.761e+02, percent-clipped=6.0 +2023-03-08 21:49:27,492 INFO [train.py:898] (0/4) Epoch 7, batch 3150, loss[loss=0.177, simple_loss=0.2589, pruned_loss=0.04756, over 18269.00 frames. ], tot_loss[loss=0.2109, simple_loss=0.2908, pruned_loss=0.06556, over 3588700.17 frames. ], batch size: 47, lr: 1.57e-02, grad_scale: 2.0 +2023-03-08 21:50:09,920 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-03-08 21:50:15,338 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=24995.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:50:26,650 INFO [train.py:898] (0/4) Epoch 7, batch 3200, loss[loss=0.2164, simple_loss=0.3018, pruned_loss=0.06552, over 18484.00 frames. ], tot_loss[loss=0.2102, simple_loss=0.29, pruned_loss=0.06522, over 3593076.08 frames. ], batch size: 59, lr: 1.57e-02, grad_scale: 4.0 +2023-03-08 21:50:39,547 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-08 21:50:44,796 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8271, 4.9720, 4.9745, 4.7103, 4.7495, 4.7781, 5.1920, 5.0917], + device='cuda:0'), covar=tensor([0.0065, 0.0066, 0.0075, 0.0100, 0.0069, 0.0098, 0.0059, 0.0096], + device='cuda:0'), in_proj_covar=tensor([0.0070, 0.0051, 0.0053, 0.0067, 0.0056, 0.0075, 0.0065, 0.0063], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0002, 0.0003, 0.0002, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 21:50:59,182 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4386, 3.1898, 1.7432, 4.2439, 2.7665, 4.2734, 2.0681, 3.4780], + device='cuda:0'), covar=tensor([0.0478, 0.0720, 0.1378, 0.0312, 0.0856, 0.0209, 0.1076, 0.0454], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0194, 0.0168, 0.0196, 0.0170, 0.0183, 0.0175, 0.0166], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 21:51:03,253 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.251e+02 3.758e+02 4.426e+02 5.545e+02 1.381e+03, threshold=8.852e+02, percent-clipped=2.0 +2023-03-08 21:51:25,676 INFO [train.py:898] (0/4) Epoch 7, batch 3250, loss[loss=0.2195, simple_loss=0.3066, pruned_loss=0.06614, over 18357.00 frames. ], tot_loss[loss=0.2097, simple_loss=0.2897, pruned_loss=0.06488, over 3593758.42 frames. ], batch size: 55, lr: 1.57e-02, grad_scale: 4.0 +2023-03-08 21:51:27,269 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=25056.0, num_to_drop=1, layers_to_drop={3} +2023-03-08 21:51:49,739 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9995, 3.8065, 5.1220, 3.2281, 4.1472, 2.6527, 3.0364, 2.0464], + device='cuda:0'), covar=tensor([0.0696, 0.0609, 0.0040, 0.0465, 0.0497, 0.1997, 0.2051, 0.1411], + device='cuda:0'), in_proj_covar=tensor([0.0178, 0.0190, 0.0093, 0.0149, 0.0204, 0.0233, 0.0241, 0.0190], + device='cuda:0'), out_proj_covar=tensor([1.6121e-04, 1.7882e-04, 8.9945e-05, 1.3958e-04, 1.9187e-04, 2.1869e-04, + 2.2785e-04, 1.8104e-04], device='cuda:0') +2023-03-08 21:52:24,642 INFO [train.py:898] (0/4) Epoch 7, batch 3300, loss[loss=0.1903, simple_loss=0.2851, pruned_loss=0.04772, over 18489.00 frames. ], tot_loss[loss=0.2098, simple_loss=0.2899, pruned_loss=0.06484, over 3593884.27 frames. ], batch size: 53, lr: 1.57e-02, grad_scale: 4.0 +2023-03-08 21:52:42,711 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25121.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:52:55,235 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2151, 5.4366, 2.8551, 5.0853, 4.9828, 5.4109, 5.1183, 2.8763], + device='cuda:0'), covar=tensor([0.0137, 0.0033, 0.0674, 0.0066, 0.0063, 0.0049, 0.0083, 0.0875], + device='cuda:0'), in_proj_covar=tensor([0.0069, 0.0057, 0.0084, 0.0072, 0.0068, 0.0057, 0.0072, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 21:53:01,192 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.469e+02 3.921e+02 4.598e+02 5.927e+02 2.644e+03, threshold=9.195e+02, percent-clipped=9.0 +2023-03-08 21:53:01,530 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3308, 5.3258, 4.8090, 5.2484, 5.2409, 4.6805, 5.1462, 4.9182], + device='cuda:0'), covar=tensor([0.0404, 0.0363, 0.1442, 0.0679, 0.0537, 0.0392, 0.0375, 0.0782], + device='cuda:0'), in_proj_covar=tensor([0.0348, 0.0388, 0.0542, 0.0319, 0.0291, 0.0368, 0.0393, 0.0491], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-08 21:53:08,508 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6545, 5.1640, 5.2141, 5.2287, 4.7090, 5.0719, 3.8879, 5.0650], + device='cuda:0'), covar=tensor([0.0250, 0.0428, 0.0298, 0.0313, 0.0462, 0.0325, 0.2029, 0.0326], + device='cuda:0'), in_proj_covar=tensor([0.0145, 0.0192, 0.0177, 0.0188, 0.0185, 0.0192, 0.0258, 0.0177], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0005, 0.0005, 0.0005, 0.0006, 0.0005], + device='cuda:0') +2023-03-08 21:53:23,206 INFO [train.py:898] (0/4) Epoch 7, batch 3350, loss[loss=0.1976, simple_loss=0.293, pruned_loss=0.05114, over 18507.00 frames. ], tot_loss[loss=0.2096, simple_loss=0.2899, pruned_loss=0.0647, over 3603845.71 frames. ], batch size: 53, lr: 1.57e-02, grad_scale: 4.0 +2023-03-08 21:54:19,479 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25203.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:54:20,380 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25204.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:54:21,382 INFO [train.py:898] (0/4) Epoch 7, batch 3400, loss[loss=0.1963, simple_loss=0.2775, pruned_loss=0.05754, over 18546.00 frames. ], tot_loss[loss=0.2093, simple_loss=0.2895, pruned_loss=0.06453, over 3609889.65 frames. ], batch size: 49, lr: 1.57e-02, grad_scale: 4.0 +2023-03-08 21:54:44,300 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25225.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:54:56,995 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.655e+02 3.734e+02 4.394e+02 5.547e+02 1.008e+03, threshold=8.789e+02, percent-clipped=3.0 +2023-03-08 21:55:11,371 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6136, 3.5218, 3.3123, 2.8470, 3.3120, 2.6058, 2.4509, 3.5513], + device='cuda:0'), covar=tensor([0.0032, 0.0047, 0.0070, 0.0106, 0.0062, 0.0151, 0.0167, 0.0057], + device='cuda:0'), in_proj_covar=tensor([0.0068, 0.0091, 0.0081, 0.0127, 0.0082, 0.0128, 0.0134, 0.0074], + device='cuda:0'), out_proj_covar=tensor([9.3816e-05, 1.3750e-04, 1.2057e-04, 1.9845e-04, 1.2221e-04, 1.9751e-04, + 2.0489e-04, 1.0864e-04], device='cuda:0') +2023-03-08 21:55:14,351 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=25251.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:55:19,392 INFO [train.py:898] (0/4) Epoch 7, batch 3450, loss[loss=0.2095, simple_loss=0.2921, pruned_loss=0.06343, over 18297.00 frames. ], tot_loss[loss=0.2098, simple_loss=0.2899, pruned_loss=0.06489, over 3597755.23 frames. ], batch size: 57, lr: 1.56e-02, grad_scale: 4.0 +2023-03-08 21:55:39,659 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=25273.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:55:59,516 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-08 21:56:17,142 INFO [train.py:898] (0/4) Epoch 7, batch 3500, loss[loss=0.1895, simple_loss=0.2587, pruned_loss=0.06015, over 18474.00 frames. ], tot_loss[loss=0.209, simple_loss=0.2891, pruned_loss=0.06449, over 3599096.51 frames. ], batch size: 44, lr: 1.56e-02, grad_scale: 2.0 +2023-03-08 21:56:32,163 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-08 21:56:53,684 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.862e+02 4.141e+02 4.748e+02 6.314e+02 1.477e+03, threshold=9.496e+02, percent-clipped=11.0 +2023-03-08 21:56:59,164 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1832, 5.8326, 5.2849, 5.5252, 5.2939, 5.2842, 5.8463, 5.8132], + device='cuda:0'), covar=tensor([0.1309, 0.0599, 0.0513, 0.0697, 0.1392, 0.0673, 0.0521, 0.0578], + device='cuda:0'), in_proj_covar=tensor([0.0470, 0.0375, 0.0289, 0.0413, 0.0570, 0.0414, 0.0515, 0.0392], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-08 21:56:59,412 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5434, 1.8311, 2.6242, 2.6747, 3.3630, 4.9593, 4.3547, 4.0351], + device='cuda:0'), covar=tensor([0.0811, 0.1600, 0.1672, 0.1029, 0.1396, 0.0052, 0.0317, 0.0274], + device='cuda:0'), in_proj_covar=tensor([0.0198, 0.0250, 0.0250, 0.0223, 0.0330, 0.0145, 0.0223, 0.0169], + device='cuda:0'), out_proj_covar=tensor([1.3138e-04, 1.6764e-04, 1.7327e-04, 1.3746e-04, 2.2089e-04, 9.4445e-05, + 1.4301e-04, 1.1265e-04], device='cuda:0') +2023-03-08 21:57:08,484 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25351.0, num_to_drop=1, layers_to_drop={3} +2023-03-08 21:57:12,756 INFO [train.py:898] (0/4) Epoch 7, batch 3550, loss[loss=0.2308, simple_loss=0.3104, pruned_loss=0.07564, over 17942.00 frames. ], tot_loss[loss=0.2089, simple_loss=0.289, pruned_loss=0.06444, over 3603840.54 frames. ], batch size: 65, lr: 1.56e-02, grad_scale: 2.0 +2023-03-08 21:57:52,128 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5187, 3.2645, 1.9472, 4.1787, 2.8491, 4.3106, 1.8032, 3.5626], + device='cuda:0'), covar=tensor([0.0484, 0.0793, 0.1355, 0.0429, 0.0925, 0.0229, 0.1343, 0.0411], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0194, 0.0169, 0.0196, 0.0168, 0.0185, 0.0177, 0.0169], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 21:58:07,629 INFO [train.py:898] (0/4) Epoch 7, batch 3600, loss[loss=0.177, simple_loss=0.2509, pruned_loss=0.05153, over 16775.00 frames. ], tot_loss[loss=0.2081, simple_loss=0.288, pruned_loss=0.06412, over 3591918.32 frames. ], batch size: 37, lr: 1.56e-02, grad_scale: 4.0 +2023-03-08 21:58:24,784 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25421.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 21:58:31,815 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=25428.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 21:58:40,130 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.475e+02 3.916e+02 4.844e+02 6.068e+02 1.506e+03, threshold=9.689e+02, percent-clipped=7.0 +2023-03-08 21:58:42,536 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-7.pt +2023-03-08 21:59:12,577 INFO [train.py:898] (0/4) Epoch 8, batch 0, loss[loss=0.2163, simple_loss=0.2997, pruned_loss=0.0665, over 18353.00 frames. ], tot_loss[loss=0.2163, simple_loss=0.2997, pruned_loss=0.0665, over 18353.00 frames. ], batch size: 55, lr: 1.47e-02, grad_scale: 8.0 +2023-03-08 21:59:12,579 INFO [train.py:923] (0/4) Computing validation loss +2023-03-08 21:59:24,299 INFO [train.py:932] (0/4) Epoch 8, validation: loss=0.17, simple_loss=0.2728, pruned_loss=0.03358, over 944034.00 frames. +2023-03-08 21:59:24,300 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-08 21:59:59,983 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=25469.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:00:22,359 INFO [train.py:898] (0/4) Epoch 8, batch 50, loss[loss=0.2127, simple_loss=0.3023, pruned_loss=0.06152, over 17944.00 frames. ], tot_loss[loss=0.2117, simple_loss=0.2939, pruned_loss=0.0648, over 809917.35 frames. ], batch size: 65, lr: 1.47e-02, grad_scale: 8.0 +2023-03-08 22:00:22,728 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=25489.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 22:00:39,150 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25504.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:01:18,379 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.473e+02 3.465e+02 4.274e+02 5.083e+02 8.127e+02, threshold=8.548e+02, percent-clipped=0.0 +2023-03-08 22:01:20,766 INFO [train.py:898] (0/4) Epoch 8, batch 100, loss[loss=0.2043, simple_loss=0.2812, pruned_loss=0.06371, over 18510.00 frames. ], tot_loss[loss=0.207, simple_loss=0.2882, pruned_loss=0.06292, over 1441527.06 frames. ], batch size: 47, lr: 1.47e-02, grad_scale: 8.0 +2023-03-08 22:01:35,581 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=25552.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:02:19,594 INFO [train.py:898] (0/4) Epoch 8, batch 150, loss[loss=0.2048, simple_loss=0.2905, pruned_loss=0.05954, over 18477.00 frames. ], tot_loss[loss=0.2079, simple_loss=0.2892, pruned_loss=0.06332, over 1908885.59 frames. ], batch size: 51, lr: 1.46e-02, grad_scale: 8.0 +2023-03-08 22:02:47,523 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-08 22:03:16,567 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.739e+02 3.654e+02 4.521e+02 5.323e+02 1.367e+03, threshold=9.043e+02, percent-clipped=1.0 +2023-03-08 22:03:18,861 INFO [train.py:898] (0/4) Epoch 8, batch 200, loss[loss=0.2108, simple_loss=0.2952, pruned_loss=0.06323, over 18297.00 frames. ], tot_loss[loss=0.2079, simple_loss=0.2893, pruned_loss=0.06329, over 2292133.69 frames. ], batch size: 49, lr: 1.46e-02, grad_scale: 8.0 +2023-03-08 22:03:30,061 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.6385, 5.5281, 5.1136, 5.5088, 5.5688, 4.9452, 5.4549, 5.1798], + device='cuda:0'), covar=tensor([0.0289, 0.0325, 0.1374, 0.0717, 0.0427, 0.0376, 0.0312, 0.0754], + device='cuda:0'), in_proj_covar=tensor([0.0342, 0.0396, 0.0543, 0.0313, 0.0286, 0.0369, 0.0383, 0.0495], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-08 22:03:32,340 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=25651.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 22:03:35,945 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5281, 2.9603, 2.5843, 2.7933, 3.5497, 3.4718, 3.0238, 2.8814], + device='cuda:0'), covar=tensor([0.0250, 0.0217, 0.0691, 0.0386, 0.0186, 0.0169, 0.0383, 0.0426], + device='cuda:0'), in_proj_covar=tensor([0.0109, 0.0087, 0.0140, 0.0119, 0.0085, 0.0072, 0.0113, 0.0114], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 22:04:17,974 INFO [train.py:898] (0/4) Epoch 8, batch 250, loss[loss=0.2017, simple_loss=0.2679, pruned_loss=0.06771, over 18499.00 frames. ], tot_loss[loss=0.2067, simple_loss=0.2878, pruned_loss=0.0628, over 2581547.08 frames. ], batch size: 44, lr: 1.46e-02, grad_scale: 8.0 +2023-03-08 22:04:29,462 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=25699.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:04:58,370 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-08 22:05:02,815 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4797, 3.9582, 4.1257, 3.4399, 3.5034, 3.2561, 2.5032, 1.8444], + device='cuda:0'), covar=tensor([0.0218, 0.0227, 0.0072, 0.0212, 0.0316, 0.0257, 0.0728, 0.1033], + device='cuda:0'), in_proj_covar=tensor([0.0048, 0.0043, 0.0040, 0.0052, 0.0073, 0.0049, 0.0068, 0.0074], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0005, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 22:05:14,434 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.583e+02 3.732e+02 4.597e+02 5.455e+02 9.874e+02, threshold=9.193e+02, percent-clipped=1.0 +2023-03-08 22:05:17,281 INFO [train.py:898] (0/4) Epoch 8, batch 300, loss[loss=0.2108, simple_loss=0.2887, pruned_loss=0.06646, over 16144.00 frames. ], tot_loss[loss=0.2066, simple_loss=0.288, pruned_loss=0.06255, over 2799237.95 frames. ], batch size: 94, lr: 1.46e-02, grad_scale: 8.0 +2023-03-08 22:05:31,306 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-08 22:06:10,045 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=25784.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 22:06:15,820 INFO [train.py:898] (0/4) Epoch 8, batch 350, loss[loss=0.216, simple_loss=0.3002, pruned_loss=0.06594, over 18340.00 frames. ], tot_loss[loss=0.2067, simple_loss=0.2881, pruned_loss=0.06267, over 2977714.19 frames. ], batch size: 56, lr: 1.46e-02, grad_scale: 8.0 +2023-03-08 22:06:59,022 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1454, 4.9481, 5.1701, 5.3170, 5.0363, 5.8356, 5.5425, 5.1841], + device='cuda:0'), covar=tensor([0.0807, 0.0700, 0.0648, 0.0527, 0.1325, 0.0706, 0.0501, 0.1496], + device='cuda:0'), in_proj_covar=tensor([0.0270, 0.0211, 0.0213, 0.0213, 0.0255, 0.0305, 0.0201, 0.0299], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 22:07:11,597 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.442e+02 3.462e+02 4.130e+02 5.163e+02 1.142e+03, threshold=8.260e+02, percent-clipped=1.0 +2023-03-08 22:07:12,067 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5005, 5.2150, 5.2196, 5.0928, 4.8023, 5.0818, 4.3772, 5.0536], + device='cuda:0'), covar=tensor([0.0252, 0.0274, 0.0196, 0.0307, 0.0346, 0.0228, 0.1317, 0.0257], + device='cuda:0'), in_proj_covar=tensor([0.0149, 0.0197, 0.0180, 0.0196, 0.0188, 0.0197, 0.0262, 0.0183], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0006, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-08 22:07:14,543 INFO [train.py:898] (0/4) Epoch 8, batch 400, loss[loss=0.2057, simple_loss=0.2897, pruned_loss=0.06086, over 18499.00 frames. ], tot_loss[loss=0.2069, simple_loss=0.2886, pruned_loss=0.06259, over 3126196.30 frames. ], batch size: 51, lr: 1.46e-02, grad_scale: 8.0 +2023-03-08 22:07:37,280 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=25858.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 22:08:13,029 INFO [train.py:898] (0/4) Epoch 8, batch 450, loss[loss=0.2307, simple_loss=0.3097, pruned_loss=0.07582, over 18337.00 frames. ], tot_loss[loss=0.2076, simple_loss=0.2892, pruned_loss=0.06297, over 3236678.42 frames. ], batch size: 56, lr: 1.46e-02, grad_scale: 8.0 +2023-03-08 22:08:48,649 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=25919.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 22:09:09,624 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.689e+02 3.778e+02 4.755e+02 5.799e+02 1.474e+03, threshold=9.510e+02, percent-clipped=4.0 +2023-03-08 22:09:11,981 INFO [train.py:898] (0/4) Epoch 8, batch 500, loss[loss=0.1853, simple_loss=0.27, pruned_loss=0.05025, over 18559.00 frames. ], tot_loss[loss=0.2077, simple_loss=0.289, pruned_loss=0.06319, over 3299745.68 frames. ], batch size: 49, lr: 1.45e-02, grad_scale: 8.0 +2023-03-08 22:09:16,881 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8462, 3.5732, 2.2174, 4.4077, 3.0696, 4.5803, 2.3963, 4.1849], + device='cuda:0'), covar=tensor([0.0411, 0.0598, 0.1188, 0.0373, 0.0760, 0.0179, 0.0958, 0.0269], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0199, 0.0174, 0.0201, 0.0172, 0.0193, 0.0178, 0.0171], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 22:09:28,348 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8365, 5.0015, 4.9838, 4.7310, 4.7730, 4.7124, 5.1470, 5.1369], + device='cuda:0'), covar=tensor([0.0053, 0.0063, 0.0062, 0.0078, 0.0064, 0.0096, 0.0066, 0.0078], + device='cuda:0'), in_proj_covar=tensor([0.0071, 0.0052, 0.0053, 0.0067, 0.0055, 0.0076, 0.0065, 0.0063], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0002, 0.0003, 0.0002, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 22:09:47,766 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6917, 4.0709, 4.1078, 4.1007, 3.8589, 4.0499, 3.6211, 3.9947], + device='cuda:0'), covar=tensor([0.0291, 0.0360, 0.0258, 0.0334, 0.0352, 0.0236, 0.1014, 0.0325], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0199, 0.0183, 0.0199, 0.0193, 0.0200, 0.0265, 0.0186], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0006, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-08 22:10:10,519 INFO [train.py:898] (0/4) Epoch 8, batch 550, loss[loss=0.2141, simple_loss=0.2974, pruned_loss=0.06539, over 17734.00 frames. ], tot_loss[loss=0.2072, simple_loss=0.2885, pruned_loss=0.06293, over 3367998.49 frames. ], batch size: 70, lr: 1.45e-02, grad_scale: 8.0 +2023-03-08 22:10:24,347 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-26000.pt +2023-03-08 22:11:10,803 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.250e+02 3.451e+02 4.216e+02 4.967e+02 1.068e+03, threshold=8.432e+02, percent-clipped=2.0 +2023-03-08 22:11:13,149 INFO [train.py:898] (0/4) Epoch 8, batch 600, loss[loss=0.2418, simple_loss=0.3179, pruned_loss=0.08285, over 18328.00 frames. ], tot_loss[loss=0.2069, simple_loss=0.2881, pruned_loss=0.06292, over 3419652.17 frames. ], batch size: 56, lr: 1.45e-02, grad_scale: 8.0 +2023-03-08 22:11:15,886 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6081, 3.2994, 4.1108, 2.9248, 3.5969, 2.6985, 2.5818, 2.3128], + device='cuda:0'), covar=tensor([0.0634, 0.0568, 0.0085, 0.0405, 0.0542, 0.1541, 0.1673, 0.1124], + device='cuda:0'), in_proj_covar=tensor([0.0181, 0.0194, 0.0096, 0.0153, 0.0207, 0.0235, 0.0247, 0.0196], + device='cuda:0'), out_proj_covar=tensor([1.6326e-04, 1.8158e-04, 9.0486e-05, 1.4236e-04, 1.9334e-04, 2.2095e-04, + 2.3161e-04, 1.8492e-04], device='cuda:0') +2023-03-08 22:11:18,692 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6052, 3.4389, 1.8882, 4.4327, 2.7232, 4.5418, 2.0745, 3.9368], + device='cuda:0'), covar=tensor([0.0507, 0.0781, 0.1496, 0.0413, 0.0972, 0.0281, 0.1248, 0.0385], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0197, 0.0172, 0.0202, 0.0172, 0.0193, 0.0179, 0.0171], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 22:11:46,105 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3720, 3.1336, 1.7890, 4.1026, 2.5410, 4.1128, 1.9654, 3.5932], + device='cuda:0'), covar=tensor([0.0475, 0.0803, 0.1447, 0.0328, 0.0969, 0.0208, 0.1158, 0.0353], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0194, 0.0169, 0.0199, 0.0168, 0.0190, 0.0175, 0.0168], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 22:12:06,627 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26084.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 22:12:12,075 INFO [train.py:898] (0/4) Epoch 8, batch 650, loss[loss=0.2389, simple_loss=0.3193, pruned_loss=0.07928, over 18492.00 frames. ], tot_loss[loss=0.2078, simple_loss=0.2886, pruned_loss=0.06347, over 3442973.07 frames. ], batch size: 59, lr: 1.45e-02, grad_scale: 8.0 +2023-03-08 22:12:28,534 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26102.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:13:03,425 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=26132.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 22:13:08,681 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.333e+02 3.685e+02 4.568e+02 5.569e+02 1.081e+03, threshold=9.136e+02, percent-clipped=5.0 +2023-03-08 22:13:11,017 INFO [train.py:898] (0/4) Epoch 8, batch 700, loss[loss=0.1766, simple_loss=0.259, pruned_loss=0.04709, over 18265.00 frames. ], tot_loss[loss=0.2076, simple_loss=0.2885, pruned_loss=0.06338, over 3471305.77 frames. ], batch size: 45, lr: 1.45e-02, grad_scale: 8.0 +2023-03-08 22:13:39,855 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26163.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:13:46,042 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26168.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:14:09,253 INFO [train.py:898] (0/4) Epoch 8, batch 750, loss[loss=0.2217, simple_loss=0.3134, pruned_loss=0.06505, over 18295.00 frames. ], tot_loss[loss=0.208, simple_loss=0.2892, pruned_loss=0.06345, over 3505429.90 frames. ], batch size: 57, lr: 1.45e-02, grad_scale: 8.0 +2023-03-08 22:14:39,869 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26214.0, num_to_drop=1, layers_to_drop={3} +2023-03-08 22:14:57,286 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26229.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:15:05,904 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.506e+02 3.613e+02 4.168e+02 4.953e+02 1.109e+03, threshold=8.337e+02, percent-clipped=3.0 +2023-03-08 22:15:08,155 INFO [train.py:898] (0/4) Epoch 8, batch 800, loss[loss=0.2643, simple_loss=0.3201, pruned_loss=0.1043, over 12276.00 frames. ], tot_loss[loss=0.2072, simple_loss=0.2883, pruned_loss=0.06304, over 3531498.02 frames. ], batch size: 129, lr: 1.45e-02, grad_scale: 8.0 +2023-03-08 22:15:21,462 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26250.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:16:05,489 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.68 vs. limit=5.0 +2023-03-08 22:16:07,214 INFO [train.py:898] (0/4) Epoch 8, batch 850, loss[loss=0.1525, simple_loss=0.2374, pruned_loss=0.03382, over 18576.00 frames. ], tot_loss[loss=0.2068, simple_loss=0.2878, pruned_loss=0.06288, over 3548100.42 frames. ], batch size: 45, lr: 1.45e-02, grad_scale: 8.0 +2023-03-08 22:16:33,363 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26311.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:17:04,267 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.475e+02 3.702e+02 4.449e+02 5.668e+02 1.488e+03, threshold=8.898e+02, percent-clipped=3.0 +2023-03-08 22:17:06,541 INFO [train.py:898] (0/4) Epoch 8, batch 900, loss[loss=0.2275, simple_loss=0.3027, pruned_loss=0.07621, over 17745.00 frames. ], tot_loss[loss=0.2065, simple_loss=0.2877, pruned_loss=0.0627, over 3561284.88 frames. ], batch size: 70, lr: 1.44e-02, grad_scale: 8.0 +2023-03-08 22:17:46,087 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0711, 4.3035, 5.3730, 3.4895, 4.2858, 2.7263, 3.1954, 2.2133], + device='cuda:0'), covar=tensor([0.0715, 0.0498, 0.0050, 0.0462, 0.0575, 0.2190, 0.2022, 0.1462], + device='cuda:0'), in_proj_covar=tensor([0.0179, 0.0193, 0.0096, 0.0150, 0.0204, 0.0233, 0.0247, 0.0193], + device='cuda:0'), out_proj_covar=tensor([1.6151e-04, 1.8041e-04, 9.1124e-05, 1.3963e-04, 1.9054e-04, 2.1882e-04, + 2.3041e-04, 1.8185e-04], device='cuda:0') +2023-03-08 22:18:00,262 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7551, 3.6498, 5.1292, 3.3375, 4.2792, 2.5307, 2.9899, 1.9652], + device='cuda:0'), covar=tensor([0.0897, 0.0698, 0.0080, 0.0502, 0.0501, 0.2169, 0.2146, 0.1706], + device='cuda:0'), in_proj_covar=tensor([0.0180, 0.0193, 0.0096, 0.0150, 0.0204, 0.0233, 0.0248, 0.0193], + device='cuda:0'), out_proj_covar=tensor([1.6160e-04, 1.8076e-04, 9.1099e-05, 1.3970e-04, 1.9084e-04, 2.1848e-04, + 2.3075e-04, 1.8187e-04], device='cuda:0') +2023-03-08 22:18:06,763 INFO [train.py:898] (0/4) Epoch 8, batch 950, loss[loss=0.1952, simple_loss=0.2735, pruned_loss=0.05846, over 18357.00 frames. ], tot_loss[loss=0.2075, simple_loss=0.289, pruned_loss=0.06299, over 3566390.75 frames. ], batch size: 46, lr: 1.44e-02, grad_scale: 8.0 +2023-03-08 22:18:30,604 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7774, 4.7861, 4.8169, 4.6207, 4.6313, 4.6083, 5.0643, 5.0530], + device='cuda:0'), covar=tensor([0.0060, 0.0072, 0.0074, 0.0097, 0.0067, 0.0114, 0.0076, 0.0095], + device='cuda:0'), in_proj_covar=tensor([0.0069, 0.0051, 0.0051, 0.0065, 0.0054, 0.0076, 0.0063, 0.0062], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0003, 0.0002, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 22:18:36,980 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26414.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:18:58,413 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-08 22:19:04,445 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.644e+02 3.699e+02 4.361e+02 5.008e+02 1.213e+03, threshold=8.721e+02, percent-clipped=3.0 +2023-03-08 22:19:06,723 INFO [train.py:898] (0/4) Epoch 8, batch 1000, loss[loss=0.2204, simple_loss=0.3028, pruned_loss=0.06895, over 18324.00 frames. ], tot_loss[loss=0.2081, simple_loss=0.2898, pruned_loss=0.06318, over 3568729.03 frames. ], batch size: 54, lr: 1.44e-02, grad_scale: 8.0 +2023-03-08 22:19:29,034 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26458.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:19:37,936 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9391, 4.9447, 5.0632, 4.8358, 4.7764, 4.8403, 5.2597, 5.1739], + device='cuda:0'), covar=tensor([0.0058, 0.0063, 0.0049, 0.0088, 0.0068, 0.0098, 0.0055, 0.0097], + device='cuda:0'), in_proj_covar=tensor([0.0071, 0.0051, 0.0052, 0.0066, 0.0055, 0.0077, 0.0063, 0.0063], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0002, 0.0003, 0.0002, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 22:19:50,132 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26475.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:20:06,563 INFO [train.py:898] (0/4) Epoch 8, batch 1050, loss[loss=0.1879, simple_loss=0.2563, pruned_loss=0.05974, over 18455.00 frames. ], tot_loss[loss=0.2082, simple_loss=0.2897, pruned_loss=0.06337, over 3565939.25 frames. ], batch size: 43, lr: 1.44e-02, grad_scale: 8.0 +2023-03-08 22:20:35,313 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26514.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 22:20:47,865 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26524.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:21:03,507 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.558e+02 3.535e+02 4.269e+02 5.489e+02 1.390e+03, threshold=8.539e+02, percent-clipped=8.0 +2023-03-08 22:21:05,720 INFO [train.py:898] (0/4) Epoch 8, batch 1100, loss[loss=0.2319, simple_loss=0.3094, pruned_loss=0.07715, over 18259.00 frames. ], tot_loss[loss=0.2088, simple_loss=0.2904, pruned_loss=0.06364, over 3578124.88 frames. ], batch size: 60, lr: 1.44e-02, grad_scale: 8.0 +2023-03-08 22:21:22,564 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3687, 4.4764, 2.6476, 4.4817, 5.3379, 2.4894, 3.9863, 3.9718], + device='cuda:0'), covar=tensor([0.0068, 0.0901, 0.1400, 0.0454, 0.0051, 0.1296, 0.0576, 0.0729], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0201, 0.0182, 0.0181, 0.0080, 0.0170, 0.0194, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0003, 0.0002, 0.0002, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 22:21:32,869 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=26562.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 22:21:47,618 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1603, 4.2392, 2.5338, 4.3253, 5.1854, 2.4432, 3.8267, 3.8391], + device='cuda:0'), covar=tensor([0.0072, 0.0933, 0.1432, 0.0451, 0.0050, 0.1357, 0.0602, 0.0667], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0200, 0.0182, 0.0180, 0.0080, 0.0170, 0.0193, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0003, 0.0002, 0.0002, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 22:22:05,738 INFO [train.py:898] (0/4) Epoch 8, batch 1150, loss[loss=0.2144, simple_loss=0.3058, pruned_loss=0.06155, over 18392.00 frames. ], tot_loss[loss=0.2085, simple_loss=0.2899, pruned_loss=0.06348, over 3567995.94 frames. ], batch size: 52, lr: 1.44e-02, grad_scale: 8.0 +2023-03-08 22:22:24,956 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26606.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:22:26,625 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-08 22:23:02,044 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.150e+02 3.587e+02 4.399e+02 5.427e+02 1.423e+03, threshold=8.799e+02, percent-clipped=5.0 +2023-03-08 22:23:04,989 INFO [train.py:898] (0/4) Epoch 8, batch 1200, loss[loss=0.2191, simple_loss=0.3025, pruned_loss=0.06788, over 17238.00 frames. ], tot_loss[loss=0.2087, simple_loss=0.2901, pruned_loss=0.06366, over 3579710.99 frames. ], batch size: 78, lr: 1.44e-02, grad_scale: 8.0 +2023-03-08 22:24:03,513 INFO [train.py:898] (0/4) Epoch 8, batch 1250, loss[loss=0.179, simple_loss=0.2576, pruned_loss=0.05018, over 18392.00 frames. ], tot_loss[loss=0.2084, simple_loss=0.2898, pruned_loss=0.06355, over 3582615.78 frames. ], batch size: 42, lr: 1.43e-02, grad_scale: 8.0 +2023-03-08 22:24:59,403 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.195e+02 3.433e+02 4.104e+02 5.033e+02 1.173e+03, threshold=8.208e+02, percent-clipped=2.0 +2023-03-08 22:25:01,385 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7969, 5.3815, 5.3712, 5.3098, 4.9264, 5.2614, 4.6102, 5.1160], + device='cuda:0'), covar=tensor([0.0167, 0.0232, 0.0165, 0.0271, 0.0361, 0.0196, 0.1032, 0.0266], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0198, 0.0183, 0.0204, 0.0194, 0.0202, 0.0265, 0.0188], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0006, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-08 22:25:02,168 INFO [train.py:898] (0/4) Epoch 8, batch 1300, loss[loss=0.1966, simple_loss=0.2779, pruned_loss=0.05766, over 18499.00 frames. ], tot_loss[loss=0.2092, simple_loss=0.2904, pruned_loss=0.06397, over 3588627.55 frames. ], batch size: 47, lr: 1.43e-02, grad_scale: 8.0 +2023-03-08 22:25:24,993 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26758.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:25:35,229 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6335, 2.9283, 4.2324, 3.9712, 2.8418, 4.5760, 3.9843, 2.8608], + device='cuda:0'), covar=tensor([0.0345, 0.1113, 0.0226, 0.0216, 0.1148, 0.0162, 0.0318, 0.0789], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0208, 0.0126, 0.0129, 0.0201, 0.0171, 0.0181, 0.0182], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 22:25:38,683 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=26770.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:25:59,946 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1076, 5.2197, 2.7850, 5.1045, 4.9089, 5.2622, 5.0980, 2.3794], + device='cuda:0'), covar=tensor([0.0138, 0.0067, 0.0678, 0.0069, 0.0090, 0.0076, 0.0094, 0.1092], + device='cuda:0'), in_proj_covar=tensor([0.0070, 0.0060, 0.0087, 0.0074, 0.0069, 0.0057, 0.0072, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0005, 0.0004, 0.0003, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 22:26:00,818 INFO [train.py:898] (0/4) Epoch 8, batch 1350, loss[loss=0.2309, simple_loss=0.3158, pruned_loss=0.07295, over 18356.00 frames. ], tot_loss[loss=0.2076, simple_loss=0.2885, pruned_loss=0.06331, over 3590022.95 frames. ], batch size: 56, lr: 1.43e-02, grad_scale: 8.0 +2023-03-08 22:26:21,851 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=26806.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:26:42,230 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26824.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:26:57,539 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.883e+02 3.529e+02 4.433e+02 5.428e+02 1.307e+03, threshold=8.866e+02, percent-clipped=5.0 +2023-03-08 22:27:00,014 INFO [train.py:898] (0/4) Epoch 8, batch 1400, loss[loss=0.2249, simple_loss=0.3023, pruned_loss=0.07373, over 17124.00 frames. ], tot_loss[loss=0.2061, simple_loss=0.287, pruned_loss=0.06261, over 3602008.01 frames. ], batch size: 78, lr: 1.43e-02, grad_scale: 8.0 +2023-03-08 22:27:39,755 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=26872.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:27:50,201 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6149, 3.2282, 4.2056, 4.0622, 2.7812, 4.7478, 4.0331, 2.9507], + device='cuda:0'), covar=tensor([0.0414, 0.1046, 0.0251, 0.0231, 0.1292, 0.0125, 0.0397, 0.0809], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0207, 0.0126, 0.0128, 0.0202, 0.0171, 0.0181, 0.0182], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 22:27:54,800 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.60 vs. limit=5.0 +2023-03-08 22:27:59,674 INFO [train.py:898] (0/4) Epoch 8, batch 1450, loss[loss=0.1985, simple_loss=0.2857, pruned_loss=0.05563, over 18129.00 frames. ], tot_loss[loss=0.2066, simple_loss=0.2875, pruned_loss=0.0628, over 3576628.83 frames. ], batch size: 62, lr: 1.43e-02, grad_scale: 8.0 +2023-03-08 22:28:18,318 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8605, 4.6836, 4.8891, 4.6048, 4.6825, 4.6750, 5.1373, 5.0152], + device='cuda:0'), covar=tensor([0.0061, 0.0093, 0.0075, 0.0092, 0.0065, 0.0112, 0.0072, 0.0087], + device='cuda:0'), in_proj_covar=tensor([0.0071, 0.0051, 0.0052, 0.0066, 0.0055, 0.0077, 0.0064, 0.0063], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0002, 0.0003, 0.0002, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 22:28:18,356 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26904.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 22:28:21,092 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=26906.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:28:39,454 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9900, 3.7544, 5.2271, 3.0731, 4.3623, 2.7405, 3.0275, 1.9817], + device='cuda:0'), covar=tensor([0.0718, 0.0636, 0.0041, 0.0489, 0.0466, 0.1768, 0.2021, 0.1512], + device='cuda:0'), in_proj_covar=tensor([0.0182, 0.0196, 0.0097, 0.0152, 0.0208, 0.0234, 0.0250, 0.0195], + device='cuda:0'), out_proj_covar=tensor([1.6311e-04, 1.8262e-04, 9.1294e-05, 1.4081e-04, 1.9387e-04, 2.1977e-04, + 2.3313e-04, 1.8358e-04], device='cuda:0') +2023-03-08 22:28:56,608 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.356e+02 3.688e+02 4.427e+02 5.301e+02 1.390e+03, threshold=8.853e+02, percent-clipped=1.0 +2023-03-08 22:28:58,834 INFO [train.py:898] (0/4) Epoch 8, batch 1500, loss[loss=0.2528, simple_loss=0.3179, pruned_loss=0.09389, over 12449.00 frames. ], tot_loss[loss=0.2059, simple_loss=0.2868, pruned_loss=0.06248, over 3583908.92 frames. ], batch size: 129, lr: 1.43e-02, grad_scale: 8.0 +2023-03-08 22:29:12,512 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=26950.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:29:17,100 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=26954.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:29:30,357 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=26965.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 22:29:52,142 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7190, 3.6940, 3.4066, 3.0693, 3.3402, 2.7607, 2.6155, 3.7525], + device='cuda:0'), covar=tensor([0.0034, 0.0054, 0.0083, 0.0115, 0.0065, 0.0152, 0.0182, 0.0040], + device='cuda:0'), in_proj_covar=tensor([0.0069, 0.0095, 0.0086, 0.0129, 0.0084, 0.0129, 0.0137, 0.0073], + device='cuda:0'), out_proj_covar=tensor([9.4430e-05, 1.4027e-04, 1.2584e-04, 1.9974e-04, 1.2278e-04, 1.9688e-04, + 2.0923e-04, 1.0537e-04], device='cuda:0') +2023-03-08 22:29:58,009 INFO [train.py:898] (0/4) Epoch 8, batch 1550, loss[loss=0.2044, simple_loss=0.2811, pruned_loss=0.06381, over 18546.00 frames. ], tot_loss[loss=0.206, simple_loss=0.2869, pruned_loss=0.06254, over 3596501.62 frames. ], batch size: 49, lr: 1.43e-02, grad_scale: 8.0 +2023-03-08 22:30:24,858 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=27011.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:30:54,391 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.371e+02 3.938e+02 4.762e+02 5.471e+02 1.136e+03, threshold=9.525e+02, percent-clipped=4.0 +2023-03-08 22:30:56,673 INFO [train.py:898] (0/4) Epoch 8, batch 1600, loss[loss=0.1904, simple_loss=0.2844, pruned_loss=0.04817, over 18504.00 frames. ], tot_loss[loss=0.2069, simple_loss=0.2877, pruned_loss=0.06301, over 3589881.61 frames. ], batch size: 53, lr: 1.43e-02, grad_scale: 8.0 +2023-03-08 22:31:00,786 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3320, 5.0160, 5.4163, 5.4008, 5.1958, 5.9961, 5.5918, 5.4216], + device='cuda:0'), covar=tensor([0.0883, 0.0655, 0.0669, 0.0501, 0.1308, 0.0719, 0.0644, 0.1456], + device='cuda:0'), in_proj_covar=tensor([0.0292, 0.0221, 0.0224, 0.0227, 0.0273, 0.0326, 0.0218, 0.0320], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 22:31:34,746 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=27070.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:31:35,005 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5070, 1.7568, 2.5469, 2.5022, 3.2680, 5.0148, 4.3707, 4.0454], + device='cuda:0'), covar=tensor([0.0981, 0.1915, 0.1921, 0.1240, 0.1721, 0.0061, 0.0325, 0.0330], + device='cuda:0'), in_proj_covar=tensor([0.0207, 0.0259, 0.0264, 0.0227, 0.0336, 0.0152, 0.0229, 0.0177], + device='cuda:0'), out_proj_covar=tensor([1.3545e-04, 1.7065e-04, 1.7820e-04, 1.3801e-04, 2.2042e-04, 9.8485e-05, + 1.4265e-04, 1.1632e-04], device='cuda:0') +2023-03-08 22:31:56,222 INFO [train.py:898] (0/4) Epoch 8, batch 1650, loss[loss=0.1977, simple_loss=0.2855, pruned_loss=0.05494, over 18478.00 frames. ], tot_loss[loss=0.2061, simple_loss=0.2873, pruned_loss=0.06245, over 3595771.23 frames. ], batch size: 53, lr: 1.42e-02, grad_scale: 8.0 +2023-03-08 22:32:31,915 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=27118.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:32:53,450 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.332e+02 3.429e+02 4.052e+02 5.174e+02 1.566e+03, threshold=8.105e+02, percent-clipped=2.0 +2023-03-08 22:32:55,968 INFO [train.py:898] (0/4) Epoch 8, batch 1700, loss[loss=0.2029, simple_loss=0.2968, pruned_loss=0.05455, over 18580.00 frames. ], tot_loss[loss=0.2062, simple_loss=0.2875, pruned_loss=0.06247, over 3591575.44 frames. ], batch size: 54, lr: 1.42e-02, grad_scale: 8.0 +2023-03-08 22:33:07,209 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=27148.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:33:55,118 INFO [train.py:898] (0/4) Epoch 8, batch 1750, loss[loss=0.2151, simple_loss=0.296, pruned_loss=0.0671, over 18223.00 frames. ], tot_loss[loss=0.2059, simple_loss=0.2872, pruned_loss=0.06232, over 3585911.68 frames. ], batch size: 60, lr: 1.42e-02, grad_scale: 8.0 +2023-03-08 22:33:56,624 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7142, 4.3252, 4.4108, 3.1909, 3.5028, 3.5529, 2.1423, 2.0381], + device='cuda:0'), covar=tensor([0.0222, 0.0128, 0.0066, 0.0268, 0.0378, 0.0211, 0.0863, 0.1000], + device='cuda:0'), in_proj_covar=tensor([0.0049, 0.0043, 0.0041, 0.0053, 0.0073, 0.0049, 0.0069, 0.0073], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0005, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 22:34:02,611 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.29 vs. limit=5.0 +2023-03-08 22:34:19,909 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=27209.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:34:52,836 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.362e+02 3.433e+02 4.337e+02 5.594e+02 1.098e+03, threshold=8.674e+02, percent-clipped=7.0 +2023-03-08 22:34:55,192 INFO [train.py:898] (0/4) Epoch 8, batch 1800, loss[loss=0.2088, simple_loss=0.2922, pruned_loss=0.06264, over 13400.00 frames. ], tot_loss[loss=0.2052, simple_loss=0.2866, pruned_loss=0.06192, over 3590119.43 frames. ], batch size: 130, lr: 1.42e-02, grad_scale: 8.0 +2023-03-08 22:35:19,926 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=27260.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 22:35:54,553 INFO [train.py:898] (0/4) Epoch 8, batch 1850, loss[loss=0.1727, simple_loss=0.2444, pruned_loss=0.05054, over 18504.00 frames. ], tot_loss[loss=0.2044, simple_loss=0.2856, pruned_loss=0.0616, over 3580636.41 frames. ], batch size: 44, lr: 1.42e-02, grad_scale: 8.0 +2023-03-08 22:36:10,700 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4438, 2.7600, 2.5562, 2.7464, 3.4750, 3.4483, 3.0198, 2.9603], + device='cuda:0'), covar=tensor([0.0212, 0.0366, 0.0602, 0.0356, 0.0199, 0.0134, 0.0343, 0.0317], + device='cuda:0'), in_proj_covar=tensor([0.0106, 0.0091, 0.0142, 0.0122, 0.0091, 0.0073, 0.0118, 0.0112], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 22:36:13,716 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=27306.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:36:21,709 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-08 22:36:30,459 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.91 vs. limit=2.0 +2023-03-08 22:36:43,634 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0738, 4.5277, 4.3193, 4.2500, 4.1007, 4.6827, 4.4281, 4.2649], + device='cuda:0'), covar=tensor([0.1158, 0.0964, 0.0728, 0.0728, 0.1464, 0.1093, 0.0724, 0.1450], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0216, 0.0220, 0.0223, 0.0263, 0.0318, 0.0211, 0.0311], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 22:36:50,629 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.586e+02 3.923e+02 4.847e+02 5.995e+02 1.610e+03, threshold=9.695e+02, percent-clipped=7.0 +2023-03-08 22:36:53,070 INFO [train.py:898] (0/4) Epoch 8, batch 1900, loss[loss=0.1995, simple_loss=0.2786, pruned_loss=0.0602, over 18559.00 frames. ], tot_loss[loss=0.2053, simple_loss=0.2864, pruned_loss=0.06204, over 3587218.05 frames. ], batch size: 49, lr: 1.42e-02, grad_scale: 16.0 +2023-03-08 22:37:34,697 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8029, 3.7903, 5.3104, 4.5421, 3.5103, 3.2126, 4.6551, 5.3631], + device='cuda:0'), covar=tensor([0.0769, 0.1512, 0.0061, 0.0287, 0.0713, 0.0954, 0.0231, 0.0113], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0224, 0.0085, 0.0152, 0.0171, 0.0173, 0.0161, 0.0118], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0001, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 22:37:51,492 INFO [train.py:898] (0/4) Epoch 8, batch 1950, loss[loss=0.1647, simple_loss=0.2442, pruned_loss=0.04257, over 18537.00 frames. ], tot_loss[loss=0.205, simple_loss=0.286, pruned_loss=0.06202, over 3591842.82 frames. ], batch size: 45, lr: 1.42e-02, grad_scale: 16.0 +2023-03-08 22:38:47,712 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.257e+02 3.355e+02 4.079e+02 5.085e+02 1.650e+03, threshold=8.157e+02, percent-clipped=2.0 +2023-03-08 22:38:49,969 INFO [train.py:898] (0/4) Epoch 8, batch 2000, loss[loss=0.213, simple_loss=0.2964, pruned_loss=0.06484, over 18372.00 frames. ], tot_loss[loss=0.2059, simple_loss=0.2871, pruned_loss=0.06234, over 3594125.61 frames. ], batch size: 55, lr: 1.42e-02, grad_scale: 16.0 +2023-03-08 22:38:54,248 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3549, 5.6044, 3.1046, 5.3694, 5.2770, 5.6565, 5.5174, 2.8246], + device='cuda:0'), covar=tensor([0.0129, 0.0037, 0.0578, 0.0056, 0.0053, 0.0038, 0.0060, 0.0827], + device='cuda:0'), in_proj_covar=tensor([0.0070, 0.0059, 0.0086, 0.0073, 0.0069, 0.0056, 0.0072, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0005, 0.0004, 0.0003, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 22:39:48,961 INFO [train.py:898] (0/4) Epoch 8, batch 2050, loss[loss=0.2078, simple_loss=0.297, pruned_loss=0.05928, over 18620.00 frames. ], tot_loss[loss=0.2049, simple_loss=0.2858, pruned_loss=0.06195, over 3604465.71 frames. ], batch size: 52, lr: 1.41e-02, grad_scale: 8.0 +2023-03-08 22:40:06,982 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=27504.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:40:14,759 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3394, 5.2589, 5.2846, 5.4389, 5.2593, 5.9925, 5.5815, 5.3771], + device='cuda:0'), covar=tensor([0.0853, 0.0571, 0.0756, 0.0517, 0.1310, 0.0725, 0.0661, 0.1496], + device='cuda:0'), in_proj_covar=tensor([0.0285, 0.0215, 0.0221, 0.0222, 0.0265, 0.0325, 0.0213, 0.0312], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 22:40:46,826 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.547e+02 3.620e+02 4.320e+02 6.156e+02 2.106e+03, threshold=8.640e+02, percent-clipped=12.0 +2023-03-08 22:40:47,998 INFO [train.py:898] (0/4) Epoch 8, batch 2100, loss[loss=0.2119, simple_loss=0.2927, pruned_loss=0.06554, over 18380.00 frames. ], tot_loss[loss=0.2057, simple_loss=0.2866, pruned_loss=0.06245, over 3606349.67 frames. ], batch size: 50, lr: 1.41e-02, grad_scale: 8.0 +2023-03-08 22:41:12,648 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=27560.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 22:41:47,087 INFO [train.py:898] (0/4) Epoch 8, batch 2150, loss[loss=0.2293, simple_loss=0.315, pruned_loss=0.07175, over 18285.00 frames. ], tot_loss[loss=0.2044, simple_loss=0.2853, pruned_loss=0.06174, over 3604790.70 frames. ], batch size: 54, lr: 1.41e-02, grad_scale: 8.0 +2023-03-08 22:42:07,690 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=27606.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:42:09,879 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=27608.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 22:42:45,316 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.528e+02 3.451e+02 4.193e+02 5.044e+02 8.620e+02, threshold=8.386e+02, percent-clipped=0.0 +2023-03-08 22:42:46,502 INFO [train.py:898] (0/4) Epoch 8, batch 2200, loss[loss=0.1764, simple_loss=0.256, pruned_loss=0.0484, over 18250.00 frames. ], tot_loss[loss=0.2046, simple_loss=0.2856, pruned_loss=0.06178, over 3602632.32 frames. ], batch size: 45, lr: 1.41e-02, grad_scale: 8.0 +2023-03-08 22:42:48,547 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-08 22:43:03,735 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=27654.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:43:46,148 INFO [train.py:898] (0/4) Epoch 8, batch 2250, loss[loss=0.1912, simple_loss=0.27, pruned_loss=0.05619, over 18410.00 frames. ], tot_loss[loss=0.2046, simple_loss=0.2856, pruned_loss=0.06181, over 3600838.93 frames. ], batch size: 48, lr: 1.41e-02, grad_scale: 8.0 +2023-03-08 22:44:15,655 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=27714.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 22:44:44,344 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.230e+02 4.253e+02 4.744e+02 6.262e+02 1.251e+03, threshold=9.489e+02, percent-clipped=5.0 +2023-03-08 22:44:45,491 INFO [train.py:898] (0/4) Epoch 8, batch 2300, loss[loss=0.205, simple_loss=0.2905, pruned_loss=0.05978, over 18282.00 frames. ], tot_loss[loss=0.2045, simple_loss=0.2854, pruned_loss=0.06176, over 3601903.30 frames. ], batch size: 57, lr: 1.41e-02, grad_scale: 8.0 +2023-03-08 22:44:49,288 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7923, 5.4257, 5.4478, 5.3493, 4.9984, 5.2862, 4.6627, 5.2607], + device='cuda:0'), covar=tensor([0.0205, 0.0219, 0.0163, 0.0260, 0.0287, 0.0251, 0.1065, 0.0239], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0198, 0.0185, 0.0205, 0.0195, 0.0207, 0.0268, 0.0188], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0006, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-08 22:45:28,221 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=27775.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 22:45:44,275 INFO [train.py:898] (0/4) Epoch 8, batch 2350, loss[loss=0.2161, simple_loss=0.3007, pruned_loss=0.06581, over 18483.00 frames. ], tot_loss[loss=0.2043, simple_loss=0.2856, pruned_loss=0.06147, over 3607106.54 frames. ], batch size: 51, lr: 1.41e-02, grad_scale: 8.0 +2023-03-08 22:46:03,002 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=27804.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:46:42,187 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.997e+02 3.557e+02 4.056e+02 4.769e+02 1.044e+03, threshold=8.112e+02, percent-clipped=2.0 +2023-03-08 22:46:43,704 INFO [train.py:898] (0/4) Epoch 8, batch 2400, loss[loss=0.2184, simple_loss=0.2899, pruned_loss=0.07339, over 18274.00 frames. ], tot_loss[loss=0.2045, simple_loss=0.2858, pruned_loss=0.06163, over 3600839.34 frames. ], batch size: 49, lr: 1.41e-02, grad_scale: 8.0 +2023-03-08 22:46:52,765 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9583, 3.9565, 5.5309, 3.8649, 4.5356, 2.8477, 2.9745, 2.4333], + device='cuda:0'), covar=tensor([0.0735, 0.0613, 0.0034, 0.0373, 0.0526, 0.1799, 0.2224, 0.1252], + device='cuda:0'), in_proj_covar=tensor([0.0179, 0.0194, 0.0096, 0.0151, 0.0208, 0.0231, 0.0251, 0.0192], + device='cuda:0'), out_proj_covar=tensor([1.6064e-04, 1.8029e-04, 9.0800e-05, 1.3917e-04, 1.9295e-04, 2.1671e-04, + 2.3174e-04, 1.8064e-04], device='cuda:0') +2023-03-08 22:46:59,299 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=27852.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:47:01,943 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3903, 2.8267, 2.4084, 2.6836, 3.4752, 3.3264, 3.0519, 3.0214], + device='cuda:0'), covar=tensor([0.0276, 0.0296, 0.0684, 0.0338, 0.0226, 0.0153, 0.0325, 0.0279], + device='cuda:0'), in_proj_covar=tensor([0.0108, 0.0093, 0.0143, 0.0123, 0.0092, 0.0073, 0.0119, 0.0117], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 22:47:10,151 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7968, 3.6698, 5.0714, 3.1991, 4.2205, 2.5729, 2.8534, 1.9713], + device='cuda:0'), covar=tensor([0.0839, 0.0736, 0.0053, 0.0509, 0.0527, 0.2028, 0.2268, 0.1594], + device='cuda:0'), in_proj_covar=tensor([0.0181, 0.0196, 0.0097, 0.0151, 0.0209, 0.0233, 0.0252, 0.0193], + device='cuda:0'), out_proj_covar=tensor([1.6196e-04, 1.8186e-04, 9.1459e-05, 1.4005e-04, 1.9417e-04, 2.1867e-04, + 2.3295e-04, 1.8204e-04], device='cuda:0') +2023-03-08 22:47:42,092 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2329, 5.3136, 4.2207, 5.2079, 5.2428, 4.8623, 5.0644, 4.7592], + device='cuda:0'), covar=tensor([0.0662, 0.0614, 0.2960, 0.0907, 0.0721, 0.0496, 0.0728, 0.1162], + device='cuda:0'), in_proj_covar=tensor([0.0368, 0.0424, 0.0578, 0.0337, 0.0309, 0.0390, 0.0412, 0.0529], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-08 22:47:42,908 INFO [train.py:898] (0/4) Epoch 8, batch 2450, loss[loss=0.199, simple_loss=0.2694, pruned_loss=0.06428, over 18567.00 frames. ], tot_loss[loss=0.204, simple_loss=0.2854, pruned_loss=0.06125, over 3600593.26 frames. ], batch size: 45, lr: 1.40e-02, grad_scale: 8.0 +2023-03-08 22:48:08,280 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3601, 2.8464, 4.0940, 3.6890, 2.3794, 4.3245, 3.7764, 2.4571], + device='cuda:0'), covar=tensor([0.0379, 0.1031, 0.0145, 0.0218, 0.1360, 0.0113, 0.0361, 0.0971], + device='cuda:0'), in_proj_covar=tensor([0.0175, 0.0209, 0.0129, 0.0126, 0.0203, 0.0169, 0.0184, 0.0185], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 22:48:41,269 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.646e+02 3.668e+02 4.454e+02 5.440e+02 2.886e+03, threshold=8.907e+02, percent-clipped=9.0 +2023-03-08 22:48:42,421 INFO [train.py:898] (0/4) Epoch 8, batch 2500, loss[loss=0.1868, simple_loss=0.2758, pruned_loss=0.04888, over 18480.00 frames. ], tot_loss[loss=0.2048, simple_loss=0.2864, pruned_loss=0.06157, over 3599312.38 frames. ], batch size: 53, lr: 1.40e-02, grad_scale: 8.0 +2023-03-08 22:49:41,158 INFO [train.py:898] (0/4) Epoch 8, batch 2550, loss[loss=0.1895, simple_loss=0.2694, pruned_loss=0.05479, over 18285.00 frames. ], tot_loss[loss=0.2041, simple_loss=0.2859, pruned_loss=0.06117, over 3611425.27 frames. ], batch size: 49, lr: 1.40e-02, grad_scale: 8.0 +2023-03-08 22:49:54,623 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-28000.pt +2023-03-08 22:50:20,667 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-08 22:50:37,311 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0605, 5.0208, 5.0891, 5.1400, 5.0406, 5.6853, 5.2874, 5.1391], + device='cuda:0'), covar=tensor([0.0898, 0.0624, 0.0766, 0.0627, 0.1254, 0.0682, 0.0640, 0.1388], + device='cuda:0'), in_proj_covar=tensor([0.0283, 0.0217, 0.0224, 0.0223, 0.0264, 0.0324, 0.0213, 0.0314], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 22:50:45,005 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.293e+02 3.705e+02 4.339e+02 5.125e+02 8.562e+02, threshold=8.678e+02, percent-clipped=0.0 +2023-03-08 22:50:45,030 INFO [train.py:898] (0/4) Epoch 8, batch 2600, loss[loss=0.2173, simple_loss=0.3024, pruned_loss=0.06609, over 18323.00 frames. ], tot_loss[loss=0.2041, simple_loss=0.2859, pruned_loss=0.06115, over 3605418.50 frames. ], batch size: 54, lr: 1.40e-02, grad_scale: 4.0 +2023-03-08 22:51:22,116 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28070.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 22:51:43,982 INFO [train.py:898] (0/4) Epoch 8, batch 2650, loss[loss=0.1759, simple_loss=0.2574, pruned_loss=0.04718, over 18248.00 frames. ], tot_loss[loss=0.2043, simple_loss=0.286, pruned_loss=0.06127, over 3603088.12 frames. ], batch size: 45, lr: 1.40e-02, grad_scale: 4.0 +2023-03-08 22:51:47,940 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.93 vs. limit=5.0 +2023-03-08 22:52:42,844 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.340e+02 3.798e+02 4.426e+02 5.240e+02 9.211e+02, threshold=8.852e+02, percent-clipped=2.0 +2023-03-08 22:52:42,869 INFO [train.py:898] (0/4) Epoch 8, batch 2700, loss[loss=0.2135, simple_loss=0.3008, pruned_loss=0.06313, over 18503.00 frames. ], tot_loss[loss=0.2045, simple_loss=0.2866, pruned_loss=0.06123, over 3605249.88 frames. ], batch size: 53, lr: 1.40e-02, grad_scale: 4.0 +2023-03-08 22:53:41,235 INFO [train.py:898] (0/4) Epoch 8, batch 2750, loss[loss=0.226, simple_loss=0.3042, pruned_loss=0.07389, over 16971.00 frames. ], tot_loss[loss=0.2045, simple_loss=0.2864, pruned_loss=0.06131, over 3593267.75 frames. ], batch size: 78, lr: 1.40e-02, grad_scale: 4.0 +2023-03-08 22:54:00,758 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.43 vs. limit=5.0 +2023-03-08 22:54:29,087 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28229.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:54:31,573 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-08 22:54:40,880 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.331e+02 3.538e+02 4.348e+02 5.069e+02 1.374e+03, threshold=8.696e+02, percent-clipped=5.0 +2023-03-08 22:54:40,916 INFO [train.py:898] (0/4) Epoch 8, batch 2800, loss[loss=0.1906, simple_loss=0.2689, pruned_loss=0.05617, over 18368.00 frames. ], tot_loss[loss=0.2035, simple_loss=0.2857, pruned_loss=0.06066, over 3594390.34 frames. ], batch size: 46, lr: 1.40e-02, grad_scale: 8.0 +2023-03-08 22:54:59,981 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28255.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 22:55:18,157 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7642, 4.7453, 4.9078, 4.5574, 4.7294, 4.6646, 5.0654, 5.0081], + device='cuda:0'), covar=tensor([0.0072, 0.0074, 0.0092, 0.0112, 0.0059, 0.0131, 0.0074, 0.0093], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0054, 0.0055, 0.0070, 0.0057, 0.0080, 0.0067, 0.0067], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0002, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 22:55:37,508 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28287.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:55:39,233 INFO [train.py:898] (0/4) Epoch 8, batch 2850, loss[loss=0.2219, simple_loss=0.3111, pruned_loss=0.06631, over 18302.00 frames. ], tot_loss[loss=0.2031, simple_loss=0.2855, pruned_loss=0.06038, over 3600244.12 frames. ], batch size: 54, lr: 1.39e-02, grad_scale: 8.0 +2023-03-08 22:55:40,828 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28290.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:55:55,447 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2026, 4.2674, 2.8108, 4.3898, 5.2568, 2.6230, 3.8890, 3.9067], + device='cuda:0'), covar=tensor([0.0087, 0.1016, 0.1363, 0.0458, 0.0040, 0.1245, 0.0575, 0.0684], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0203, 0.0180, 0.0179, 0.0079, 0.0169, 0.0191, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0003, 0.0002, 0.0002, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 22:56:11,223 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28316.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 22:56:24,251 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-08 22:56:26,642 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-08 22:56:38,153 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.182e+02 3.670e+02 4.425e+02 5.374e+02 1.143e+03, threshold=8.851e+02, percent-clipped=3.0 +2023-03-08 22:56:38,178 INFO [train.py:898] (0/4) Epoch 8, batch 2900, loss[loss=0.2041, simple_loss=0.2658, pruned_loss=0.07117, over 18498.00 frames. ], tot_loss[loss=0.2029, simple_loss=0.2852, pruned_loss=0.06035, over 3599629.67 frames. ], batch size: 43, lr: 1.39e-02, grad_scale: 8.0 +2023-03-08 22:56:49,647 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28348.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:57:14,575 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28370.0, num_to_drop=1, layers_to_drop={2} +2023-03-08 22:57:36,803 INFO [train.py:898] (0/4) Epoch 8, batch 2950, loss[loss=0.2135, simple_loss=0.3008, pruned_loss=0.06308, over 17005.00 frames. ], tot_loss[loss=0.2035, simple_loss=0.2855, pruned_loss=0.06078, over 3586416.52 frames. ], batch size: 78, lr: 1.39e-02, grad_scale: 8.0 +2023-03-08 22:58:08,067 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.00 vs. limit=5.0 +2023-03-08 22:58:11,175 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=28418.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 22:58:13,424 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5777, 5.3068, 5.2344, 5.1177, 4.7707, 5.1747, 4.4136, 5.1092], + device='cuda:0'), covar=tensor([0.0285, 0.0245, 0.0226, 0.0312, 0.0404, 0.0228, 0.1420, 0.0313], + device='cuda:0'), in_proj_covar=tensor([0.0158, 0.0200, 0.0187, 0.0211, 0.0200, 0.0208, 0.0271, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0006, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-08 22:58:19,642 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7292, 4.7646, 4.8520, 4.5350, 4.6111, 4.4994, 5.0504, 4.9712], + device='cuda:0'), covar=tensor([0.0070, 0.0070, 0.0078, 0.0109, 0.0065, 0.0154, 0.0086, 0.0096], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0055, 0.0056, 0.0071, 0.0058, 0.0081, 0.0068, 0.0068], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 22:58:36,008 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.441e+02 3.351e+02 4.042e+02 5.395e+02 3.528e+03, threshold=8.084e+02, percent-clipped=8.0 +2023-03-08 22:58:36,050 INFO [train.py:898] (0/4) Epoch 8, batch 3000, loss[loss=0.1706, simple_loss=0.2573, pruned_loss=0.04201, over 18518.00 frames. ], tot_loss[loss=0.2038, simple_loss=0.2857, pruned_loss=0.06097, over 3587378.78 frames. ], batch size: 47, lr: 1.39e-02, grad_scale: 8.0 +2023-03-08 22:58:36,053 INFO [train.py:923] (0/4) Computing validation loss +2023-03-08 22:58:47,832 INFO [train.py:932] (0/4) Epoch 8, validation: loss=0.165, simple_loss=0.2676, pruned_loss=0.03118, over 944034.00 frames. +2023-03-08 22:58:47,833 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-08 22:59:42,661 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28486.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 22:59:46,137 INFO [train.py:898] (0/4) Epoch 8, batch 3050, loss[loss=0.2065, simple_loss=0.2855, pruned_loss=0.06377, over 18417.00 frames. ], tot_loss[loss=0.2049, simple_loss=0.2865, pruned_loss=0.06163, over 3584384.87 frames. ], batch size: 48, lr: 1.39e-02, grad_scale: 8.0 +2023-03-08 22:59:49,927 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28492.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:00:44,064 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.579e+02 3.667e+02 4.403e+02 5.909e+02 1.221e+03, threshold=8.806e+02, percent-clipped=6.0 +2023-03-08 23:00:44,107 INFO [train.py:898] (0/4) Epoch 8, batch 3100, loss[loss=0.2031, simple_loss=0.2906, pruned_loss=0.05775, over 18572.00 frames. ], tot_loss[loss=0.2049, simple_loss=0.2867, pruned_loss=0.06158, over 3578032.33 frames. ], batch size: 54, lr: 1.39e-02, grad_scale: 8.0 +2023-03-08 23:00:54,158 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28547.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:01:01,055 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28553.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:01:39,164 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28585.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:01:43,473 INFO [train.py:898] (0/4) Epoch 8, batch 3150, loss[loss=0.2089, simple_loss=0.2923, pruned_loss=0.06272, over 17219.00 frames. ], tot_loss[loss=0.2045, simple_loss=0.2862, pruned_loss=0.06141, over 3571335.11 frames. ], batch size: 78, lr: 1.39e-02, grad_scale: 8.0 +2023-03-08 23:02:10,510 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28611.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 23:02:43,327 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.977e+02 3.448e+02 4.056e+02 5.168e+02 1.166e+03, threshold=8.112e+02, percent-clipped=4.0 +2023-03-08 23:02:43,366 INFO [train.py:898] (0/4) Epoch 8, batch 3200, loss[loss=0.1923, simple_loss=0.2762, pruned_loss=0.05418, over 16084.00 frames. ], tot_loss[loss=0.2043, simple_loss=0.286, pruned_loss=0.06129, over 3570956.14 frames. ], batch size: 94, lr: 1.39e-02, grad_scale: 8.0 +2023-03-08 23:02:48,226 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28643.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:02:48,555 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1865, 4.1893, 5.2731, 3.5260, 4.5294, 2.8646, 2.9699, 1.9809], + device='cuda:0'), covar=tensor([0.0727, 0.0564, 0.0057, 0.0478, 0.0459, 0.1967, 0.2581, 0.1641], + device='cuda:0'), in_proj_covar=tensor([0.0186, 0.0199, 0.0100, 0.0155, 0.0216, 0.0236, 0.0257, 0.0198], + device='cuda:0'), out_proj_covar=tensor([1.6628e-04, 1.8376e-04, 9.3604e-05, 1.4180e-04, 2.0040e-04, 2.2035e-04, + 2.3695e-04, 1.8562e-04], device='cuda:0') +2023-03-08 23:03:23,647 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4682, 3.3134, 1.8690, 4.2328, 2.7460, 4.2188, 2.0577, 3.4800], + device='cuda:0'), covar=tensor([0.0517, 0.0742, 0.1486, 0.0323, 0.0946, 0.0261, 0.1260, 0.0427], + device='cuda:0'), in_proj_covar=tensor([0.0173, 0.0201, 0.0173, 0.0209, 0.0172, 0.0208, 0.0184, 0.0174], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 23:03:27,498 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-08 23:03:42,535 INFO [train.py:898] (0/4) Epoch 8, batch 3250, loss[loss=0.2332, simple_loss=0.3211, pruned_loss=0.07268, over 16987.00 frames. ], tot_loss[loss=0.204, simple_loss=0.2852, pruned_loss=0.06139, over 3576790.33 frames. ], batch size: 78, lr: 1.39e-02, grad_scale: 4.0 +2023-03-08 23:03:48,062 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-03-08 23:04:14,197 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.65 vs. limit=5.0 +2023-03-08 23:04:42,243 INFO [train.py:898] (0/4) Epoch 8, batch 3300, loss[loss=0.2154, simple_loss=0.3078, pruned_loss=0.06154, over 18361.00 frames. ], tot_loss[loss=0.2042, simple_loss=0.2855, pruned_loss=0.06147, over 3589236.58 frames. ], batch size: 56, lr: 1.38e-02, grad_scale: 4.0 +2023-03-08 23:04:43,380 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.421e+02 3.488e+02 4.190e+02 5.293e+02 2.938e+03, threshold=8.380e+02, percent-clipped=5.0 +2023-03-08 23:05:41,146 INFO [train.py:898] (0/4) Epoch 8, batch 3350, loss[loss=0.2037, simple_loss=0.2943, pruned_loss=0.05655, over 18565.00 frames. ], tot_loss[loss=0.2039, simple_loss=0.2854, pruned_loss=0.0612, over 3597473.11 frames. ], batch size: 54, lr: 1.38e-02, grad_scale: 4.0 +2023-03-08 23:05:51,727 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5970, 2.7709, 3.9802, 3.6069, 2.7069, 4.5045, 3.8847, 2.8618], + device='cuda:0'), covar=tensor([0.0467, 0.1377, 0.0285, 0.0290, 0.1394, 0.0163, 0.0552, 0.0980], + device='cuda:0'), in_proj_covar=tensor([0.0175, 0.0205, 0.0132, 0.0128, 0.0201, 0.0170, 0.0185, 0.0183], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 23:06:16,657 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.86 vs. limit=2.0 +2023-03-08 23:06:40,386 INFO [train.py:898] (0/4) Epoch 8, batch 3400, loss[loss=0.1991, simple_loss=0.2794, pruned_loss=0.05938, over 17678.00 frames. ], tot_loss[loss=0.2036, simple_loss=0.285, pruned_loss=0.06113, over 3594970.64 frames. ], batch size: 39, lr: 1.38e-02, grad_scale: 4.0 +2023-03-08 23:06:41,526 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.584e+02 3.993e+02 4.817e+02 5.813e+02 1.322e+03, threshold=9.634e+02, percent-clipped=7.0 +2023-03-08 23:06:44,142 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28842.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:06:47,912 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4366, 2.6957, 2.4998, 2.6540, 3.4707, 3.3863, 2.8941, 2.8934], + device='cuda:0'), covar=tensor([0.0138, 0.0268, 0.0614, 0.0363, 0.0164, 0.0159, 0.0342, 0.0317], + device='cuda:0'), in_proj_covar=tensor([0.0107, 0.0093, 0.0145, 0.0124, 0.0091, 0.0074, 0.0119, 0.0117], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0001, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 23:06:50,918 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=28848.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:07:07,352 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=28861.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:07:35,106 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28885.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:07:36,590 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-08 23:07:39,399 INFO [train.py:898] (0/4) Epoch 8, batch 3450, loss[loss=0.1681, simple_loss=0.2544, pruned_loss=0.04097, over 18258.00 frames. ], tot_loss[loss=0.2045, simple_loss=0.286, pruned_loss=0.06152, over 3577647.91 frames. ], batch size: 45, lr: 1.38e-02, grad_scale: 4.0 +2023-03-08 23:08:05,070 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8779, 3.1077, 4.3574, 3.9849, 2.8143, 4.6961, 4.2005, 2.9550], + device='cuda:0'), covar=tensor([0.0325, 0.1082, 0.0165, 0.0232, 0.1223, 0.0119, 0.0326, 0.0850], + device='cuda:0'), in_proj_covar=tensor([0.0173, 0.0204, 0.0132, 0.0126, 0.0202, 0.0169, 0.0183, 0.0182], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 23:08:06,020 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28911.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 23:08:13,375 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-08 23:08:18,886 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=28922.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:08:31,075 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=28933.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:08:38,988 INFO [train.py:898] (0/4) Epoch 8, batch 3500, loss[loss=0.2064, simple_loss=0.2938, pruned_loss=0.05954, over 18408.00 frames. ], tot_loss[loss=0.2047, simple_loss=0.2863, pruned_loss=0.0615, over 3588171.76 frames. ], batch size: 52, lr: 1.38e-02, grad_scale: 4.0 +2023-03-08 23:08:40,151 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.088e+02 3.743e+02 4.474e+02 5.691e+02 1.966e+03, threshold=8.949e+02, percent-clipped=4.0 +2023-03-08 23:08:43,914 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=28943.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:09:02,281 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=28959.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 23:09:31,437 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.4990, 4.5280, 4.5545, 4.3108, 4.3778, 4.4061, 4.7477, 4.6345], + device='cuda:0'), covar=tensor([0.0066, 0.0066, 0.0073, 0.0090, 0.0063, 0.0123, 0.0061, 0.0085], + device='cuda:0'), in_proj_covar=tensor([0.0073, 0.0054, 0.0055, 0.0068, 0.0058, 0.0080, 0.0067, 0.0067], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0002, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 23:09:35,649 INFO [train.py:898] (0/4) Epoch 8, batch 3550, loss[loss=0.1815, simple_loss=0.2593, pruned_loss=0.05179, over 18402.00 frames. ], tot_loss[loss=0.2051, simple_loss=0.2867, pruned_loss=0.06175, over 3578060.23 frames. ], batch size: 42, lr: 1.38e-02, grad_scale: 4.0 +2023-03-08 23:09:37,859 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=28991.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:09:50,619 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29002.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:10:20,203 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.94 vs. limit=5.0 +2023-03-08 23:10:30,399 INFO [train.py:898] (0/4) Epoch 8, batch 3600, loss[loss=0.2025, simple_loss=0.2905, pruned_loss=0.05727, over 18505.00 frames. ], tot_loss[loss=0.206, simple_loss=0.2873, pruned_loss=0.06234, over 3551524.18 frames. ], batch size: 51, lr: 1.38e-02, grad_scale: 8.0 +2023-03-08 23:10:31,433 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.295e+02 3.435e+02 4.194e+02 5.068e+02 1.055e+03, threshold=8.389e+02, percent-clipped=1.0 +2023-03-08 23:10:32,556 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4660, 6.0382, 5.5130, 5.7920, 5.4956, 5.5432, 6.0469, 5.9814], + device='cuda:0'), covar=tensor([0.1110, 0.0693, 0.0355, 0.0630, 0.1466, 0.0655, 0.0591, 0.0735], + device='cuda:0'), in_proj_covar=tensor([0.0468, 0.0380, 0.0301, 0.0424, 0.0583, 0.0423, 0.0532, 0.0414], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-08 23:10:56,775 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29063.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:11:06,612 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-8.pt +2023-03-08 23:11:37,008 INFO [train.py:898] (0/4) Epoch 9, batch 0, loss[loss=0.2408, simple_loss=0.3198, pruned_loss=0.08087, over 18500.00 frames. ], tot_loss[loss=0.2408, simple_loss=0.3198, pruned_loss=0.08087, over 18500.00 frames. ], batch size: 59, lr: 1.30e-02, grad_scale: 8.0 +2023-03-08 23:11:37,010 INFO [train.py:923] (0/4) Computing validation loss +2023-03-08 23:11:48,957 INFO [train.py:932] (0/4) Epoch 9, validation: loss=0.1674, simple_loss=0.2698, pruned_loss=0.03254, over 944034.00 frames. +2023-03-08 23:11:48,958 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-08 23:12:38,126 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3852, 4.8975, 4.8838, 4.8489, 4.5382, 4.7623, 4.2647, 4.7681], + device='cuda:0'), covar=tensor([0.0202, 0.0270, 0.0198, 0.0302, 0.0276, 0.0226, 0.0956, 0.0260], + device='cuda:0'), in_proj_covar=tensor([0.0156, 0.0197, 0.0187, 0.0213, 0.0197, 0.0207, 0.0264, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0006, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-08 23:12:44,131 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1625, 2.5809, 2.3187, 2.4639, 3.2104, 3.0354, 2.6954, 2.6207], + device='cuda:0'), covar=tensor([0.0161, 0.0281, 0.0626, 0.0404, 0.0179, 0.0168, 0.0365, 0.0361], + device='cuda:0'), in_proj_covar=tensor([0.0106, 0.0093, 0.0146, 0.0125, 0.0091, 0.0075, 0.0120, 0.0115], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-08 23:12:48,238 INFO [train.py:898] (0/4) Epoch 9, batch 50, loss[loss=0.2166, simple_loss=0.2997, pruned_loss=0.06675, over 18364.00 frames. ], tot_loss[loss=0.2065, simple_loss=0.2891, pruned_loss=0.0619, over 819269.11 frames. ], batch size: 55, lr: 1.30e-02, grad_scale: 8.0 +2023-03-08 23:13:08,328 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.388e+02 3.677e+02 4.281e+02 4.949e+02 1.360e+03, threshold=8.563e+02, percent-clipped=6.0 +2023-03-08 23:13:10,856 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29142.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:13:18,810 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29148.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:13:25,701 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5210, 3.3395, 2.1443, 4.2147, 2.7823, 4.3366, 2.3671, 3.9017], + device='cuda:0'), covar=tensor([0.0503, 0.0770, 0.1319, 0.0500, 0.0933, 0.0272, 0.1110, 0.0340], + device='cuda:0'), in_proj_covar=tensor([0.0174, 0.0202, 0.0172, 0.0210, 0.0170, 0.0207, 0.0185, 0.0174], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 23:13:47,210 INFO [train.py:898] (0/4) Epoch 9, batch 100, loss[loss=0.1756, simple_loss=0.2545, pruned_loss=0.04838, over 18178.00 frames. ], tot_loss[loss=0.2015, simple_loss=0.283, pruned_loss=0.06, over 1448923.10 frames. ], batch size: 44, lr: 1.30e-02, grad_scale: 8.0 +2023-03-08 23:14:07,588 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=29190.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:14:14,295 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=29196.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:14:36,134 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9816, 4.1174, 2.2623, 4.1841, 5.0580, 2.2949, 3.7456, 3.7122], + device='cuda:0'), covar=tensor([0.0076, 0.1131, 0.1599, 0.0496, 0.0050, 0.1386, 0.0634, 0.0761], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0209, 0.0184, 0.0184, 0.0081, 0.0168, 0.0198, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0003, 0.0002, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 23:14:39,387 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29217.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:14:46,033 INFO [train.py:898] (0/4) Epoch 9, batch 150, loss[loss=0.1998, simple_loss=0.2882, pruned_loss=0.05564, over 18281.00 frames. ], tot_loss[loss=0.2013, simple_loss=0.2839, pruned_loss=0.05932, over 1933885.68 frames. ], batch size: 57, lr: 1.30e-02, grad_scale: 8.0 +2023-03-08 23:14:56,826 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-08 23:15:00,920 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29236.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:15:05,100 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.617e+02 3.595e+02 4.206e+02 4.990e+02 1.116e+03, threshold=8.412e+02, percent-clipped=1.0 +2023-03-08 23:15:44,251 INFO [train.py:898] (0/4) Epoch 9, batch 200, loss[loss=0.1931, simple_loss=0.2796, pruned_loss=0.05328, over 18286.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.283, pruned_loss=0.05944, over 2302073.62 frames. ], batch size: 49, lr: 1.30e-02, grad_scale: 8.0 +2023-03-08 23:15:52,818 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3298, 5.1899, 5.4059, 5.3394, 5.2124, 6.0024, 5.7346, 5.4707], + device='cuda:0'), covar=tensor([0.0835, 0.0607, 0.0641, 0.0674, 0.1361, 0.0733, 0.0591, 0.1504], + device='cuda:0'), in_proj_covar=tensor([0.0283, 0.0218, 0.0225, 0.0224, 0.0268, 0.0327, 0.0211, 0.0310], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 23:16:13,773 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29297.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:16:44,290 INFO [train.py:898] (0/4) Epoch 9, batch 250, loss[loss=0.1918, simple_loss=0.2708, pruned_loss=0.05644, over 18411.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.2825, pruned_loss=0.05871, over 2595498.71 frames. ], batch size: 48, lr: 1.30e-02, grad_scale: 8.0 +2023-03-08 23:17:03,791 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.373e+02 3.539e+02 4.403e+02 5.304e+02 1.212e+03, threshold=8.805e+02, percent-clipped=3.0 +2023-03-08 23:17:19,005 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6491, 3.6063, 5.2940, 4.4346, 3.3690, 2.8669, 4.8268, 5.3504], + device='cuda:0'), covar=tensor([0.0800, 0.1760, 0.0076, 0.0289, 0.0757, 0.1110, 0.0214, 0.0163], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0222, 0.0085, 0.0148, 0.0166, 0.0168, 0.0157, 0.0120], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 23:17:25,526 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29358.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:17:36,883 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6455, 4.0305, 5.4681, 4.5992, 2.9242, 2.5412, 4.6257, 5.5524], + device='cuda:0'), covar=tensor([0.0769, 0.1378, 0.0054, 0.0250, 0.0902, 0.1234, 0.0300, 0.0064], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0223, 0.0085, 0.0149, 0.0167, 0.0169, 0.0158, 0.0121], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 23:17:43,325 INFO [train.py:898] (0/4) Epoch 9, batch 300, loss[loss=0.1976, simple_loss=0.283, pruned_loss=0.0561, over 18567.00 frames. ], tot_loss[loss=0.2005, simple_loss=0.2827, pruned_loss=0.05912, over 2813911.75 frames. ], batch size: 54, lr: 1.30e-02, grad_scale: 8.0 +2023-03-08 23:17:55,303 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3246, 5.4316, 2.7606, 5.2153, 5.0282, 5.4901, 5.2893, 2.6464], + device='cuda:0'), covar=tensor([0.0135, 0.0058, 0.0705, 0.0058, 0.0068, 0.0047, 0.0070, 0.0950], + device='cuda:0'), in_proj_covar=tensor([0.0070, 0.0060, 0.0086, 0.0074, 0.0071, 0.0058, 0.0070, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0005, 0.0004, 0.0004, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 23:18:42,975 INFO [train.py:898] (0/4) Epoch 9, batch 350, loss[loss=0.225, simple_loss=0.3107, pruned_loss=0.06961, over 17728.00 frames. ], tot_loss[loss=0.201, simple_loss=0.2834, pruned_loss=0.05934, over 2988946.64 frames. ], batch size: 70, lr: 1.30e-02, grad_scale: 8.0 +2023-03-08 23:19:02,289 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.167e+02 3.494e+02 4.009e+02 5.079e+02 1.236e+03, threshold=8.018e+02, percent-clipped=2.0 +2023-03-08 23:19:24,890 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.12 vs. limit=5.0 +2023-03-08 23:19:41,978 INFO [train.py:898] (0/4) Epoch 9, batch 400, loss[loss=0.2241, simple_loss=0.3092, pruned_loss=0.06948, over 18038.00 frames. ], tot_loss[loss=0.2001, simple_loss=0.2826, pruned_loss=0.05878, over 3133852.67 frames. ], batch size: 62, lr: 1.29e-02, grad_scale: 8.0 +2023-03-08 23:20:34,025 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29517.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:20:40,637 INFO [train.py:898] (0/4) Epoch 9, batch 450, loss[loss=0.1884, simple_loss=0.2661, pruned_loss=0.0553, over 18416.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.2834, pruned_loss=0.05924, over 3230478.04 frames. ], batch size: 48, lr: 1.29e-02, grad_scale: 8.0 +2023-03-08 23:20:59,766 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.081e+02 3.569e+02 4.141e+02 5.253e+02 9.990e+02, threshold=8.283e+02, percent-clipped=5.0 +2023-03-08 23:21:12,947 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29551.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:21:29,964 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=29565.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:21:40,050 INFO [train.py:898] (0/4) Epoch 9, batch 500, loss[loss=0.1535, simple_loss=0.2332, pruned_loss=0.03687, over 18396.00 frames. ], tot_loss[loss=0.2005, simple_loss=0.283, pruned_loss=0.059, over 3307683.06 frames. ], batch size: 42, lr: 1.29e-02, grad_scale: 8.0 +2023-03-08 23:22:01,696 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29592.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:22:25,393 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29612.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:22:38,335 INFO [train.py:898] (0/4) Epoch 9, batch 550, loss[loss=0.1861, simple_loss=0.2706, pruned_loss=0.05077, over 18280.00 frames. ], tot_loss[loss=0.2008, simple_loss=0.2834, pruned_loss=0.05911, over 3374040.93 frames. ], batch size: 49, lr: 1.29e-02, grad_scale: 8.0 +2023-03-08 23:22:58,658 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.384e+02 3.638e+02 4.636e+02 5.710e+02 1.392e+03, threshold=9.272e+02, percent-clipped=6.0 +2023-03-08 23:23:06,223 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4847, 2.6651, 2.3456, 2.5276, 3.3433, 3.3468, 2.9208, 2.6219], + device='cuda:0'), covar=tensor([0.0145, 0.0326, 0.0754, 0.0406, 0.0216, 0.0167, 0.0393, 0.0460], + device='cuda:0'), in_proj_covar=tensor([0.0109, 0.0093, 0.0145, 0.0125, 0.0094, 0.0077, 0.0121, 0.0119], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 23:23:19,627 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29658.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:23:24,131 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29662.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:23:37,407 INFO [train.py:898] (0/4) Epoch 9, batch 600, loss[loss=0.2083, simple_loss=0.294, pruned_loss=0.0613, over 17299.00 frames. ], tot_loss[loss=0.2013, simple_loss=0.2837, pruned_loss=0.05945, over 3421133.08 frames. ], batch size: 78, lr: 1.29e-02, grad_scale: 8.0 +2023-03-08 23:23:50,452 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.23 vs. limit=5.0 +2023-03-08 23:23:59,523 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9224, 4.6509, 4.8835, 3.6487, 3.8106, 3.6974, 2.8747, 2.4318], + device='cuda:0'), covar=tensor([0.0231, 0.0168, 0.0057, 0.0236, 0.0295, 0.0205, 0.0645, 0.0911], + device='cuda:0'), in_proj_covar=tensor([0.0052, 0.0044, 0.0042, 0.0055, 0.0074, 0.0051, 0.0068, 0.0075], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0005, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 23:24:13,370 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3230, 5.1320, 5.4861, 5.3779, 5.1775, 6.0534, 5.6993, 5.4654], + device='cuda:0'), covar=tensor([0.0907, 0.0597, 0.0583, 0.0612, 0.1369, 0.0664, 0.0643, 0.1543], + device='cuda:0'), in_proj_covar=tensor([0.0291, 0.0220, 0.0229, 0.0226, 0.0269, 0.0326, 0.0218, 0.0318], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 23:24:16,689 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=29706.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:24:18,945 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3677, 5.3806, 2.7896, 5.2435, 5.0305, 5.3962, 5.1588, 2.5924], + device='cuda:0'), covar=tensor([0.0126, 0.0088, 0.0803, 0.0074, 0.0090, 0.0104, 0.0134, 0.1127], + device='cuda:0'), in_proj_covar=tensor([0.0072, 0.0062, 0.0088, 0.0076, 0.0072, 0.0060, 0.0073, 0.0090], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0005, 0.0004, 0.0004, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 23:24:36,229 INFO [train.py:898] (0/4) Epoch 9, batch 650, loss[loss=0.2149, simple_loss=0.2979, pruned_loss=0.06592, over 17764.00 frames. ], tot_loss[loss=0.2016, simple_loss=0.2841, pruned_loss=0.05957, over 3454272.01 frames. ], batch size: 70, lr: 1.29e-02, grad_scale: 8.0 +2023-03-08 23:24:36,671 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29723.0, num_to_drop=1, layers_to_drop={1} +2023-03-08 23:24:57,201 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.211e+02 3.283e+02 4.124e+02 5.101e+02 2.453e+03, threshold=8.247e+02, percent-clipped=7.0 +2023-03-08 23:25:35,095 INFO [train.py:898] (0/4) Epoch 9, batch 700, loss[loss=0.2185, simple_loss=0.292, pruned_loss=0.07244, over 18383.00 frames. ], tot_loss[loss=0.2019, simple_loss=0.2842, pruned_loss=0.05984, over 3479976.28 frames. ], batch size: 50, lr: 1.29e-02, grad_scale: 8.0 +2023-03-08 23:25:43,327 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-08 23:26:10,097 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=29802.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:26:14,083 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.71 vs. limit=2.0 +2023-03-08 23:26:34,261 INFO [train.py:898] (0/4) Epoch 9, batch 750, loss[loss=0.2016, simple_loss=0.2967, pruned_loss=0.05328, over 18632.00 frames. ], tot_loss[loss=0.2008, simple_loss=0.2831, pruned_loss=0.05932, over 3506741.91 frames. ], batch size: 52, lr: 1.29e-02, grad_scale: 8.0 +2023-03-08 23:26:55,145 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.192e+02 3.216e+02 4.021e+02 4.703e+02 9.794e+02, threshold=8.041e+02, percent-clipped=2.0 +2023-03-08 23:27:21,832 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=29863.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:27:33,005 INFO [train.py:898] (0/4) Epoch 9, batch 800, loss[loss=0.1907, simple_loss=0.2722, pruned_loss=0.05465, over 18412.00 frames. ], tot_loss[loss=0.2004, simple_loss=0.283, pruned_loss=0.05889, over 3539242.12 frames. ], batch size: 48, lr: 1.29e-02, grad_scale: 8.0 +2023-03-08 23:27:56,478 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=29892.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:28:13,930 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=29907.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:28:32,031 INFO [train.py:898] (0/4) Epoch 9, batch 850, loss[loss=0.1909, simple_loss=0.2824, pruned_loss=0.04973, over 18367.00 frames. ], tot_loss[loss=0.2003, simple_loss=0.2831, pruned_loss=0.05877, over 3557068.13 frames. ], batch size: 55, lr: 1.29e-02, grad_scale: 8.0 +2023-03-08 23:28:40,945 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-08 23:28:52,841 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.076e+02 3.769e+02 4.369e+02 5.199e+02 8.545e+02, threshold=8.737e+02, percent-clipped=2.0 +2023-03-08 23:28:53,031 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=29940.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:29:13,481 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2695, 4.7411, 4.3975, 4.5629, 4.3021, 4.3468, 4.8112, 4.7429], + device='cuda:0'), covar=tensor([0.1094, 0.0703, 0.1657, 0.0768, 0.1535, 0.0773, 0.0651, 0.0696], + device='cuda:0'), in_proj_covar=tensor([0.0487, 0.0391, 0.0303, 0.0431, 0.0603, 0.0431, 0.0554, 0.0424], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-08 23:29:26,222 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8463, 4.8825, 5.0214, 5.0021, 4.7958, 5.5561, 5.2161, 4.8813], + device='cuda:0'), covar=tensor([0.0998, 0.0771, 0.0659, 0.0598, 0.1352, 0.0729, 0.0616, 0.1566], + device='cuda:0'), in_proj_covar=tensor([0.0286, 0.0219, 0.0226, 0.0224, 0.0268, 0.0324, 0.0213, 0.0312], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 23:29:31,723 INFO [train.py:898] (0/4) Epoch 9, batch 900, loss[loss=0.2762, simple_loss=0.3371, pruned_loss=0.1077, over 12374.00 frames. ], tot_loss[loss=0.201, simple_loss=0.2837, pruned_loss=0.05912, over 3549697.55 frames. ], batch size: 130, lr: 1.28e-02, grad_scale: 8.0 +2023-03-08 23:30:05,146 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-30000.pt +2023-03-08 23:30:30,151 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30018.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 23:30:31,555 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5174, 3.5935, 5.0495, 4.1497, 2.9757, 2.7909, 4.0994, 5.0490], + device='cuda:0'), covar=tensor([0.0894, 0.1463, 0.0072, 0.0341, 0.0962, 0.1223, 0.0397, 0.0139], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0224, 0.0087, 0.0149, 0.0168, 0.0170, 0.0157, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 23:30:35,519 INFO [train.py:898] (0/4) Epoch 9, batch 950, loss[loss=0.2037, simple_loss=0.2866, pruned_loss=0.06042, over 18239.00 frames. ], tot_loss[loss=0.2003, simple_loss=0.2832, pruned_loss=0.0587, over 3569760.07 frames. ], batch size: 60, lr: 1.28e-02, grad_scale: 8.0 +2023-03-08 23:30:56,116 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.525e+02 3.421e+02 3.887e+02 4.758e+02 7.533e+02, threshold=7.773e+02, percent-clipped=0.0 +2023-03-08 23:31:08,719 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7053, 4.1095, 4.0765, 4.0964, 3.8322, 4.0080, 3.6525, 4.0245], + device='cuda:0'), covar=tensor([0.0284, 0.0354, 0.0284, 0.0454, 0.0360, 0.0265, 0.1092, 0.0340], + device='cuda:0'), in_proj_covar=tensor([0.0158, 0.0201, 0.0189, 0.0215, 0.0200, 0.0209, 0.0271, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0006, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-08 23:31:34,973 INFO [train.py:898] (0/4) Epoch 9, batch 1000, loss[loss=0.2053, simple_loss=0.2858, pruned_loss=0.06238, over 18354.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.2823, pruned_loss=0.0587, over 3564189.19 frames. ], batch size: 56, lr: 1.28e-02, grad_scale: 8.0 +2023-03-08 23:32:19,880 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-08 23:32:34,164 INFO [train.py:898] (0/4) Epoch 9, batch 1050, loss[loss=0.1906, simple_loss=0.2771, pruned_loss=0.0521, over 18263.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.2824, pruned_loss=0.0587, over 3570182.82 frames. ], batch size: 47, lr: 1.28e-02, grad_scale: 8.0 +2023-03-08 23:32:36,846 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9442, 4.7003, 2.4593, 4.5109, 4.4838, 4.7307, 4.5481, 2.6565], + device='cuda:0'), covar=tensor([0.0151, 0.0055, 0.0773, 0.0098, 0.0069, 0.0069, 0.0105, 0.0902], + device='cuda:0'), in_proj_covar=tensor([0.0072, 0.0061, 0.0086, 0.0076, 0.0071, 0.0058, 0.0073, 0.0090], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0005, 0.0004, 0.0004, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 23:32:36,890 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30125.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:32:53,968 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.406e+02 3.588e+02 4.236e+02 5.310e+02 1.075e+03, threshold=8.473e+02, percent-clipped=3.0 +2023-03-08 23:33:15,627 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30158.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:33:32,457 INFO [train.py:898] (0/4) Epoch 9, batch 1100, loss[loss=0.156, simple_loss=0.2363, pruned_loss=0.03783, over 17624.00 frames. ], tot_loss[loss=0.2009, simple_loss=0.283, pruned_loss=0.0594, over 3558663.96 frames. ], batch size: 39, lr: 1.28e-02, grad_scale: 8.0 +2023-03-08 23:33:34,312 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-08 23:33:37,812 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-03-08 23:33:40,033 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-03-08 23:33:47,841 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30186.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:34:12,781 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30207.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:34:15,619 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30209.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:34:31,360 INFO [train.py:898] (0/4) Epoch 9, batch 1150, loss[loss=0.1901, simple_loss=0.268, pruned_loss=0.05615, over 18495.00 frames. ], tot_loss[loss=0.2018, simple_loss=0.2839, pruned_loss=0.05986, over 3558959.67 frames. ], batch size: 47, lr: 1.28e-02, grad_scale: 8.0 +2023-03-08 23:34:50,756 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.527e+02 3.691e+02 4.407e+02 5.401e+02 1.436e+03, threshold=8.814e+02, percent-clipped=4.0 +2023-03-08 23:35:09,065 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=30255.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:35:27,456 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30270.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:35:30,583 INFO [train.py:898] (0/4) Epoch 9, batch 1200, loss[loss=0.1869, simple_loss=0.2711, pruned_loss=0.05138, over 18552.00 frames. ], tot_loss[loss=0.202, simple_loss=0.2842, pruned_loss=0.05993, over 3554730.88 frames. ], batch size: 49, lr: 1.28e-02, grad_scale: 8.0 +2023-03-08 23:36:24,725 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30318.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 23:36:30,184 INFO [train.py:898] (0/4) Epoch 9, batch 1250, loss[loss=0.1696, simple_loss=0.2455, pruned_loss=0.0469, over 18416.00 frames. ], tot_loss[loss=0.2011, simple_loss=0.2832, pruned_loss=0.05951, over 3558931.33 frames. ], batch size: 42, lr: 1.28e-02, grad_scale: 8.0 +2023-03-08 23:36:49,689 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.180e+02 3.477e+02 4.151e+02 5.169e+02 1.110e+03, threshold=8.302e+02, percent-clipped=2.0 +2023-03-08 23:37:09,847 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.29 vs. limit=5.0 +2023-03-08 23:37:14,222 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30360.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:37:21,618 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=30366.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:37:29,899 INFO [train.py:898] (0/4) Epoch 9, batch 1300, loss[loss=0.2151, simple_loss=0.2957, pruned_loss=0.06722, over 17883.00 frames. ], tot_loss[loss=0.2012, simple_loss=0.2834, pruned_loss=0.05951, over 3564612.85 frames. ], batch size: 70, lr: 1.28e-02, grad_scale: 8.0 +2023-03-08 23:38:26,508 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30421.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:38:28,946 INFO [train.py:898] (0/4) Epoch 9, batch 1350, loss[loss=0.1951, simple_loss=0.2855, pruned_loss=0.05239, over 18312.00 frames. ], tot_loss[loss=0.2012, simple_loss=0.2836, pruned_loss=0.0594, over 3561043.96 frames. ], batch size: 54, lr: 1.27e-02, grad_scale: 8.0 +2023-03-08 23:38:48,275 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.096e+02 3.366e+02 4.141e+02 5.097e+02 1.343e+03, threshold=8.282e+02, percent-clipped=5.0 +2023-03-08 23:39:09,355 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30458.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:39:27,650 INFO [train.py:898] (0/4) Epoch 9, batch 1400, loss[loss=0.2589, simple_loss=0.3269, pruned_loss=0.09546, over 12958.00 frames. ], tot_loss[loss=0.2013, simple_loss=0.2836, pruned_loss=0.05949, over 3564320.27 frames. ], batch size: 131, lr: 1.27e-02, grad_scale: 8.0 +2023-03-08 23:39:37,251 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30481.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:39:53,686 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8597, 3.4855, 4.7363, 3.0389, 3.8549, 2.5334, 2.8465, 2.0793], + device='cuda:0'), covar=tensor([0.0786, 0.0761, 0.0079, 0.0556, 0.0690, 0.2063, 0.2171, 0.1588], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0205, 0.0104, 0.0161, 0.0221, 0.0240, 0.0267, 0.0203], + device='cuda:0'), out_proj_covar=tensor([1.6806e-04, 1.8813e-04, 9.6127e-05, 1.4602e-04, 2.0261e-04, 2.2210e-04, + 2.4385e-04, 1.8955e-04], device='cuda:0') +2023-03-08 23:40:06,178 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=30506.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:40:26,572 INFO [train.py:898] (0/4) Epoch 9, batch 1450, loss[loss=0.1806, simple_loss=0.2548, pruned_loss=0.05317, over 18497.00 frames. ], tot_loss[loss=0.202, simple_loss=0.2848, pruned_loss=0.05963, over 3561755.24 frames. ], batch size: 44, lr: 1.27e-02, grad_scale: 8.0 +2023-03-08 23:40:46,480 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.569e+02 3.548e+02 4.179e+02 5.117e+02 1.123e+03, threshold=8.357e+02, percent-clipped=2.0 +2023-03-08 23:41:16,094 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30565.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:41:24,909 INFO [train.py:898] (0/4) Epoch 9, batch 1500, loss[loss=0.189, simple_loss=0.2823, pruned_loss=0.04783, over 18385.00 frames. ], tot_loss[loss=0.2028, simple_loss=0.2857, pruned_loss=0.05996, over 3569695.82 frames. ], batch size: 52, lr: 1.27e-02, grad_scale: 8.0 +2023-03-08 23:42:24,073 INFO [train.py:898] (0/4) Epoch 9, batch 1550, loss[loss=0.2219, simple_loss=0.3038, pruned_loss=0.07, over 15857.00 frames. ], tot_loss[loss=0.2022, simple_loss=0.2848, pruned_loss=0.05982, over 3580060.77 frames. ], batch size: 94, lr: 1.27e-02, grad_scale: 8.0 +2023-03-08 23:42:24,475 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0039, 3.3839, 3.4332, 2.8255, 3.0153, 2.8612, 2.4074, 2.0947], + device='cuda:0'), covar=tensor([0.0234, 0.0156, 0.0108, 0.0268, 0.0330, 0.0236, 0.0581, 0.0765], + device='cuda:0'), in_proj_covar=tensor([0.0054, 0.0044, 0.0044, 0.0056, 0.0076, 0.0052, 0.0069, 0.0076], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-08 23:42:44,354 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.188e+02 3.439e+02 3.957e+02 5.356e+02 1.144e+03, threshold=7.914e+02, percent-clipped=4.0 +2023-03-08 23:42:46,588 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-08 23:43:01,897 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30655.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:43:23,618 INFO [train.py:898] (0/4) Epoch 9, batch 1600, loss[loss=0.1844, simple_loss=0.2556, pruned_loss=0.05655, over 17747.00 frames. ], tot_loss[loss=0.201, simple_loss=0.2836, pruned_loss=0.05916, over 3575593.04 frames. ], batch size: 39, lr: 1.27e-02, grad_scale: 8.0 +2023-03-08 23:43:25,489 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-08 23:43:45,353 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-08 23:44:14,602 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=30716.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:44:14,862 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30716.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 23:44:22,849 INFO [train.py:898] (0/4) Epoch 9, batch 1650, loss[loss=0.1903, simple_loss=0.263, pruned_loss=0.05884, over 18490.00 frames. ], tot_loss[loss=0.2007, simple_loss=0.2831, pruned_loss=0.05915, over 3581939.62 frames. ], batch size: 44, lr: 1.27e-02, grad_scale: 16.0 +2023-03-08 23:44:24,163 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8321, 4.9827, 5.0116, 4.9271, 4.8489, 5.5690, 5.0774, 4.9522], + device='cuda:0'), covar=tensor([0.0925, 0.0669, 0.0689, 0.0691, 0.1190, 0.0681, 0.0617, 0.1604], + device='cuda:0'), in_proj_covar=tensor([0.0293, 0.0219, 0.0229, 0.0229, 0.0272, 0.0326, 0.0215, 0.0325], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 23:44:43,344 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.525e+02 3.575e+02 4.515e+02 5.590e+02 1.202e+03, threshold=9.030e+02, percent-clipped=6.0 +2023-03-08 23:45:04,578 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.3194, 1.9756, 1.9693, 2.0823, 2.4396, 2.4546, 2.3564, 2.2114], + device='cuda:0'), covar=tensor([0.0200, 0.0228, 0.0473, 0.0336, 0.0199, 0.0129, 0.0310, 0.0252], + device='cuda:0'), in_proj_covar=tensor([0.0111, 0.0095, 0.0143, 0.0125, 0.0092, 0.0078, 0.0124, 0.0118], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 23:45:22,488 INFO [train.py:898] (0/4) Epoch 9, batch 1700, loss[loss=0.1929, simple_loss=0.2855, pruned_loss=0.05017, over 18591.00 frames. ], tot_loss[loss=0.2011, simple_loss=0.2834, pruned_loss=0.05939, over 3574189.02 frames. ], batch size: 54, lr: 1.27e-02, grad_scale: 16.0 +2023-03-08 23:45:32,838 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30781.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:46:22,137 INFO [train.py:898] (0/4) Epoch 9, batch 1750, loss[loss=0.1975, simple_loss=0.281, pruned_loss=0.057, over 18275.00 frames. ], tot_loss[loss=0.2011, simple_loss=0.2833, pruned_loss=0.05947, over 3574568.14 frames. ], batch size: 47, lr: 1.27e-02, grad_scale: 16.0 +2023-03-08 23:46:29,098 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=30829.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:46:43,007 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.320e+02 3.498e+02 4.092e+02 4.863e+02 1.038e+03, threshold=8.183e+02, percent-clipped=2.0 +2023-03-08 23:47:03,289 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-08 23:47:11,806 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=30865.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:47:20,586 INFO [train.py:898] (0/4) Epoch 9, batch 1800, loss[loss=0.1579, simple_loss=0.2326, pruned_loss=0.04155, over 18449.00 frames. ], tot_loss[loss=0.1997, simple_loss=0.282, pruned_loss=0.05871, over 3582579.99 frames. ], batch size: 43, lr: 1.27e-02, grad_scale: 8.0 +2023-03-08 23:47:57,652 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30903.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:48:05,855 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5508, 2.7653, 2.6949, 2.7696, 3.5100, 3.4929, 3.1193, 2.9430], + device='cuda:0'), covar=tensor([0.0171, 0.0334, 0.0635, 0.0320, 0.0223, 0.0150, 0.0334, 0.0331], + device='cuda:0'), in_proj_covar=tensor([0.0112, 0.0096, 0.0144, 0.0125, 0.0093, 0.0078, 0.0122, 0.0119], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 23:48:09,040 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=30913.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:48:20,086 INFO [train.py:898] (0/4) Epoch 9, batch 1850, loss[loss=0.2223, simple_loss=0.2974, pruned_loss=0.07357, over 12341.00 frames. ], tot_loss[loss=0.1986, simple_loss=0.2811, pruned_loss=0.0581, over 3581431.04 frames. ], batch size: 129, lr: 1.26e-02, grad_scale: 8.0 +2023-03-08 23:48:22,724 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30925.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:48:31,839 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7811, 3.7968, 5.0467, 2.9058, 4.2346, 2.6511, 2.9503, 1.9286], + device='cuda:0'), covar=tensor([0.0912, 0.0735, 0.0071, 0.0604, 0.0585, 0.2135, 0.2474, 0.1800], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0208, 0.0105, 0.0162, 0.0223, 0.0241, 0.0272, 0.0205], + device='cuda:0'), out_proj_covar=tensor([1.6765e-04, 1.9031e-04, 9.7246e-05, 1.4659e-04, 2.0385e-04, 2.2261e-04, + 2.4818e-04, 1.9014e-04], device='cuda:0') +2023-03-08 23:48:33,882 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=30934.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:48:42,066 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.240e+02 3.365e+02 3.795e+02 4.610e+02 7.960e+02, threshold=7.590e+02, percent-clipped=0.0 +2023-03-08 23:48:45,067 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-08 23:49:09,547 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30964.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:49:19,303 INFO [train.py:898] (0/4) Epoch 9, batch 1900, loss[loss=0.2299, simple_loss=0.3128, pruned_loss=0.07355, over 18157.00 frames. ], tot_loss[loss=0.199, simple_loss=0.2814, pruned_loss=0.05826, over 3587093.58 frames. ], batch size: 62, lr: 1.26e-02, grad_scale: 8.0 +2023-03-08 23:49:34,704 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30986.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:49:45,333 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=30995.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:50:04,635 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31011.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 23:50:04,742 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31011.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 23:50:10,173 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31016.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:50:17,975 INFO [train.py:898] (0/4) Epoch 9, batch 1950, loss[loss=0.203, simple_loss=0.2679, pruned_loss=0.06905, over 18468.00 frames. ], tot_loss[loss=0.1997, simple_loss=0.2822, pruned_loss=0.05864, over 3585108.10 frames. ], batch size: 43, lr: 1.26e-02, grad_scale: 8.0 +2023-03-08 23:50:39,080 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.073e+02 3.300e+02 4.115e+02 5.228e+02 1.013e+03, threshold=8.231e+02, percent-clipped=3.0 +2023-03-08 23:51:06,735 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=31064.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:51:12,694 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2140, 5.2613, 2.9500, 5.0716, 4.9802, 5.3596, 5.1490, 2.7630], + device='cuda:0'), covar=tensor([0.0139, 0.0060, 0.0661, 0.0082, 0.0082, 0.0063, 0.0096, 0.0918], + device='cuda:0'), in_proj_covar=tensor([0.0072, 0.0061, 0.0086, 0.0077, 0.0072, 0.0059, 0.0074, 0.0090], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0005, 0.0004, 0.0004, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 23:51:16,215 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31072.0, num_to_drop=1, layers_to_drop={3} +2023-03-08 23:51:16,963 INFO [train.py:898] (0/4) Epoch 9, batch 2000, loss[loss=0.2695, simple_loss=0.3325, pruned_loss=0.1033, over 12932.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.2818, pruned_loss=0.05846, over 3585536.18 frames. ], batch size: 129, lr: 1.26e-02, grad_scale: 8.0 +2023-03-08 23:51:18,762 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-08 23:51:49,466 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7787, 4.6919, 4.8259, 3.5888, 3.8729, 3.7848, 2.8421, 2.2096], + device='cuda:0'), covar=tensor([0.0217, 0.0139, 0.0052, 0.0211, 0.0300, 0.0171, 0.0600, 0.0905], + device='cuda:0'), in_proj_covar=tensor([0.0054, 0.0044, 0.0043, 0.0056, 0.0075, 0.0052, 0.0069, 0.0075], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-08 23:51:54,165 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5196, 2.5226, 2.4501, 2.6915, 3.4125, 3.4854, 2.9349, 2.6727], + device='cuda:0'), covar=tensor([0.0164, 0.0315, 0.0696, 0.0397, 0.0233, 0.0152, 0.0432, 0.0433], + device='cuda:0'), in_proj_covar=tensor([0.0112, 0.0096, 0.0148, 0.0126, 0.0093, 0.0079, 0.0124, 0.0118], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-08 23:52:15,798 INFO [train.py:898] (0/4) Epoch 9, batch 2050, loss[loss=0.156, simple_loss=0.2346, pruned_loss=0.03869, over 18414.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.2823, pruned_loss=0.05875, over 3577098.19 frames. ], batch size: 42, lr: 1.26e-02, grad_scale: 8.0 +2023-03-08 23:52:37,043 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.601e+02 3.647e+02 4.491e+02 5.379e+02 9.813e+02, threshold=8.982e+02, percent-clipped=3.0 +2023-03-08 23:53:15,209 INFO [train.py:898] (0/4) Epoch 9, batch 2100, loss[loss=0.1769, simple_loss=0.2531, pruned_loss=0.05033, over 18251.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.2819, pruned_loss=0.05866, over 3569055.33 frames. ], batch size: 45, lr: 1.26e-02, grad_scale: 8.0 +2023-03-08 23:54:14,808 INFO [train.py:898] (0/4) Epoch 9, batch 2150, loss[loss=0.2391, simple_loss=0.3081, pruned_loss=0.085, over 12459.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.2823, pruned_loss=0.0587, over 3571293.36 frames. ], batch size: 129, lr: 1.26e-02, grad_scale: 8.0 +2023-03-08 23:54:19,971 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6800, 3.7288, 3.5606, 3.0889, 3.5407, 2.5930, 2.8360, 3.7807], + device='cuda:0'), covar=tensor([0.0033, 0.0052, 0.0066, 0.0119, 0.0065, 0.0191, 0.0165, 0.0036], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0103, 0.0092, 0.0140, 0.0093, 0.0138, 0.0145, 0.0077], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0001, 0.0001, 0.0002, 0.0001, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-08 23:54:35,294 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.335e+02 3.491e+02 4.023e+02 4.961e+02 1.002e+03, threshold=8.046e+02, percent-clipped=2.0 +2023-03-08 23:54:56,921 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31259.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:55:13,348 INFO [train.py:898] (0/4) Epoch 9, batch 2200, loss[loss=0.197, simple_loss=0.2814, pruned_loss=0.05629, over 16144.00 frames. ], tot_loss[loss=0.1999, simple_loss=0.2824, pruned_loss=0.0587, over 3582547.82 frames. ], batch size: 94, lr: 1.26e-02, grad_scale: 4.0 +2023-03-08 23:55:15,938 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1420, 5.0347, 2.7550, 4.8922, 4.7968, 5.0829, 4.9093, 2.6997], + device='cuda:0'), covar=tensor([0.0130, 0.0051, 0.0619, 0.0067, 0.0069, 0.0054, 0.0069, 0.0866], + device='cuda:0'), in_proj_covar=tensor([0.0073, 0.0062, 0.0087, 0.0077, 0.0072, 0.0060, 0.0075, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0005, 0.0004, 0.0004, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 23:55:22,629 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31281.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:55:32,581 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31290.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:55:57,270 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31311.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:55:57,793 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-08 23:56:11,961 INFO [train.py:898] (0/4) Epoch 9, batch 2250, loss[loss=0.1828, simple_loss=0.2623, pruned_loss=0.0517, over 18504.00 frames. ], tot_loss[loss=0.2001, simple_loss=0.2825, pruned_loss=0.05882, over 3585901.90 frames. ], batch size: 47, lr: 1.26e-02, grad_scale: 4.0 +2023-03-08 23:56:33,578 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.219e+02 3.737e+02 4.347e+02 5.503e+02 2.827e+03, threshold=8.695e+02, percent-clipped=9.0 +2023-03-08 23:56:53,440 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=31359.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:57:00,817 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1494, 5.0963, 5.4415, 5.2103, 5.0988, 5.9675, 5.4494, 5.1688], + device='cuda:0'), covar=tensor([0.0970, 0.0584, 0.0613, 0.0620, 0.1353, 0.0638, 0.0583, 0.1613], + device='cuda:0'), in_proj_covar=tensor([0.0289, 0.0219, 0.0229, 0.0232, 0.0271, 0.0322, 0.0217, 0.0322], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-08 23:57:02,909 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31367.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 23:57:09,986 INFO [train.py:898] (0/4) Epoch 9, batch 2300, loss[loss=0.1877, simple_loss=0.2705, pruned_loss=0.05241, over 18421.00 frames. ], tot_loss[loss=0.2005, simple_loss=0.2825, pruned_loss=0.05923, over 3582107.75 frames. ], batch size: 48, lr: 1.26e-02, grad_scale: 4.0 +2023-03-08 23:57:50,327 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5496, 3.4000, 1.9795, 4.3199, 2.8677, 4.3051, 2.0159, 3.8245], + device='cuda:0'), covar=tensor([0.0510, 0.0755, 0.1416, 0.0325, 0.0857, 0.0236, 0.1283, 0.0367], + device='cuda:0'), in_proj_covar=tensor([0.0182, 0.0207, 0.0178, 0.0218, 0.0176, 0.0222, 0.0188, 0.0179], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-08 23:57:52,661 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31409.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:58:09,022 INFO [train.py:898] (0/4) Epoch 9, batch 2350, loss[loss=0.1899, simple_loss=0.27, pruned_loss=0.05495, over 18279.00 frames. ], tot_loss[loss=0.2001, simple_loss=0.2822, pruned_loss=0.05905, over 3579964.60 frames. ], batch size: 47, lr: 1.25e-02, grad_scale: 4.0 +2023-03-08 23:58:31,514 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.260e+02 3.212e+02 3.852e+02 4.857e+02 1.133e+03, threshold=7.704e+02, percent-clipped=2.0 +2023-03-08 23:58:59,784 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1051, 5.2150, 2.7282, 5.0413, 4.8974, 5.2465, 5.0449, 2.6769], + device='cuda:0'), covar=tensor([0.0154, 0.0048, 0.0796, 0.0076, 0.0068, 0.0054, 0.0087, 0.0990], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0062, 0.0088, 0.0077, 0.0073, 0.0061, 0.0075, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0005, 0.0004, 0.0004, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-08 23:59:04,361 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31470.0, num_to_drop=0, layers_to_drop=set() +2023-03-08 23:59:07,398 INFO [train.py:898] (0/4) Epoch 9, batch 2400, loss[loss=0.1771, simple_loss=0.2564, pruned_loss=0.04892, over 18161.00 frames. ], tot_loss[loss=0.2001, simple_loss=0.2826, pruned_loss=0.05887, over 3574854.60 frames. ], batch size: 44, lr: 1.25e-02, grad_scale: 8.0 +2023-03-08 23:59:29,262 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31491.0, num_to_drop=1, layers_to_drop={0} +2023-03-08 23:59:59,443 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6572, 5.3219, 5.4004, 5.3243, 4.9191, 5.1966, 4.5361, 5.1756], + device='cuda:0'), covar=tensor([0.0254, 0.0269, 0.0160, 0.0319, 0.0343, 0.0245, 0.1194, 0.0292], + device='cuda:0'), in_proj_covar=tensor([0.0163, 0.0203, 0.0190, 0.0218, 0.0201, 0.0210, 0.0270, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0006, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-09 00:00:06,630 INFO [train.py:898] (0/4) Epoch 9, batch 2450, loss[loss=0.1971, simple_loss=0.2836, pruned_loss=0.05536, over 17061.00 frames. ], tot_loss[loss=0.1986, simple_loss=0.2812, pruned_loss=0.05799, over 3580718.73 frames. ], batch size: 78, lr: 1.25e-02, grad_scale: 8.0 +2023-03-09 00:00:25,399 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8742, 3.9395, 5.2242, 3.0635, 4.3023, 2.5637, 3.1395, 2.0459], + device='cuda:0'), covar=tensor([0.0835, 0.0681, 0.0048, 0.0545, 0.0607, 0.2053, 0.2034, 0.1643], + device='cuda:0'), in_proj_covar=tensor([0.0188, 0.0207, 0.0105, 0.0160, 0.0220, 0.0238, 0.0267, 0.0200], + device='cuda:0'), out_proj_covar=tensor([1.6694e-04, 1.8903e-04, 9.6925e-05, 1.4424e-04, 2.0066e-04, 2.2010e-04, + 2.4357e-04, 1.8626e-04], device='cuda:0') +2023-03-09 00:00:29,271 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.799e+02 3.437e+02 4.058e+02 5.079e+02 1.109e+03, threshold=8.115e+02, percent-clipped=5.0 +2023-03-09 00:00:35,495 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8083, 2.9140, 4.1555, 4.1138, 2.5148, 4.7083, 3.8843, 2.9338], + device='cuda:0'), covar=tensor([0.0401, 0.1263, 0.0292, 0.0229, 0.1521, 0.0163, 0.0441, 0.0970], + device='cuda:0'), in_proj_covar=tensor([0.0179, 0.0210, 0.0137, 0.0132, 0.0203, 0.0176, 0.0194, 0.0188], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 00:00:40,984 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31552.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 00:00:49,205 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31559.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:01:05,460 INFO [train.py:898] (0/4) Epoch 9, batch 2500, loss[loss=0.214, simple_loss=0.2973, pruned_loss=0.06541, over 17717.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.2806, pruned_loss=0.05784, over 3582329.62 frames. ], batch size: 70, lr: 1.25e-02, grad_scale: 8.0 +2023-03-09 00:01:09,165 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6079, 5.2252, 5.2625, 5.2871, 4.8683, 5.1482, 4.4896, 5.0722], + device='cuda:0'), covar=tensor([0.0236, 0.0245, 0.0181, 0.0260, 0.0294, 0.0218, 0.1101, 0.0298], + device='cuda:0'), in_proj_covar=tensor([0.0161, 0.0203, 0.0189, 0.0218, 0.0200, 0.0210, 0.0268, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0006, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-09 00:01:15,248 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31581.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:01:25,893 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31590.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:01:45,093 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=31607.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:02:03,063 INFO [train.py:898] (0/4) Epoch 9, batch 2550, loss[loss=0.2149, simple_loss=0.299, pruned_loss=0.06537, over 17805.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.2806, pruned_loss=0.05794, over 3576580.33 frames. ], batch size: 70, lr: 1.25e-02, grad_scale: 8.0 +2023-03-09 00:02:10,576 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=31629.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:02:21,823 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=31638.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:02:24,079 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7798, 4.8362, 4.8315, 4.8274, 4.7598, 5.4778, 5.1200, 4.8277], + device='cuda:0'), covar=tensor([0.0937, 0.0758, 0.0867, 0.0714, 0.1405, 0.0878, 0.0645, 0.1540], + device='cuda:0'), in_proj_covar=tensor([0.0290, 0.0220, 0.0231, 0.0235, 0.0273, 0.0329, 0.0217, 0.0321], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 00:02:26,082 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.346e+02 3.602e+02 4.315e+02 5.683e+02 1.136e+03, threshold=8.630e+02, percent-clipped=7.0 +2023-03-09 00:02:54,678 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=31667.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 00:02:54,750 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0367, 4.1410, 2.3988, 4.2342, 5.1196, 2.4397, 3.3922, 3.5543], + device='cuda:0'), covar=tensor([0.0080, 0.1269, 0.1738, 0.0580, 0.0055, 0.1358, 0.0890, 0.0995], + device='cuda:0'), in_proj_covar=tensor([0.0104, 0.0217, 0.0187, 0.0188, 0.0085, 0.0173, 0.0201, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 00:03:01,171 INFO [train.py:898] (0/4) Epoch 9, batch 2600, loss[loss=0.2492, simple_loss=0.3169, pruned_loss=0.09073, over 12388.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.2814, pruned_loss=0.05836, over 3572439.62 frames. ], batch size: 130, lr: 1.25e-02, grad_scale: 8.0 +2023-03-09 00:03:37,734 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31704.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:03:50,198 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=31715.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 00:03:59,006 INFO [train.py:898] (0/4) Epoch 9, batch 2650, loss[loss=0.224, simple_loss=0.3136, pruned_loss=0.06726, over 18586.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.2817, pruned_loss=0.05849, over 3578567.67 frames. ], batch size: 54, lr: 1.25e-02, grad_scale: 8.0 +2023-03-09 00:04:21,589 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.212e+02 3.574e+02 4.199e+02 5.233e+02 1.424e+03, threshold=8.398e+02, percent-clipped=3.0 +2023-03-09 00:04:48,298 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31765.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:04:48,528 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31765.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:04:57,869 INFO [train.py:898] (0/4) Epoch 9, batch 2700, loss[loss=0.1986, simple_loss=0.287, pruned_loss=0.05506, over 18365.00 frames. ], tot_loss[loss=0.1997, simple_loss=0.2826, pruned_loss=0.05846, over 3570619.71 frames. ], batch size: 56, lr: 1.25e-02, grad_scale: 8.0 +2023-03-09 00:05:15,545 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31788.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:05:56,590 INFO [train.py:898] (0/4) Epoch 9, batch 2750, loss[loss=0.223, simple_loss=0.3001, pruned_loss=0.07298, over 18295.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.2825, pruned_loss=0.05818, over 3576666.84 frames. ], batch size: 57, lr: 1.25e-02, grad_scale: 8.0 +2023-03-09 00:05:56,807 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31823.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:06:02,576 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6607, 2.8634, 2.8447, 2.8621, 3.6532, 3.5196, 3.1655, 3.0505], + device='cuda:0'), covar=tensor([0.0226, 0.0307, 0.0522, 0.0355, 0.0217, 0.0209, 0.0298, 0.0314], + device='cuda:0'), in_proj_covar=tensor([0.0110, 0.0097, 0.0145, 0.0127, 0.0092, 0.0080, 0.0122, 0.0117], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 00:06:19,351 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.068e+02 3.248e+02 4.036e+02 4.734e+02 1.785e+03, threshold=8.071e+02, percent-clipped=3.0 +2023-03-09 00:06:25,185 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=31847.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 00:06:27,551 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31849.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:06:55,868 INFO [train.py:898] (0/4) Epoch 9, batch 2800, loss[loss=0.1852, simple_loss=0.273, pruned_loss=0.04877, over 18506.00 frames. ], tot_loss[loss=0.1976, simple_loss=0.2808, pruned_loss=0.05718, over 3580801.61 frames. ], batch size: 53, lr: 1.25e-02, grad_scale: 8.0 +2023-03-09 00:07:02,070 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31878.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:07:05,565 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5094, 2.7003, 2.6851, 2.6951, 3.6216, 3.4491, 2.9268, 2.8611], + device='cuda:0'), covar=tensor([0.0255, 0.0323, 0.0571, 0.0402, 0.0164, 0.0167, 0.0361, 0.0332], + device='cuda:0'), in_proj_covar=tensor([0.0111, 0.0097, 0.0145, 0.0127, 0.0091, 0.0080, 0.0123, 0.0117], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 00:07:09,074 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31884.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:07:19,955 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-09 00:07:27,543 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-09 00:07:45,696 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31915.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:07:54,184 INFO [train.py:898] (0/4) Epoch 9, batch 2850, loss[loss=0.1955, simple_loss=0.2753, pruned_loss=0.05783, over 18547.00 frames. ], tot_loss[loss=0.1978, simple_loss=0.2811, pruned_loss=0.05726, over 3587872.54 frames. ], batch size: 49, lr: 1.25e-02, grad_scale: 8.0 +2023-03-09 00:08:07,388 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8636, 4.8102, 5.0077, 4.6070, 4.6814, 4.7252, 5.1101, 5.0515], + device='cuda:0'), covar=tensor([0.0056, 0.0082, 0.0051, 0.0093, 0.0065, 0.0096, 0.0079, 0.0096], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0055, 0.0055, 0.0070, 0.0060, 0.0081, 0.0068, 0.0069], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 00:08:13,155 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31939.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:08:16,250 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.274e+02 3.441e+02 4.233e+02 5.343e+02 2.679e+03, threshold=8.467e+02, percent-clipped=8.0 +2023-03-09 00:08:26,698 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=31950.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:08:30,362 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9747, 4.6343, 4.6879, 3.2782, 3.7589, 3.6515, 2.7305, 2.7921], + device='cuda:0'), covar=tensor([0.0187, 0.0166, 0.0073, 0.0319, 0.0371, 0.0206, 0.0688, 0.0729], + device='cuda:0'), in_proj_covar=tensor([0.0055, 0.0045, 0.0044, 0.0057, 0.0076, 0.0053, 0.0070, 0.0077], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 00:08:36,579 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5832, 1.9697, 2.5815, 2.5988, 3.2934, 4.8958, 4.4066, 3.7610], + device='cuda:0'), covar=tensor([0.1080, 0.1931, 0.2290, 0.1296, 0.1677, 0.0088, 0.0367, 0.0475], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0278, 0.0288, 0.0239, 0.0348, 0.0168, 0.0245, 0.0188], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 00:08:52,779 INFO [train.py:898] (0/4) Epoch 9, batch 2900, loss[loss=0.2197, simple_loss=0.3089, pruned_loss=0.06528, over 18288.00 frames. ], tot_loss[loss=0.1978, simple_loss=0.2812, pruned_loss=0.05721, over 3590637.70 frames. ], batch size: 57, lr: 1.24e-02, grad_scale: 8.0 +2023-03-09 00:08:56,504 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=31976.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:09:24,953 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-32000.pt +2023-03-09 00:09:37,333 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7468, 5.3201, 5.1886, 5.1826, 4.7864, 5.1524, 4.4592, 5.1348], + device='cuda:0'), covar=tensor([0.0232, 0.0222, 0.0220, 0.0381, 0.0382, 0.0221, 0.1317, 0.0285], + device='cuda:0'), in_proj_covar=tensor([0.0163, 0.0203, 0.0191, 0.0223, 0.0203, 0.0210, 0.0272, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0006, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-09 00:09:43,665 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=32011.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:09:56,749 INFO [train.py:898] (0/4) Epoch 9, batch 2950, loss[loss=0.1818, simple_loss=0.255, pruned_loss=0.05435, over 18414.00 frames. ], tot_loss[loss=0.1981, simple_loss=0.2813, pruned_loss=0.05751, over 3579332.91 frames. ], batch size: 42, lr: 1.24e-02, grad_scale: 8.0 +2023-03-09 00:10:19,144 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.223e+02 3.183e+02 3.919e+02 4.606e+02 8.522e+02, threshold=7.838e+02, percent-clipped=1.0 +2023-03-09 00:10:41,317 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32060.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:10:47,579 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32065.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:10:56,316 INFO [train.py:898] (0/4) Epoch 9, batch 3000, loss[loss=0.2271, simple_loss=0.3059, pruned_loss=0.07413, over 16060.00 frames. ], tot_loss[loss=0.1983, simple_loss=0.2814, pruned_loss=0.0576, over 3581832.54 frames. ], batch size: 94, lr: 1.24e-02, grad_scale: 8.0 +2023-03-09 00:10:56,318 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 00:11:08,351 INFO [train.py:932] (0/4) Epoch 9, validation: loss=0.1618, simple_loss=0.2644, pruned_loss=0.02958, over 944034.00 frames. +2023-03-09 00:11:08,352 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-09 00:11:55,403 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=32113.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:12:07,343 INFO [train.py:898] (0/4) Epoch 9, batch 3050, loss[loss=0.2529, simple_loss=0.3237, pruned_loss=0.09107, over 12056.00 frames. ], tot_loss[loss=0.1994, simple_loss=0.2822, pruned_loss=0.05828, over 3574377.64 frames. ], batch size: 130, lr: 1.24e-02, grad_scale: 8.0 +2023-03-09 00:12:29,407 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.233e+02 3.379e+02 3.882e+02 4.685e+02 8.666e+02, threshold=7.765e+02, percent-clipped=1.0 +2023-03-09 00:12:32,450 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32144.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:12:35,837 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32147.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 00:12:46,834 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8860, 3.0579, 4.4072, 4.1154, 2.8928, 4.7590, 4.1319, 3.1885], + device='cuda:0'), covar=tensor([0.0346, 0.1194, 0.0170, 0.0237, 0.1281, 0.0172, 0.0324, 0.0803], + device='cuda:0'), in_proj_covar=tensor([0.0173, 0.0206, 0.0134, 0.0130, 0.0200, 0.0169, 0.0187, 0.0183], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 00:12:56,135 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-09 00:13:05,895 INFO [train.py:898] (0/4) Epoch 9, batch 3100, loss[loss=0.2185, simple_loss=0.297, pruned_loss=0.06996, over 18357.00 frames. ], tot_loss[loss=0.1996, simple_loss=0.2827, pruned_loss=0.05832, over 3580521.94 frames. ], batch size: 56, lr: 1.24e-02, grad_scale: 8.0 +2023-03-09 00:13:13,304 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32179.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:13:23,123 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5115, 4.4805, 2.7358, 4.5562, 5.5665, 2.9885, 4.1171, 4.2718], + device='cuda:0'), covar=tensor([0.0064, 0.1113, 0.1483, 0.0466, 0.0032, 0.1109, 0.0585, 0.0703], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0212, 0.0185, 0.0182, 0.0083, 0.0169, 0.0197, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 00:13:31,156 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2452, 4.3265, 2.3163, 4.3009, 5.3059, 2.5064, 3.7417, 4.0123], + device='cuda:0'), covar=tensor([0.0064, 0.0946, 0.1627, 0.0507, 0.0039, 0.1343, 0.0675, 0.0678], + device='cuda:0'), in_proj_covar=tensor([0.0102, 0.0210, 0.0183, 0.0181, 0.0083, 0.0168, 0.0195, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0003, 0.0002, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 00:13:32,009 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=32195.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 00:14:05,295 INFO [train.py:898] (0/4) Epoch 9, batch 3150, loss[loss=0.1861, simple_loss=0.2665, pruned_loss=0.05291, over 18264.00 frames. ], tot_loss[loss=0.1982, simple_loss=0.2812, pruned_loss=0.0576, over 3589730.08 frames. ], batch size: 47, lr: 1.24e-02, grad_scale: 8.0 +2023-03-09 00:14:18,427 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32234.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:14:28,070 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.206e+02 3.166e+02 3.799e+02 4.714e+02 1.226e+03, threshold=7.598e+02, percent-clipped=2.0 +2023-03-09 00:15:02,010 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32271.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:15:04,020 INFO [train.py:898] (0/4) Epoch 9, batch 3200, loss[loss=0.1968, simple_loss=0.2813, pruned_loss=0.0562, over 16889.00 frames. ], tot_loss[loss=0.1989, simple_loss=0.282, pruned_loss=0.05793, over 3595392.79 frames. ], batch size: 78, lr: 1.24e-02, grad_scale: 8.0 +2023-03-09 00:15:43,097 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=32306.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:16:02,577 INFO [train.py:898] (0/4) Epoch 9, batch 3250, loss[loss=0.178, simple_loss=0.2567, pruned_loss=0.04958, over 18156.00 frames. ], tot_loss[loss=0.1984, simple_loss=0.2815, pruned_loss=0.0577, over 3606747.33 frames. ], batch size: 44, lr: 1.24e-02, grad_scale: 8.0 +2023-03-09 00:16:24,631 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.580e+02 3.470e+02 4.080e+02 5.186e+02 8.555e+02, threshold=8.161e+02, percent-clipped=3.0 +2023-03-09 00:16:46,344 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32360.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:17:00,677 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3553, 2.5922, 3.6472, 3.7236, 2.6880, 4.0813, 3.6315, 2.7189], + device='cuda:0'), covar=tensor([0.0476, 0.1329, 0.0315, 0.0250, 0.1292, 0.0206, 0.0487, 0.0994], + device='cuda:0'), in_proj_covar=tensor([0.0179, 0.0210, 0.0139, 0.0133, 0.0204, 0.0174, 0.0193, 0.0187], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 00:17:01,462 INFO [train.py:898] (0/4) Epoch 9, batch 3300, loss[loss=0.1995, simple_loss=0.291, pruned_loss=0.05399, over 17217.00 frames. ], tot_loss[loss=0.1979, simple_loss=0.2808, pruned_loss=0.05745, over 3612127.97 frames. ], batch size: 78, lr: 1.24e-02, grad_scale: 8.0 +2023-03-09 00:17:43,021 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=32408.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:17:56,582 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9827, 4.8878, 5.0834, 4.6937, 4.8731, 4.8917, 5.1752, 5.1613], + device='cuda:0'), covar=tensor([0.0063, 0.0092, 0.0079, 0.0112, 0.0072, 0.0107, 0.0072, 0.0084], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0054, 0.0056, 0.0071, 0.0060, 0.0081, 0.0069, 0.0069], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 00:17:58,757 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3018, 5.8874, 5.3892, 5.5924, 5.3503, 5.3278, 5.9037, 5.8842], + device='cuda:0'), covar=tensor([0.1189, 0.0650, 0.0528, 0.0729, 0.1442, 0.0748, 0.0535, 0.0609], + device='cuda:0'), in_proj_covar=tensor([0.0504, 0.0408, 0.0310, 0.0455, 0.0611, 0.0452, 0.0578, 0.0430], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 00:18:00,699 INFO [train.py:898] (0/4) Epoch 9, batch 3350, loss[loss=0.1849, simple_loss=0.2721, pruned_loss=0.04888, over 18355.00 frames. ], tot_loss[loss=0.1973, simple_loss=0.2802, pruned_loss=0.05723, over 3609016.93 frames. ], batch size: 55, lr: 1.24e-02, grad_scale: 8.0 +2023-03-09 00:18:00,967 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4540, 5.4483, 5.0088, 5.3931, 5.4705, 4.6771, 5.3412, 5.0343], + device='cuda:0'), covar=tensor([0.0378, 0.0368, 0.1282, 0.0719, 0.0451, 0.0447, 0.0349, 0.0915], + device='cuda:0'), in_proj_covar=tensor([0.0387, 0.0433, 0.0583, 0.0344, 0.0320, 0.0400, 0.0421, 0.0554], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 00:18:22,502 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.160e+02 3.179e+02 3.976e+02 5.107e+02 1.332e+03, threshold=7.951e+02, percent-clipped=3.0 +2023-03-09 00:18:25,074 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32444.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:18:35,534 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8263, 3.7893, 3.6043, 3.0691, 3.5377, 2.7277, 3.0303, 3.8672], + device='cuda:0'), covar=tensor([0.0029, 0.0054, 0.0051, 0.0117, 0.0068, 0.0140, 0.0131, 0.0036], + device='cuda:0'), in_proj_covar=tensor([0.0082, 0.0108, 0.0098, 0.0145, 0.0096, 0.0142, 0.0150, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 00:18:59,649 INFO [train.py:898] (0/4) Epoch 9, batch 3400, loss[loss=0.1932, simple_loss=0.2743, pruned_loss=0.05602, over 18394.00 frames. ], tot_loss[loss=0.1968, simple_loss=0.2798, pruned_loss=0.05687, over 3608542.15 frames. ], batch size: 48, lr: 1.23e-02, grad_scale: 8.0 +2023-03-09 00:19:06,796 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32479.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:19:16,287 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.49 vs. limit=2.0 +2023-03-09 00:19:18,066 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7541, 4.5293, 4.6478, 3.4630, 3.6773, 3.7159, 2.4688, 2.1394], + device='cuda:0'), covar=tensor([0.0213, 0.0147, 0.0070, 0.0279, 0.0301, 0.0194, 0.0790, 0.0963], + device='cuda:0'), in_proj_covar=tensor([0.0056, 0.0046, 0.0045, 0.0057, 0.0076, 0.0053, 0.0071, 0.0077], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 00:19:21,679 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=32492.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:19:51,415 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.2634, 3.2012, 4.5486, 3.8943, 2.9721, 2.7449, 3.8746, 4.4891], + device='cuda:0'), covar=tensor([0.0899, 0.1471, 0.0100, 0.0330, 0.0838, 0.1091, 0.0374, 0.0223], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0231, 0.0090, 0.0153, 0.0169, 0.0171, 0.0161, 0.0129], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0002, 0.0003, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 00:19:57,959 INFO [train.py:898] (0/4) Epoch 9, batch 3450, loss[loss=0.1882, simple_loss=0.2652, pruned_loss=0.05553, over 18270.00 frames. ], tot_loss[loss=0.1973, simple_loss=0.2801, pruned_loss=0.05722, over 3596910.74 frames. ], batch size: 45, lr: 1.23e-02, grad_scale: 8.0 +2023-03-09 00:20:02,583 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=32527.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:20:10,639 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32534.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:20:19,239 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.059e+02 3.448e+02 3.896e+02 4.764e+02 9.293e+02, threshold=7.793e+02, percent-clipped=1.0 +2023-03-09 00:20:53,909 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32571.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:20:56,532 INFO [train.py:898] (0/4) Epoch 9, batch 3500, loss[loss=0.1997, simple_loss=0.2804, pruned_loss=0.05952, over 16419.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2791, pruned_loss=0.05702, over 3603640.91 frames. ], batch size: 94, lr: 1.23e-02, grad_scale: 8.0 +2023-03-09 00:21:06,827 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=32582.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:21:34,290 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=32606.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:21:47,754 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=32619.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:21:52,046 INFO [train.py:898] (0/4) Epoch 9, batch 3550, loss[loss=0.1787, simple_loss=0.2725, pruned_loss=0.04244, over 18513.00 frames. ], tot_loss[loss=0.1971, simple_loss=0.2796, pruned_loss=0.05728, over 3594388.87 frames. ], batch size: 51, lr: 1.23e-02, grad_scale: 8.0 +2023-03-09 00:22:09,004 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.81 vs. limit=5.0 +2023-03-09 00:22:12,489 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.120e+02 3.503e+02 4.111e+02 5.019e+02 1.415e+03, threshold=8.222e+02, percent-clipped=4.0 +2023-03-09 00:22:22,694 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6997, 3.4619, 5.1473, 2.9733, 4.3047, 2.4353, 2.9888, 2.0478], + device='cuda:0'), covar=tensor([0.0898, 0.0826, 0.0062, 0.0529, 0.0467, 0.2202, 0.2206, 0.1679], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0209, 0.0109, 0.0162, 0.0224, 0.0240, 0.0269, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 00:22:25,441 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=32654.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:22:37,368 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0716, 5.0954, 5.0432, 4.8314, 4.8579, 4.9403, 5.3228, 5.2647], + device='cuda:0'), covar=tensor([0.0053, 0.0068, 0.0061, 0.0084, 0.0055, 0.0097, 0.0079, 0.0090], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0055, 0.0056, 0.0071, 0.0060, 0.0082, 0.0069, 0.0070], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 00:22:45,803 INFO [train.py:898] (0/4) Epoch 9, batch 3600, loss[loss=0.1723, simple_loss=0.2532, pruned_loss=0.04569, over 18376.00 frames. ], tot_loss[loss=0.1964, simple_loss=0.2791, pruned_loss=0.05687, over 3605428.28 frames. ], batch size: 42, lr: 1.23e-02, grad_scale: 8.0 +2023-03-09 00:23:13,222 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1928, 4.2207, 3.9430, 4.1540, 4.1890, 3.7750, 4.1672, 3.9848], + device='cuda:0'), covar=tensor([0.0472, 0.0599, 0.1327, 0.0750, 0.0534, 0.0451, 0.0451, 0.0938], + device='cuda:0'), in_proj_covar=tensor([0.0381, 0.0431, 0.0577, 0.0343, 0.0318, 0.0395, 0.0423, 0.0545], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 00:23:22,110 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-9.pt +2023-03-09 00:23:53,331 INFO [train.py:898] (0/4) Epoch 10, batch 0, loss[loss=0.2102, simple_loss=0.2934, pruned_loss=0.06351, over 18393.00 frames. ], tot_loss[loss=0.2102, simple_loss=0.2934, pruned_loss=0.06351, over 18393.00 frames. ], batch size: 52, lr: 1.17e-02, grad_scale: 8.0 +2023-03-09 00:23:53,333 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 00:24:05,219 INFO [train.py:932] (0/4) Epoch 10, validation: loss=0.1621, simple_loss=0.2651, pruned_loss=0.02958, over 944034.00 frames. +2023-03-09 00:24:05,219 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-09 00:24:32,921 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2486, 3.0884, 1.7476, 3.8788, 2.6401, 3.9109, 2.1319, 3.3984], + device='cuda:0'), covar=tensor([0.0576, 0.0819, 0.1526, 0.0537, 0.0865, 0.0264, 0.1196, 0.0419], + device='cuda:0'), in_proj_covar=tensor([0.0182, 0.0207, 0.0175, 0.0221, 0.0175, 0.0221, 0.0187, 0.0180], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 00:24:46,644 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.116e+02 3.572e+02 4.322e+02 5.704e+02 1.376e+03, threshold=8.645e+02, percent-clipped=6.0 +2023-03-09 00:25:02,037 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5156, 2.1297, 4.0758, 4.0678, 2.2201, 4.4227, 3.8027, 2.4861], + device='cuda:0'), covar=tensor([0.0442, 0.2088, 0.0224, 0.0201, 0.1986, 0.0206, 0.0409, 0.1337], + device='cuda:0'), in_proj_covar=tensor([0.0182, 0.0220, 0.0142, 0.0138, 0.0210, 0.0179, 0.0198, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 00:25:03,986 INFO [train.py:898] (0/4) Epoch 10, batch 50, loss[loss=0.1665, simple_loss=0.2506, pruned_loss=0.04123, over 18288.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.2814, pruned_loss=0.05844, over 800265.09 frames. ], batch size: 49, lr: 1.17e-02, grad_scale: 8.0 +2023-03-09 00:25:16,552 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-09 00:25:34,757 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5572, 1.9340, 2.6022, 2.5230, 3.3274, 5.1470, 4.6076, 4.2535], + device='cuda:0'), covar=tensor([0.1034, 0.1963, 0.2206, 0.1358, 0.1665, 0.0073, 0.0314, 0.0339], + device='cuda:0'), in_proj_covar=tensor([0.0225, 0.0282, 0.0291, 0.0240, 0.0348, 0.0171, 0.0243, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 00:25:57,167 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8680, 4.8237, 4.9213, 4.6204, 4.5606, 4.7406, 5.0837, 5.0108], + device='cuda:0'), covar=tensor([0.0056, 0.0079, 0.0075, 0.0095, 0.0073, 0.0101, 0.0073, 0.0099], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0055, 0.0056, 0.0070, 0.0060, 0.0081, 0.0068, 0.0069], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 00:26:02,543 INFO [train.py:898] (0/4) Epoch 10, batch 100, loss[loss=0.2078, simple_loss=0.2914, pruned_loss=0.06209, over 18109.00 frames. ], tot_loss[loss=0.1991, simple_loss=0.2814, pruned_loss=0.05843, over 1406943.64 frames. ], batch size: 62, lr: 1.17e-02, grad_scale: 8.0 +2023-03-09 00:26:44,170 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.347e+02 3.358e+02 3.908e+02 4.733e+02 8.989e+02, threshold=7.816e+02, percent-clipped=2.0 +2023-03-09 00:27:01,146 INFO [train.py:898] (0/4) Epoch 10, batch 150, loss[loss=0.2042, simple_loss=0.2831, pruned_loss=0.0626, over 18512.00 frames. ], tot_loss[loss=0.197, simple_loss=0.2805, pruned_loss=0.05681, over 1904992.42 frames. ], batch size: 47, lr: 1.17e-02, grad_scale: 8.0 +2023-03-09 00:27:31,441 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.3903, 3.5772, 5.0798, 4.1699, 3.1036, 3.0031, 4.4731, 5.1455], + device='cuda:0'), covar=tensor([0.0875, 0.1547, 0.0082, 0.0336, 0.0870, 0.1061, 0.0298, 0.0110], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0232, 0.0093, 0.0155, 0.0172, 0.0172, 0.0164, 0.0130], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 00:27:38,627 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-09 00:27:56,360 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-09 00:28:00,323 INFO [train.py:898] (0/4) Epoch 10, batch 200, loss[loss=0.2257, simple_loss=0.3092, pruned_loss=0.0711, over 16014.00 frames. ], tot_loss[loss=0.1972, simple_loss=0.2804, pruned_loss=0.05704, over 2270949.18 frames. ], batch size: 94, lr: 1.17e-02, grad_scale: 8.0 +2023-03-09 00:28:42,827 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.147e+02 3.153e+02 3.905e+02 4.894e+02 1.439e+03, threshold=7.811e+02, percent-clipped=2.0 +2023-03-09 00:28:43,231 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8282, 3.0102, 4.1237, 4.2451, 2.7534, 4.7354, 4.0362, 3.0483], + device='cuda:0'), covar=tensor([0.0376, 0.1337, 0.0216, 0.0233, 0.1453, 0.0158, 0.0425, 0.0928], + device='cuda:0'), in_proj_covar=tensor([0.0181, 0.0220, 0.0142, 0.0137, 0.0210, 0.0178, 0.0198, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 00:28:50,956 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.69 vs. limit=2.0 +2023-03-09 00:28:59,438 INFO [train.py:898] (0/4) Epoch 10, batch 250, loss[loss=0.2058, simple_loss=0.2962, pruned_loss=0.05767, over 18307.00 frames. ], tot_loss[loss=0.1962, simple_loss=0.2794, pruned_loss=0.05649, over 2568161.61 frames. ], batch size: 54, lr: 1.17e-02, grad_scale: 4.0 +2023-03-09 00:29:09,914 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=32966.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:29:16,215 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-03-09 00:29:30,832 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5497, 5.1752, 5.0649, 5.1797, 4.7223, 4.9725, 4.3662, 5.0007], + device='cuda:0'), covar=tensor([0.0247, 0.0262, 0.0218, 0.0308, 0.0322, 0.0230, 0.1166, 0.0264], + device='cuda:0'), in_proj_covar=tensor([0.0161, 0.0206, 0.0191, 0.0221, 0.0204, 0.0211, 0.0269, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0006, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-09 00:29:58,487 INFO [train.py:898] (0/4) Epoch 10, batch 300, loss[loss=0.1874, simple_loss=0.2708, pruned_loss=0.05203, over 18296.00 frames. ], tot_loss[loss=0.1966, simple_loss=0.2801, pruned_loss=0.05658, over 2797661.37 frames. ], batch size: 49, lr: 1.16e-02, grad_scale: 4.0 +2023-03-09 00:30:21,473 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33027.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:30:40,677 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.108e+02 3.425e+02 4.260e+02 4.893e+02 1.115e+03, threshold=8.520e+02, percent-clipped=1.0 +2023-03-09 00:30:57,414 INFO [train.py:898] (0/4) Epoch 10, batch 350, loss[loss=0.2175, simple_loss=0.2941, pruned_loss=0.07044, over 12573.00 frames. ], tot_loss[loss=0.1961, simple_loss=0.2798, pruned_loss=0.05623, over 2973409.01 frames. ], batch size: 129, lr: 1.16e-02, grad_scale: 4.0 +2023-03-09 00:31:11,096 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33069.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:31:40,428 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2191, 4.2285, 2.5030, 4.1770, 5.2011, 2.4646, 3.7757, 3.7676], + device='cuda:0'), covar=tensor([0.0064, 0.1020, 0.1550, 0.0483, 0.0037, 0.1372, 0.0696, 0.0785], + device='cuda:0'), in_proj_covar=tensor([0.0106, 0.0215, 0.0185, 0.0185, 0.0084, 0.0172, 0.0198, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 00:31:55,095 INFO [train.py:898] (0/4) Epoch 10, batch 400, loss[loss=0.1831, simple_loss=0.2649, pruned_loss=0.05066, over 18401.00 frames. ], tot_loss[loss=0.1958, simple_loss=0.2796, pruned_loss=0.05601, over 3103728.62 frames. ], batch size: 48, lr: 1.16e-02, grad_scale: 8.0 +2023-03-09 00:32:13,115 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.42 vs. limit=5.0 +2023-03-09 00:32:22,432 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33130.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:32:37,084 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.379e+02 3.296e+02 4.038e+02 4.885e+02 1.161e+03, threshold=8.076e+02, percent-clipped=2.0 +2023-03-09 00:32:37,596 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6753, 3.7058, 3.4659, 3.0029, 3.4934, 2.7269, 2.7976, 3.7792], + device='cuda:0'), covar=tensor([0.0034, 0.0055, 0.0065, 0.0127, 0.0069, 0.0166, 0.0157, 0.0043], + device='cuda:0'), in_proj_covar=tensor([0.0084, 0.0108, 0.0098, 0.0145, 0.0097, 0.0142, 0.0149, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 00:32:51,598 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33155.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:32:53,564 INFO [train.py:898] (0/4) Epoch 10, batch 450, loss[loss=0.1736, simple_loss=0.2502, pruned_loss=0.04849, over 18571.00 frames. ], tot_loss[loss=0.1958, simple_loss=0.2794, pruned_loss=0.05612, over 3193358.17 frames. ], batch size: 45, lr: 1.16e-02, grad_scale: 8.0 +2023-03-09 00:33:19,550 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3551, 4.3736, 2.7090, 4.3296, 5.2818, 2.5696, 3.8298, 3.8651], + device='cuda:0'), covar=tensor([0.0051, 0.0935, 0.1358, 0.0529, 0.0037, 0.1286, 0.0649, 0.0721], + device='cuda:0'), in_proj_covar=tensor([0.0105, 0.0216, 0.0183, 0.0184, 0.0084, 0.0170, 0.0196, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0003, 0.0002, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 00:33:24,734 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-03-09 00:33:52,247 INFO [train.py:898] (0/4) Epoch 10, batch 500, loss[loss=0.1696, simple_loss=0.2463, pruned_loss=0.04647, over 18413.00 frames. ], tot_loss[loss=0.1957, simple_loss=0.2793, pruned_loss=0.05605, over 3292199.95 frames. ], batch size: 43, lr: 1.16e-02, grad_scale: 8.0 +2023-03-09 00:34:03,220 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33216.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:34:33,503 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.970e+02 3.300e+02 3.836e+02 4.921e+02 1.033e+03, threshold=7.671e+02, percent-clipped=3.0 +2023-03-09 00:34:35,157 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1315, 4.0612, 5.2362, 2.9656, 4.2105, 2.6295, 2.9657, 2.2128], + device='cuda:0'), covar=tensor([0.0807, 0.0647, 0.0061, 0.0623, 0.0596, 0.2186, 0.2398, 0.1574], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0208, 0.0109, 0.0162, 0.0223, 0.0242, 0.0268, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 00:34:49,826 INFO [train.py:898] (0/4) Epoch 10, batch 550, loss[loss=0.1973, simple_loss=0.2878, pruned_loss=0.05337, over 18220.00 frames. ], tot_loss[loss=0.1952, simple_loss=0.2789, pruned_loss=0.05581, over 3362106.78 frames. ], batch size: 60, lr: 1.16e-02, grad_scale: 8.0 +2023-03-09 00:35:13,310 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.4530, 3.6875, 5.2665, 4.4993, 3.4578, 3.2569, 4.3893, 5.3107], + device='cuda:0'), covar=tensor([0.0898, 0.1468, 0.0070, 0.0297, 0.0811, 0.1003, 0.0377, 0.0174], + device='cuda:0'), in_proj_covar=tensor([0.0137, 0.0237, 0.0093, 0.0157, 0.0175, 0.0174, 0.0167, 0.0134], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 00:35:48,960 INFO [train.py:898] (0/4) Epoch 10, batch 600, loss[loss=0.1765, simple_loss=0.2609, pruned_loss=0.04607, over 18394.00 frames. ], tot_loss[loss=0.1946, simple_loss=0.2786, pruned_loss=0.05529, over 3414154.46 frames. ], batch size: 48, lr: 1.16e-02, grad_scale: 8.0 +2023-03-09 00:36:06,493 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33322.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:36:14,100 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7679, 2.3091, 2.9275, 2.9668, 3.5925, 5.3719, 4.8482, 4.3794], + device='cuda:0'), covar=tensor([0.0990, 0.1651, 0.1916, 0.1061, 0.1463, 0.0057, 0.0286, 0.0332], + device='cuda:0'), in_proj_covar=tensor([0.0229, 0.0287, 0.0297, 0.0243, 0.0355, 0.0173, 0.0248, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 00:36:26,593 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7422, 3.5593, 4.9429, 2.7841, 4.0884, 2.4849, 3.1322, 2.0045], + device='cuda:0'), covar=tensor([0.0964, 0.0796, 0.0082, 0.0688, 0.0561, 0.2223, 0.2064, 0.1674], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0210, 0.0110, 0.0163, 0.0224, 0.0244, 0.0269, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 00:36:30,341 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.217e+02 3.224e+02 3.633e+02 4.391e+02 9.720e+02, threshold=7.266e+02, percent-clipped=2.0 +2023-03-09 00:36:34,336 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.34 vs. limit=5.0 +2023-03-09 00:36:46,438 INFO [train.py:898] (0/4) Epoch 10, batch 650, loss[loss=0.2094, simple_loss=0.2869, pruned_loss=0.06591, over 18261.00 frames. ], tot_loss[loss=0.1946, simple_loss=0.2785, pruned_loss=0.05537, over 3463168.56 frames. ], batch size: 47, lr: 1.16e-02, grad_scale: 8.0 +2023-03-09 00:36:52,821 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4028, 5.9387, 5.4696, 5.7282, 5.4946, 5.4738, 6.0139, 5.9816], + device='cuda:0'), covar=tensor([0.1096, 0.0784, 0.0449, 0.0750, 0.1412, 0.0581, 0.0489, 0.0595], + device='cuda:0'), in_proj_covar=tensor([0.0500, 0.0410, 0.0306, 0.0453, 0.0605, 0.0455, 0.0579, 0.0433], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 00:37:13,535 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7714, 2.2261, 2.7958, 2.9614, 3.5152, 5.3838, 4.9007, 4.1191], + device='cuda:0'), covar=tensor([0.0967, 0.1754, 0.2147, 0.1114, 0.1608, 0.0071, 0.0275, 0.0411], + device='cuda:0'), in_proj_covar=tensor([0.0229, 0.0286, 0.0295, 0.0244, 0.0355, 0.0173, 0.0248, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 00:37:28,086 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1110, 5.1351, 2.5972, 4.9922, 4.8511, 5.1352, 4.9377, 2.6399], + device='cuda:0'), covar=tensor([0.0155, 0.0067, 0.0754, 0.0073, 0.0076, 0.0090, 0.0098, 0.1011], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0063, 0.0086, 0.0079, 0.0074, 0.0062, 0.0075, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 00:37:31,585 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5280, 3.6142, 3.3723, 3.0441, 3.2946, 2.7166, 2.6484, 3.6743], + device='cuda:0'), covar=tensor([0.0045, 0.0057, 0.0075, 0.0118, 0.0072, 0.0149, 0.0167, 0.0037], + device='cuda:0'), in_proj_covar=tensor([0.0083, 0.0108, 0.0097, 0.0144, 0.0098, 0.0140, 0.0148, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 00:37:45,672 INFO [train.py:898] (0/4) Epoch 10, batch 700, loss[loss=0.1823, simple_loss=0.2698, pruned_loss=0.04736, over 18492.00 frames. ], tot_loss[loss=0.195, simple_loss=0.2789, pruned_loss=0.05554, over 3501112.31 frames. ], batch size: 51, lr: 1.16e-02, grad_scale: 8.0 +2023-03-09 00:38:07,123 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33425.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:38:10,918 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.63 vs. limit=2.0 +2023-03-09 00:38:27,850 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.134e+02 3.209e+02 3.744e+02 4.761e+02 1.041e+03, threshold=7.488e+02, percent-clipped=6.0 +2023-03-09 00:38:44,051 INFO [train.py:898] (0/4) Epoch 10, batch 750, loss[loss=0.2137, simple_loss=0.3099, pruned_loss=0.0587, over 18632.00 frames. ], tot_loss[loss=0.1946, simple_loss=0.2785, pruned_loss=0.05528, over 3528025.48 frames. ], batch size: 52, lr: 1.16e-02, grad_scale: 8.0 +2023-03-09 00:39:01,829 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33472.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:39:42,272 INFO [train.py:898] (0/4) Epoch 10, batch 800, loss[loss=0.1863, simple_loss=0.2609, pruned_loss=0.0559, over 18279.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2785, pruned_loss=0.05563, over 3532981.95 frames. ], batch size: 45, lr: 1.16e-02, grad_scale: 8.0 +2023-03-09 00:39:47,430 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33511.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:40:13,206 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33533.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:40:24,513 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.088e+02 3.240e+02 4.102e+02 4.783e+02 1.200e+03, threshold=8.205e+02, percent-clipped=1.0 +2023-03-09 00:40:40,563 INFO [train.py:898] (0/4) Epoch 10, batch 850, loss[loss=0.2147, simple_loss=0.295, pruned_loss=0.06724, over 16023.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2786, pruned_loss=0.05558, over 3547031.49 frames. ], batch size: 94, lr: 1.16e-02, grad_scale: 8.0 +2023-03-09 00:41:39,909 INFO [train.py:898] (0/4) Epoch 10, batch 900, loss[loss=0.2079, simple_loss=0.2944, pruned_loss=0.06066, over 18208.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2791, pruned_loss=0.05534, over 3564288.53 frames. ], batch size: 60, lr: 1.15e-02, grad_scale: 8.0 +2023-03-09 00:41:57,523 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=33622.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:42:22,395 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.227e+02 3.532e+02 4.038e+02 4.740e+02 8.485e+02, threshold=8.075e+02, percent-clipped=1.0 +2023-03-09 00:42:38,550 INFO [train.py:898] (0/4) Epoch 10, batch 950, loss[loss=0.252, simple_loss=0.3169, pruned_loss=0.09359, over 12796.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.2793, pruned_loss=0.05521, over 3574941.06 frames. ], batch size: 129, lr: 1.15e-02, grad_scale: 8.0 +2023-03-09 00:42:53,705 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=33670.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:43:01,707 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33677.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:43:33,062 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-09 00:43:36,486 INFO [train.py:898] (0/4) Epoch 10, batch 1000, loss[loss=0.1742, simple_loss=0.2563, pruned_loss=0.04604, over 18492.00 frames. ], tot_loss[loss=0.1949, simple_loss=0.279, pruned_loss=0.05536, over 3564536.17 frames. ], batch size: 44, lr: 1.15e-02, grad_scale: 8.0 +2023-03-09 00:43:57,444 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=33725.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:44:13,314 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=33738.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 00:44:18,699 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.319e+02 3.380e+02 3.936e+02 4.816e+02 9.502e+02, threshold=7.872e+02, percent-clipped=1.0 +2023-03-09 00:44:35,204 INFO [train.py:898] (0/4) Epoch 10, batch 1050, loss[loss=0.1992, simple_loss=0.2781, pruned_loss=0.06016, over 18405.00 frames. ], tot_loss[loss=0.1952, simple_loss=0.2793, pruned_loss=0.05554, over 3564429.20 frames. ], batch size: 52, lr: 1.15e-02, grad_scale: 8.0 +2023-03-09 00:44:53,659 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=33773.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:45:34,466 INFO [train.py:898] (0/4) Epoch 10, batch 1100, loss[loss=0.2039, simple_loss=0.2867, pruned_loss=0.06052, over 18485.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2782, pruned_loss=0.05508, over 3574468.20 frames. ], batch size: 51, lr: 1.15e-02, grad_scale: 8.0 +2023-03-09 00:45:39,197 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=33811.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:45:58,743 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=33828.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:46:16,688 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.260e+02 3.207e+02 4.126e+02 4.735e+02 1.316e+03, threshold=8.252e+02, percent-clipped=4.0 +2023-03-09 00:46:25,764 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-09 00:46:32,948 INFO [train.py:898] (0/4) Epoch 10, batch 1150, loss[loss=0.1786, simple_loss=0.2585, pruned_loss=0.0494, over 18250.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2786, pruned_loss=0.05545, over 3579981.09 frames. ], batch size: 45, lr: 1.15e-02, grad_scale: 8.0 +2023-03-09 00:46:36,022 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=33859.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:47:10,551 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5344, 2.0138, 2.4411, 2.5831, 3.2392, 5.1655, 4.6534, 4.0037], + device='cuda:0'), covar=tensor([0.1121, 0.2003, 0.2502, 0.1375, 0.1861, 0.0073, 0.0304, 0.0403], + device='cuda:0'), in_proj_covar=tensor([0.0229, 0.0287, 0.0297, 0.0243, 0.0356, 0.0174, 0.0248, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 00:47:22,070 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-09 00:47:31,463 INFO [train.py:898] (0/4) Epoch 10, batch 1200, loss[loss=0.2357, simple_loss=0.3206, pruned_loss=0.07545, over 16049.00 frames. ], tot_loss[loss=0.1953, simple_loss=0.279, pruned_loss=0.05579, over 3577090.38 frames. ], batch size: 95, lr: 1.15e-02, grad_scale: 8.0 +2023-03-09 00:47:54,066 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8565, 3.8843, 4.9772, 2.8634, 4.1469, 2.6693, 2.8801, 1.9683], + device='cuda:0'), covar=tensor([0.0925, 0.0710, 0.0078, 0.0718, 0.0587, 0.2216, 0.2688, 0.1812], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0211, 0.0111, 0.0163, 0.0225, 0.0243, 0.0274, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 00:48:01,316 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-09 00:48:09,160 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=33939.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:48:13,373 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.163e+02 3.321e+02 3.896e+02 4.849e+02 9.830e+02, threshold=7.793e+02, percent-clipped=3.0 +2023-03-09 00:48:30,003 INFO [train.py:898] (0/4) Epoch 10, batch 1250, loss[loss=0.2082, simple_loss=0.2918, pruned_loss=0.06227, over 18003.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2783, pruned_loss=0.05564, over 3584395.71 frames. ], batch size: 65, lr: 1.15e-02, grad_scale: 8.0 +2023-03-09 00:48:46,809 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8306, 3.8310, 3.6599, 3.1480, 3.6314, 2.9932, 2.8766, 3.7888], + device='cuda:0'), covar=tensor([0.0030, 0.0053, 0.0055, 0.0096, 0.0065, 0.0134, 0.0146, 0.0045], + device='cuda:0'), in_proj_covar=tensor([0.0084, 0.0109, 0.0098, 0.0146, 0.0098, 0.0142, 0.0151, 0.0081], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 00:49:18,392 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4519, 3.1442, 1.7260, 4.2271, 2.7129, 4.0273, 1.9184, 3.5804], + device='cuda:0'), covar=tensor([0.0535, 0.0931, 0.1526, 0.0389, 0.0943, 0.0330, 0.1450, 0.0461], + device='cuda:0'), in_proj_covar=tensor([0.0182, 0.0208, 0.0172, 0.0225, 0.0176, 0.0227, 0.0186, 0.0179], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 00:49:19,529 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-34000.pt +2023-03-09 00:49:24,458 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34000.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:49:32,599 INFO [train.py:898] (0/4) Epoch 10, batch 1300, loss[loss=0.1974, simple_loss=0.2841, pruned_loss=0.05534, over 18389.00 frames. ], tot_loss[loss=0.1945, simple_loss=0.2782, pruned_loss=0.05541, over 3586680.39 frames. ], batch size: 52, lr: 1.15e-02, grad_scale: 8.0 +2023-03-09 00:49:34,112 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1166, 5.0832, 4.6900, 5.0300, 5.0015, 4.4799, 4.9611, 4.6741], + device='cuda:0'), covar=tensor([0.0312, 0.0392, 0.1206, 0.0649, 0.0470, 0.0348, 0.0354, 0.0861], + device='cuda:0'), in_proj_covar=tensor([0.0382, 0.0448, 0.0590, 0.0349, 0.0331, 0.0403, 0.0429, 0.0556], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 00:50:02,497 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34033.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 00:50:14,325 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.927e+02 3.198e+02 3.729e+02 4.411e+02 1.072e+03, threshold=7.459e+02, percent-clipped=4.0 +2023-03-09 00:50:15,950 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34044.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:50:30,793 INFO [train.py:898] (0/4) Epoch 10, batch 1350, loss[loss=0.1789, simple_loss=0.2623, pruned_loss=0.04778, over 18561.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2774, pruned_loss=0.05495, over 3588043.53 frames. ], batch size: 49, lr: 1.15e-02, grad_scale: 8.0 +2023-03-09 00:51:27,363 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34105.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:51:29,261 INFO [train.py:898] (0/4) Epoch 10, batch 1400, loss[loss=0.234, simple_loss=0.3034, pruned_loss=0.08232, over 12514.00 frames. ], tot_loss[loss=0.1929, simple_loss=0.2766, pruned_loss=0.05456, over 3591482.34 frames. ], batch size: 129, lr: 1.15e-02, grad_scale: 8.0 +2023-03-09 00:51:54,979 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34128.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:52:11,995 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.227e+02 3.479e+02 4.188e+02 5.259e+02 1.052e+03, threshold=8.376e+02, percent-clipped=5.0 +2023-03-09 00:52:15,224 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34145.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:52:28,422 INFO [train.py:898] (0/4) Epoch 10, batch 1450, loss[loss=0.2062, simple_loss=0.2903, pruned_loss=0.06106, over 17853.00 frames. ], tot_loss[loss=0.1929, simple_loss=0.2766, pruned_loss=0.05459, over 3602013.60 frames. ], batch size: 70, lr: 1.15e-02, grad_scale: 8.0 +2023-03-09 00:52:50,503 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=34176.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:53:26,236 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34206.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:53:26,962 INFO [train.py:898] (0/4) Epoch 10, batch 1500, loss[loss=0.1663, simple_loss=0.2479, pruned_loss=0.04234, over 18427.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.2763, pruned_loss=0.05423, over 3609031.01 frames. ], batch size: 43, lr: 1.14e-02, grad_scale: 8.0 +2023-03-09 00:53:45,089 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8650, 5.4229, 5.4227, 5.4527, 5.0149, 5.3369, 4.7118, 5.3463], + device='cuda:0'), covar=tensor([0.0233, 0.0244, 0.0172, 0.0228, 0.0277, 0.0202, 0.1102, 0.0251], + device='cuda:0'), in_proj_covar=tensor([0.0166, 0.0210, 0.0198, 0.0229, 0.0208, 0.0215, 0.0274, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0006, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-09 00:54:08,715 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.894e+02 3.400e+02 4.165e+02 5.471e+02 1.005e+03, threshold=8.329e+02, percent-clipped=3.0 +2023-03-09 00:54:25,078 INFO [train.py:898] (0/4) Epoch 10, batch 1550, loss[loss=0.2015, simple_loss=0.2818, pruned_loss=0.06064, over 16141.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.276, pruned_loss=0.05431, over 3610617.30 frames. ], batch size: 94, lr: 1.14e-02, grad_scale: 8.0 +2023-03-09 00:55:09,414 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34295.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:55:23,438 INFO [train.py:898] (0/4) Epoch 10, batch 1600, loss[loss=0.1988, simple_loss=0.2856, pruned_loss=0.05595, over 18619.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.276, pruned_loss=0.05462, over 3609336.97 frames. ], batch size: 52, lr: 1.14e-02, grad_scale: 8.0 +2023-03-09 00:55:48,320 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6614, 1.9872, 2.8227, 2.7318, 3.5252, 5.2210, 4.6196, 4.0503], + device='cuda:0'), covar=tensor([0.1116, 0.1971, 0.2166, 0.1254, 0.1489, 0.0102, 0.0340, 0.0441], + device='cuda:0'), in_proj_covar=tensor([0.0228, 0.0285, 0.0293, 0.0241, 0.0354, 0.0174, 0.0251, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 00:55:53,832 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34333.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:55:57,786 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.83 vs. limit=2.0 +2023-03-09 00:56:05,596 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.752e+02 3.318e+02 3.747e+02 4.447e+02 1.015e+03, threshold=7.495e+02, percent-clipped=2.0 +2023-03-09 00:56:20,933 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7557, 4.8309, 4.8123, 4.8172, 4.7926, 5.3590, 4.9452, 4.7985], + device='cuda:0'), covar=tensor([0.0929, 0.0719, 0.0694, 0.0614, 0.1093, 0.0703, 0.0674, 0.1415], + device='cuda:0'), in_proj_covar=tensor([0.0300, 0.0227, 0.0242, 0.0240, 0.0276, 0.0338, 0.0223, 0.0332], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 00:56:21,842 INFO [train.py:898] (0/4) Epoch 10, batch 1650, loss[loss=0.2018, simple_loss=0.2849, pruned_loss=0.05937, over 16107.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2765, pruned_loss=0.0548, over 3599382.34 frames. ], batch size: 95, lr: 1.14e-02, grad_scale: 8.0 +2023-03-09 00:56:49,662 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=34381.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:57:11,885 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34400.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:57:20,095 INFO [train.py:898] (0/4) Epoch 10, batch 1700, loss[loss=0.1914, simple_loss=0.2815, pruned_loss=0.05062, over 18471.00 frames. ], tot_loss[loss=0.1935, simple_loss=0.277, pruned_loss=0.05502, over 3597190.57 frames. ], batch size: 59, lr: 1.14e-02, grad_scale: 4.0 +2023-03-09 00:58:03,526 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.339e+02 3.311e+02 3.796e+02 4.465e+02 1.114e+03, threshold=7.593e+02, percent-clipped=2.0 +2023-03-09 00:58:18,863 INFO [train.py:898] (0/4) Epoch 10, batch 1750, loss[loss=0.1769, simple_loss=0.2576, pruned_loss=0.04806, over 18489.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.277, pruned_loss=0.05491, over 3589801.51 frames. ], batch size: 47, lr: 1.14e-02, grad_scale: 4.0 +2023-03-09 00:58:19,563 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-09 00:59:11,448 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34501.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 00:59:15,432 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-09 00:59:18,088 INFO [train.py:898] (0/4) Epoch 10, batch 1800, loss[loss=0.1977, simple_loss=0.2838, pruned_loss=0.05576, over 16224.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2769, pruned_loss=0.05457, over 3591625.28 frames. ], batch size: 94, lr: 1.14e-02, grad_scale: 4.0 +2023-03-09 00:59:53,805 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1001, 5.0879, 2.4816, 4.8833, 4.7956, 5.0468, 4.8618, 2.5085], + device='cuda:0'), covar=tensor([0.0147, 0.0052, 0.0836, 0.0082, 0.0068, 0.0080, 0.0097, 0.1041], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0064, 0.0088, 0.0081, 0.0075, 0.0064, 0.0075, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0005, 0.0005, 0.0004, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 01:00:01,445 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.848e+02 3.049e+02 3.588e+02 4.503e+02 1.027e+03, threshold=7.176e+02, percent-clipped=5.0 +2023-03-09 01:00:13,958 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3472, 4.3470, 2.4555, 4.2791, 5.3021, 2.5305, 3.9402, 3.9742], + device='cuda:0'), covar=tensor([0.0074, 0.0897, 0.1565, 0.0511, 0.0044, 0.1264, 0.0592, 0.0760], + device='cuda:0'), in_proj_covar=tensor([0.0108, 0.0222, 0.0186, 0.0188, 0.0085, 0.0172, 0.0199, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 01:00:16,675 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-09 01:00:16,828 INFO [train.py:898] (0/4) Epoch 10, batch 1850, loss[loss=0.2221, simple_loss=0.3041, pruned_loss=0.07012, over 17037.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2769, pruned_loss=0.05458, over 3588209.71 frames. ], batch size: 78, lr: 1.14e-02, grad_scale: 4.0 +2023-03-09 01:00:53,994 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2432, 4.2088, 2.4248, 4.2643, 5.2019, 2.2897, 3.7213, 4.0191], + device='cuda:0'), covar=tensor([0.0088, 0.1184, 0.1703, 0.0522, 0.0051, 0.1476, 0.0771, 0.0784], + device='cuda:0'), in_proj_covar=tensor([0.0107, 0.0220, 0.0184, 0.0186, 0.0085, 0.0170, 0.0197, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0002, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 01:00:59,486 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34593.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:00:59,877 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.52 vs. limit=5.0 +2023-03-09 01:01:01,698 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34595.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:01:13,787 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5691, 2.0096, 2.6522, 2.6378, 3.3068, 4.9456, 4.3873, 4.1411], + device='cuda:0'), covar=tensor([0.1130, 0.1949, 0.2241, 0.1372, 0.1750, 0.0122, 0.0371, 0.0375], + device='cuda:0'), in_proj_covar=tensor([0.0229, 0.0286, 0.0294, 0.0242, 0.0352, 0.0176, 0.0249, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 01:01:15,604 INFO [train.py:898] (0/4) Epoch 10, batch 1900, loss[loss=0.1987, simple_loss=0.2803, pruned_loss=0.0585, over 18381.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.277, pruned_loss=0.05488, over 3584840.28 frames. ], batch size: 50, lr: 1.14e-02, grad_scale: 4.0 +2023-03-09 01:01:33,763 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34622.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:01:38,137 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34626.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:01:58,494 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=34643.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:01:59,387 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.282e+02 3.244e+02 3.774e+02 4.780e+02 1.001e+03, threshold=7.549e+02, percent-clipped=4.0 +2023-03-09 01:02:11,289 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34654.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:02:14,314 INFO [train.py:898] (0/4) Epoch 10, batch 1950, loss[loss=0.1988, simple_loss=0.2829, pruned_loss=0.05729, over 18300.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2768, pruned_loss=0.05483, over 3584917.62 frames. ], batch size: 49, lr: 1.14e-02, grad_scale: 4.0 +2023-03-09 01:02:14,780 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5499, 3.5077, 4.9840, 4.0729, 3.2162, 2.8637, 4.2581, 5.0866], + device='cuda:0'), covar=tensor([0.0825, 0.1704, 0.0090, 0.0368, 0.0875, 0.1109, 0.0319, 0.0136], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0240, 0.0097, 0.0158, 0.0176, 0.0174, 0.0168, 0.0137], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 01:02:16,946 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3656, 5.9136, 5.3998, 5.6975, 5.4509, 5.3895, 5.9600, 5.8921], + device='cuda:0'), covar=tensor([0.1221, 0.0619, 0.0545, 0.0654, 0.1421, 0.0675, 0.0530, 0.0569], + device='cuda:0'), in_proj_covar=tensor([0.0523, 0.0424, 0.0329, 0.0468, 0.0643, 0.0469, 0.0602, 0.0443], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 01:02:44,906 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34683.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 01:02:49,885 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34687.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:03:05,169 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34700.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:03:12,471 INFO [train.py:898] (0/4) Epoch 10, batch 2000, loss[loss=0.193, simple_loss=0.2743, pruned_loss=0.05585, over 18294.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.2771, pruned_loss=0.05474, over 3596837.00 frames. ], batch size: 49, lr: 1.14e-02, grad_scale: 8.0 +2023-03-09 01:03:16,340 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=34710.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 01:03:56,812 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.230e+02 3.366e+02 4.037e+02 4.949e+02 9.171e+02, threshold=8.073e+02, percent-clipped=4.0 +2023-03-09 01:04:01,480 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=34748.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:04:11,483 INFO [train.py:898] (0/4) Epoch 10, batch 2050, loss[loss=0.1905, simple_loss=0.2775, pruned_loss=0.05174, over 18300.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.2765, pruned_loss=0.05446, over 3603079.08 frames. ], batch size: 54, lr: 1.14e-02, grad_scale: 8.0 +2023-03-09 01:04:14,136 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2436, 5.3521, 2.7591, 5.1504, 5.0459, 5.3777, 5.1827, 2.6080], + device='cuda:0'), covar=tensor([0.0129, 0.0043, 0.0732, 0.0073, 0.0054, 0.0049, 0.0078, 0.0976], + device='cuda:0'), in_proj_covar=tensor([0.0073, 0.0063, 0.0086, 0.0078, 0.0072, 0.0063, 0.0073, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 01:04:28,557 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=34771.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 01:05:04,161 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=34801.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:05:10,833 INFO [train.py:898] (0/4) Epoch 10, batch 2100, loss[loss=0.1894, simple_loss=0.2731, pruned_loss=0.05287, over 18289.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2759, pruned_loss=0.05435, over 3595511.32 frames. ], batch size: 49, lr: 1.14e-02, grad_scale: 8.0 +2023-03-09 01:05:54,598 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.339e+02 3.283e+02 3.862e+02 4.779e+02 1.024e+03, threshold=7.723e+02, percent-clipped=2.0 +2023-03-09 01:06:01,029 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=34849.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:06:09,706 INFO [train.py:898] (0/4) Epoch 10, batch 2150, loss[loss=0.2003, simple_loss=0.2877, pruned_loss=0.05646, over 18308.00 frames. ], tot_loss[loss=0.1924, simple_loss=0.276, pruned_loss=0.05439, over 3592004.97 frames. ], batch size: 57, lr: 1.13e-02, grad_scale: 8.0 +2023-03-09 01:07:01,058 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2598, 4.9780, 5.3521, 5.4263, 5.1241, 5.9649, 5.6833, 5.2505], + device='cuda:0'), covar=tensor([0.0856, 0.0646, 0.0714, 0.0538, 0.1403, 0.0613, 0.0466, 0.1582], + device='cuda:0'), in_proj_covar=tensor([0.0296, 0.0222, 0.0240, 0.0236, 0.0276, 0.0330, 0.0220, 0.0326], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 01:07:07,216 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.01 vs. limit=2.0 +2023-03-09 01:07:07,668 INFO [train.py:898] (0/4) Epoch 10, batch 2200, loss[loss=0.2232, simple_loss=0.3018, pruned_loss=0.07227, over 12679.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.2765, pruned_loss=0.05438, over 3598115.25 frames. ], batch size: 131, lr: 1.13e-02, grad_scale: 8.0 +2023-03-09 01:07:18,952 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6253, 2.7056, 2.7099, 2.6849, 3.6441, 3.5442, 3.1004, 3.0076], + device='cuda:0'), covar=tensor([0.0170, 0.0261, 0.0525, 0.0366, 0.0121, 0.0156, 0.0340, 0.0313], + device='cuda:0'), in_proj_covar=tensor([0.0116, 0.0104, 0.0152, 0.0135, 0.0099, 0.0086, 0.0132, 0.0127], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 01:07:39,183 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-09 01:07:45,795 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6346, 4.6409, 4.7835, 4.4590, 4.3635, 4.4701, 4.8435, 4.8165], + device='cuda:0'), covar=tensor([0.0062, 0.0065, 0.0056, 0.0090, 0.0071, 0.0103, 0.0080, 0.0091], + device='cuda:0'), in_proj_covar=tensor([0.0077, 0.0055, 0.0057, 0.0071, 0.0060, 0.0083, 0.0070, 0.0068], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 01:07:50,021 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.072e+02 3.371e+02 4.058e+02 4.999e+02 1.282e+03, threshold=8.115e+02, percent-clipped=5.0 +2023-03-09 01:07:57,102 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34949.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:08:05,878 INFO [train.py:898] (0/4) Epoch 10, batch 2250, loss[loss=0.1874, simple_loss=0.2767, pruned_loss=0.0491, over 18379.00 frames. ], tot_loss[loss=0.1933, simple_loss=0.2771, pruned_loss=0.05475, over 3593485.12 frames. ], batch size: 52, lr: 1.13e-02, grad_scale: 8.0 +2023-03-09 01:08:29,640 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34978.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 01:08:34,770 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=34982.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:08:43,343 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8705, 4.8262, 5.0041, 4.5842, 4.4716, 4.7427, 5.0860, 5.0036], + device='cuda:0'), covar=tensor([0.0063, 0.0087, 0.0059, 0.0109, 0.0072, 0.0097, 0.0087, 0.0108], + device='cuda:0'), in_proj_covar=tensor([0.0077, 0.0055, 0.0057, 0.0072, 0.0061, 0.0083, 0.0070, 0.0069], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 01:09:05,056 INFO [train.py:898] (0/4) Epoch 10, batch 2300, loss[loss=0.1585, simple_loss=0.2329, pruned_loss=0.04202, over 18368.00 frames. ], tot_loss[loss=0.1945, simple_loss=0.2781, pruned_loss=0.05543, over 3563421.40 frames. ], batch size: 42, lr: 1.13e-02, grad_scale: 8.0 +2023-03-09 01:09:42,948 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0522, 3.3809, 2.5198, 3.3279, 4.0486, 2.4150, 3.2102, 3.2908], + device='cuda:0'), covar=tensor([0.0119, 0.1020, 0.1327, 0.0593, 0.0077, 0.1190, 0.0741, 0.0793], + device='cuda:0'), in_proj_covar=tensor([0.0107, 0.0220, 0.0185, 0.0186, 0.0085, 0.0169, 0.0197, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 01:09:46,277 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6193, 2.7393, 4.2393, 4.1004, 2.5094, 4.6072, 3.9550, 2.6834], + device='cuda:0'), covar=tensor([0.0383, 0.1323, 0.0215, 0.0196, 0.1408, 0.0138, 0.0359, 0.1002], + device='cuda:0'), in_proj_covar=tensor([0.0181, 0.0212, 0.0142, 0.0136, 0.0207, 0.0175, 0.0198, 0.0187], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 01:09:48,163 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.223e+02 3.414e+02 4.118e+02 5.165e+02 1.398e+03, threshold=8.236e+02, percent-clipped=7.0 +2023-03-09 01:10:04,324 INFO [train.py:898] (0/4) Epoch 10, batch 2350, loss[loss=0.2276, simple_loss=0.3115, pruned_loss=0.07188, over 18088.00 frames. ], tot_loss[loss=0.1948, simple_loss=0.2785, pruned_loss=0.05552, over 3569100.97 frames. ], batch size: 62, lr: 1.13e-02, grad_scale: 8.0 +2023-03-09 01:10:14,497 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35066.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 01:10:40,167 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.32 vs. limit=5.0 +2023-03-09 01:10:45,252 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35092.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:11:03,094 INFO [train.py:898] (0/4) Epoch 10, batch 2400, loss[loss=0.1854, simple_loss=0.2664, pruned_loss=0.05225, over 18404.00 frames. ], tot_loss[loss=0.194, simple_loss=0.2777, pruned_loss=0.05513, over 3577310.77 frames. ], batch size: 48, lr: 1.13e-02, grad_scale: 8.0 +2023-03-09 01:11:20,339 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9702, 5.0546, 5.1281, 5.1027, 4.9716, 5.7354, 5.4178, 5.0035], + device='cuda:0'), covar=tensor([0.1104, 0.0689, 0.0748, 0.0729, 0.1576, 0.0808, 0.0577, 0.1728], + device='cuda:0'), in_proj_covar=tensor([0.0304, 0.0229, 0.0246, 0.0244, 0.0283, 0.0342, 0.0224, 0.0332], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 01:11:34,303 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9623, 4.4876, 4.5665, 3.2701, 3.7149, 3.8394, 2.5968, 2.3062], + device='cuda:0'), covar=tensor([0.0170, 0.0170, 0.0065, 0.0294, 0.0303, 0.0141, 0.0763, 0.0856], + device='cuda:0'), in_proj_covar=tensor([0.0057, 0.0045, 0.0045, 0.0056, 0.0078, 0.0055, 0.0071, 0.0076], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 01:11:46,353 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.146e+02 3.097e+02 3.497e+02 4.481e+02 9.242e+02, threshold=6.993e+02, percent-clipped=1.0 +2023-03-09 01:11:57,091 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35153.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:12:02,050 INFO [train.py:898] (0/4) Epoch 10, batch 2450, loss[loss=0.1935, simple_loss=0.2755, pruned_loss=0.05578, over 18141.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2784, pruned_loss=0.05523, over 3574451.78 frames. ], batch size: 62, lr: 1.13e-02, grad_scale: 8.0 +2023-03-09 01:12:02,469 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2349, 4.2218, 2.4914, 4.0410, 5.2434, 2.3479, 3.7211, 3.8276], + device='cuda:0'), covar=tensor([0.0074, 0.1023, 0.1432, 0.0568, 0.0044, 0.1309, 0.0656, 0.0720], + device='cuda:0'), in_proj_covar=tensor([0.0109, 0.0223, 0.0189, 0.0189, 0.0086, 0.0173, 0.0200, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 01:12:15,366 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-09 01:12:32,415 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.57 vs. limit=5.0 +2023-03-09 01:13:00,744 INFO [train.py:898] (0/4) Epoch 10, batch 2500, loss[loss=0.1744, simple_loss=0.2522, pruned_loss=0.04836, over 17804.00 frames. ], tot_loss[loss=0.195, simple_loss=0.279, pruned_loss=0.05553, over 3572995.64 frames. ], batch size: 39, lr: 1.13e-02, grad_scale: 8.0 +2023-03-09 01:13:06,992 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-03-09 01:13:43,835 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.285e+02 3.181e+02 4.060e+02 4.793e+02 9.479e+02, threshold=8.119e+02, percent-clipped=6.0 +2023-03-09 01:13:50,144 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35249.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:13:58,867 INFO [train.py:898] (0/4) Epoch 10, batch 2550, loss[loss=0.1807, simple_loss=0.2704, pruned_loss=0.04551, over 18632.00 frames. ], tot_loss[loss=0.1944, simple_loss=0.2785, pruned_loss=0.0551, over 3589657.12 frames. ], batch size: 52, lr: 1.13e-02, grad_scale: 8.0 +2023-03-09 01:14:23,534 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35278.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 01:14:27,838 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35282.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:14:45,831 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=35297.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:14:57,515 INFO [train.py:898] (0/4) Epoch 10, batch 2600, loss[loss=0.2218, simple_loss=0.3077, pruned_loss=0.06796, over 18377.00 frames. ], tot_loss[loss=0.1942, simple_loss=0.2782, pruned_loss=0.05507, over 3590670.66 frames. ], batch size: 56, lr: 1.13e-02, grad_scale: 8.0 +2023-03-09 01:15:17,415 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6533, 2.7889, 2.6991, 2.9028, 3.6707, 3.5107, 3.1723, 3.0869], + device='cuda:0'), covar=tensor([0.0177, 0.0346, 0.0597, 0.0394, 0.0179, 0.0200, 0.0366, 0.0332], + device='cuda:0'), in_proj_covar=tensor([0.0116, 0.0105, 0.0151, 0.0136, 0.0101, 0.0085, 0.0134, 0.0128], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 01:15:20,600 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=35326.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:15:25,250 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=35330.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:15:40,654 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.265e+02 3.407e+02 3.893e+02 4.729e+02 1.117e+03, threshold=7.786e+02, percent-clipped=1.0 +2023-03-09 01:15:42,195 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35345.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 01:15:52,426 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.4423, 3.5751, 5.1846, 4.4367, 3.1221, 3.0956, 4.5095, 5.3438], + device='cuda:0'), covar=tensor([0.0837, 0.1702, 0.0069, 0.0325, 0.0906, 0.1005, 0.0312, 0.0112], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0236, 0.0098, 0.0159, 0.0175, 0.0172, 0.0168, 0.0137], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 01:15:56,579 INFO [train.py:898] (0/4) Epoch 10, batch 2650, loss[loss=0.1808, simple_loss=0.2628, pruned_loss=0.04943, over 18533.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.277, pruned_loss=0.05486, over 3580116.87 frames. ], batch size: 49, lr: 1.13e-02, grad_scale: 8.0 +2023-03-09 01:16:07,604 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35366.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 01:16:17,218 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35374.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:16:54,588 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35406.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 01:16:55,350 INFO [train.py:898] (0/4) Epoch 10, batch 2700, loss[loss=0.1952, simple_loss=0.2842, pruned_loss=0.05303, over 17052.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2774, pruned_loss=0.055, over 3580967.91 frames. ], batch size: 78, lr: 1.13e-02, grad_scale: 8.0 +2023-03-09 01:17:03,989 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=35414.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 01:17:04,036 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0213, 5.0604, 5.2397, 4.9928, 5.0357, 5.7176, 5.2988, 5.0030], + device='cuda:0'), covar=tensor([0.0940, 0.0658, 0.0668, 0.0704, 0.1360, 0.0728, 0.0684, 0.1694], + device='cuda:0'), in_proj_covar=tensor([0.0303, 0.0228, 0.0245, 0.0245, 0.0282, 0.0343, 0.0227, 0.0334], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 01:17:28,426 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35435.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:17:37,964 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.429e+02 3.322e+02 4.284e+02 5.081e+02 8.413e+02, threshold=8.569e+02, percent-clipped=3.0 +2023-03-09 01:17:42,697 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35448.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:17:53,629 INFO [train.py:898] (0/4) Epoch 10, batch 2750, loss[loss=0.1647, simple_loss=0.2434, pruned_loss=0.04305, over 18484.00 frames. ], tot_loss[loss=0.1939, simple_loss=0.2774, pruned_loss=0.05521, over 3578250.52 frames. ], batch size: 44, lr: 1.12e-02, grad_scale: 8.0 +2023-03-09 01:17:59,410 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-03-09 01:18:19,177 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.48 vs. limit=2.0 +2023-03-09 01:18:51,868 INFO [train.py:898] (0/4) Epoch 10, batch 2800, loss[loss=0.209, simple_loss=0.2915, pruned_loss=0.0633, over 18498.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2767, pruned_loss=0.05463, over 3583523.47 frames. ], batch size: 53, lr: 1.12e-02, grad_scale: 8.0 +2023-03-09 01:19:18,846 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35530.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:19:34,531 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.967e+02 3.480e+02 4.448e+02 5.529e+02 1.797e+03, threshold=8.895e+02, percent-clipped=6.0 +2023-03-09 01:19:49,762 INFO [train.py:898] (0/4) Epoch 10, batch 2850, loss[loss=0.1931, simple_loss=0.2731, pruned_loss=0.05658, over 16195.00 frames. ], tot_loss[loss=0.1937, simple_loss=0.2773, pruned_loss=0.05512, over 3576154.13 frames. ], batch size: 94, lr: 1.12e-02, grad_scale: 8.0 +2023-03-09 01:20:29,945 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35591.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:20:43,403 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4184, 5.3853, 5.0231, 5.3725, 5.3935, 4.7945, 5.2799, 5.0088], + device='cuda:0'), covar=tensor([0.0364, 0.0401, 0.1205, 0.0598, 0.0451, 0.0369, 0.0343, 0.0965], + device='cuda:0'), in_proj_covar=tensor([0.0395, 0.0449, 0.0594, 0.0348, 0.0336, 0.0408, 0.0438, 0.0569], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 01:20:47,591 INFO [train.py:898] (0/4) Epoch 10, batch 2900, loss[loss=0.186, simple_loss=0.267, pruned_loss=0.05257, over 18372.00 frames. ], tot_loss[loss=0.1939, simple_loss=0.2775, pruned_loss=0.05521, over 3578403.57 frames. ], batch size: 46, lr: 1.12e-02, grad_scale: 8.0 +2023-03-09 01:21:29,080 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7098, 2.9094, 4.2834, 3.7398, 2.6450, 4.6344, 3.9255, 2.8280], + device='cuda:0'), covar=tensor([0.0411, 0.1350, 0.0205, 0.0324, 0.1598, 0.0166, 0.0443, 0.1016], + device='cuda:0'), in_proj_covar=tensor([0.0180, 0.0213, 0.0145, 0.0138, 0.0211, 0.0180, 0.0200, 0.0188], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 01:21:32,209 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.306e+02 3.157e+02 3.765e+02 4.717e+02 8.382e+02, threshold=7.531e+02, percent-clipped=0.0 +2023-03-09 01:21:47,256 INFO [train.py:898] (0/4) Epoch 10, batch 2950, loss[loss=0.2113, simple_loss=0.2956, pruned_loss=0.06346, over 18500.00 frames. ], tot_loss[loss=0.1939, simple_loss=0.278, pruned_loss=0.05493, over 3578666.44 frames. ], batch size: 59, lr: 1.12e-02, grad_scale: 8.0 +2023-03-09 01:22:39,385 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35701.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 01:22:46,135 INFO [train.py:898] (0/4) Epoch 10, batch 3000, loss[loss=0.1998, simple_loss=0.2883, pruned_loss=0.05565, over 18121.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2775, pruned_loss=0.05445, over 3583034.45 frames. ], batch size: 62, lr: 1.12e-02, grad_scale: 8.0 +2023-03-09 01:22:46,138 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 01:22:58,182 INFO [train.py:932] (0/4) Epoch 10, validation: loss=0.1597, simple_loss=0.2619, pruned_loss=0.0287, over 944034.00 frames. +2023-03-09 01:22:58,183 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-09 01:23:25,157 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35730.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:23:40,660 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.222e+02 3.466e+02 3.980e+02 4.724e+02 8.688e+02, threshold=7.961e+02, percent-clipped=2.0 +2023-03-09 01:23:43,470 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4151, 3.2046, 1.8416, 4.1483, 2.8429, 4.1794, 2.2023, 3.6726], + device='cuda:0'), covar=tensor([0.0562, 0.0856, 0.1612, 0.0447, 0.0983, 0.0232, 0.1160, 0.0416], + device='cuda:0'), in_proj_covar=tensor([0.0188, 0.0210, 0.0177, 0.0231, 0.0180, 0.0232, 0.0189, 0.0182], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 01:23:44,447 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7048, 4.6845, 4.2882, 4.5374, 4.6822, 4.0954, 4.6251, 4.3048], + device='cuda:0'), covar=tensor([0.0376, 0.0507, 0.1382, 0.0908, 0.0521, 0.0422, 0.0360, 0.1007], + device='cuda:0'), in_proj_covar=tensor([0.0398, 0.0451, 0.0599, 0.0353, 0.0340, 0.0412, 0.0441, 0.0574], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 01:23:45,589 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=35748.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:23:56,264 INFO [train.py:898] (0/4) Epoch 10, batch 3050, loss[loss=0.2085, simple_loss=0.2904, pruned_loss=0.06325, over 18234.00 frames. ], tot_loss[loss=0.1934, simple_loss=0.2773, pruned_loss=0.05474, over 3580135.97 frames. ], batch size: 60, lr: 1.12e-02, grad_scale: 8.0 +2023-03-09 01:24:13,354 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1229, 5.2553, 4.2551, 5.0620, 5.1866, 4.7645, 5.0635, 4.6615], + device='cuda:0'), covar=tensor([0.0798, 0.0758, 0.3100, 0.1246, 0.0829, 0.0498, 0.0782, 0.1538], + device='cuda:0'), in_proj_covar=tensor([0.0398, 0.0449, 0.0596, 0.0349, 0.0339, 0.0409, 0.0439, 0.0572], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 01:24:13,454 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35771.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:24:42,180 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=35796.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:24:42,483 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.4434, 3.6305, 5.0671, 4.1720, 3.0909, 2.8512, 4.1598, 5.2590], + device='cuda:0'), covar=tensor([0.0852, 0.1710, 0.0119, 0.0422, 0.0982, 0.1184, 0.0448, 0.0147], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0237, 0.0099, 0.0159, 0.0174, 0.0173, 0.0171, 0.0139], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 01:24:55,271 INFO [train.py:898] (0/4) Epoch 10, batch 3100, loss[loss=0.2041, simple_loss=0.2924, pruned_loss=0.05788, over 18161.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.2766, pruned_loss=0.05425, over 3589714.06 frames. ], batch size: 62, lr: 1.12e-02, grad_scale: 8.0 +2023-03-09 01:24:58,663 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-03-09 01:25:00,717 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=35811.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:25:07,073 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4295, 3.3690, 3.2161, 2.8098, 3.0743, 2.2490, 2.4098, 3.2037], + device='cuda:0'), covar=tensor([0.0049, 0.0115, 0.0087, 0.0152, 0.0107, 0.0248, 0.0232, 0.0091], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0114, 0.0102, 0.0149, 0.0103, 0.0146, 0.0153, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 01:25:26,933 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35832.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:25:39,853 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.088e+02 3.519e+02 4.020e+02 4.842e+02 1.442e+03, threshold=8.041e+02, percent-clipped=6.0 +2023-03-09 01:25:54,160 INFO [train.py:898] (0/4) Epoch 10, batch 3150, loss[loss=0.1961, simple_loss=0.2816, pruned_loss=0.05528, over 18355.00 frames. ], tot_loss[loss=0.1927, simple_loss=0.2769, pruned_loss=0.05423, over 3599448.03 frames. ], batch size: 56, lr: 1.12e-02, grad_scale: 8.0 +2023-03-09 01:25:57,467 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.72 vs. limit=2.0 +2023-03-09 01:26:12,903 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=35872.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:26:28,853 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=35886.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:26:52,131 INFO [train.py:898] (0/4) Epoch 10, batch 3200, loss[loss=0.1969, simple_loss=0.2879, pruned_loss=0.05294, over 18304.00 frames. ], tot_loss[loss=0.193, simple_loss=0.2771, pruned_loss=0.05445, over 3603165.12 frames. ], batch size: 54, lr: 1.12e-02, grad_scale: 8.0 +2023-03-09 01:27:35,152 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.239e+02 3.585e+02 4.251e+02 5.261e+02 1.507e+03, threshold=8.503e+02, percent-clipped=6.0 +2023-03-09 01:27:39,736 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5491, 5.0462, 5.0448, 4.9806, 4.5854, 4.8936, 4.3838, 4.8894], + device='cuda:0'), covar=tensor([0.0231, 0.0281, 0.0198, 0.0331, 0.0364, 0.0263, 0.1116, 0.0288], + device='cuda:0'), in_proj_covar=tensor([0.0172, 0.0212, 0.0200, 0.0237, 0.0211, 0.0221, 0.0277, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0005, 0.0005, 0.0006, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-09 01:27:39,990 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.29 vs. limit=5.0 +2023-03-09 01:27:49,573 INFO [train.py:898] (0/4) Epoch 10, batch 3250, loss[loss=0.1577, simple_loss=0.2369, pruned_loss=0.03925, over 18592.00 frames. ], tot_loss[loss=0.1935, simple_loss=0.2775, pruned_loss=0.05474, over 3600320.25 frames. ], batch size: 45, lr: 1.12e-02, grad_scale: 8.0 +2023-03-09 01:28:02,768 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-09 01:28:40,607 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-36000.pt +2023-03-09 01:28:46,225 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36001.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 01:28:52,531 INFO [train.py:898] (0/4) Epoch 10, batch 3300, loss[loss=0.168, simple_loss=0.2636, pruned_loss=0.03618, over 18309.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2773, pruned_loss=0.05452, over 3605570.39 frames. ], batch size: 54, lr: 1.12e-02, grad_scale: 8.0 +2023-03-09 01:29:12,650 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8101, 3.6001, 4.9249, 2.9932, 4.1875, 2.5322, 3.0282, 1.8798], + device='cuda:0'), covar=tensor([0.0963, 0.0802, 0.0092, 0.0728, 0.0625, 0.2110, 0.2308, 0.1862], + device='cuda:0'), in_proj_covar=tensor([0.0197, 0.0217, 0.0117, 0.0169, 0.0233, 0.0249, 0.0281, 0.0211], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 01:29:19,586 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36030.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:29:35,215 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.117e+02 3.405e+02 4.202e+02 5.092e+02 8.448e+02, threshold=8.405e+02, percent-clipped=0.0 +2023-03-09 01:29:40,842 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=36049.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 01:29:49,552 INFO [train.py:898] (0/4) Epoch 10, batch 3350, loss[loss=0.1756, simple_loss=0.2588, pruned_loss=0.04621, over 18244.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2771, pruned_loss=0.05462, over 3612060.32 frames. ], batch size: 45, lr: 1.12e-02, grad_scale: 8.0 +2023-03-09 01:30:07,727 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2909, 5.5221, 3.0361, 5.3286, 5.1320, 5.5306, 5.3665, 3.1393], + device='cuda:0'), covar=tensor([0.0136, 0.0043, 0.0594, 0.0063, 0.0056, 0.0053, 0.0072, 0.0757], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0064, 0.0086, 0.0079, 0.0074, 0.0062, 0.0074, 0.0090], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 01:30:13,354 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=36078.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:30:47,889 INFO [train.py:898] (0/4) Epoch 10, batch 3400, loss[loss=0.1859, simple_loss=0.2581, pruned_loss=0.05689, over 18280.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2766, pruned_loss=0.05431, over 3615753.27 frames. ], batch size: 45, lr: 1.11e-02, grad_scale: 8.0 +2023-03-09 01:31:10,747 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36127.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:31:31,611 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.609e+02 3.400e+02 4.178e+02 5.234e+02 8.202e+02, threshold=8.355e+02, percent-clipped=0.0 +2023-03-09 01:31:46,937 INFO [train.py:898] (0/4) Epoch 10, batch 3450, loss[loss=0.1983, simple_loss=0.2854, pruned_loss=0.0556, over 18392.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.2768, pruned_loss=0.05413, over 3605612.70 frames. ], batch size: 48, lr: 1.11e-02, grad_scale: 8.0 +2023-03-09 01:31:49,733 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8006, 3.5459, 4.7473, 2.7908, 3.9252, 2.4734, 2.7825, 2.0418], + device='cuda:0'), covar=tensor([0.0873, 0.0766, 0.0099, 0.0686, 0.0619, 0.2257, 0.2296, 0.1563], + device='cuda:0'), in_proj_covar=tensor([0.0196, 0.0215, 0.0116, 0.0167, 0.0231, 0.0248, 0.0278, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 01:31:58,338 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36167.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:32:14,177 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36181.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:32:20,906 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36186.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:32:45,294 INFO [train.py:898] (0/4) Epoch 10, batch 3500, loss[loss=0.1763, simple_loss=0.2623, pruned_loss=0.04516, over 18402.00 frames. ], tot_loss[loss=0.1926, simple_loss=0.2769, pruned_loss=0.05419, over 3608112.09 frames. ], batch size: 48, lr: 1.11e-02, grad_scale: 8.0 +2023-03-09 01:33:16,084 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=36234.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:33:25,100 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36242.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:33:26,936 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.239e+02 3.308e+02 3.805e+02 4.814e+02 1.268e+03, threshold=7.610e+02, percent-clipped=2.0 +2023-03-09 01:33:41,580 INFO [train.py:898] (0/4) Epoch 10, batch 3550, loss[loss=0.1947, simple_loss=0.2869, pruned_loss=0.05123, over 18362.00 frames. ], tot_loss[loss=0.1925, simple_loss=0.2766, pruned_loss=0.05414, over 3603938.48 frames. ], batch size: 55, lr: 1.11e-02, grad_scale: 8.0 +2023-03-09 01:34:11,727 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5051, 3.7194, 5.1196, 4.3945, 3.2628, 2.9805, 4.4861, 5.3145], + device='cuda:0'), covar=tensor([0.0844, 0.1385, 0.0088, 0.0323, 0.0860, 0.1053, 0.0302, 0.0116], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0235, 0.0100, 0.0157, 0.0174, 0.0172, 0.0170, 0.0140], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 01:34:31,824 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.94 vs. limit=5.0 +2023-03-09 01:34:36,260 INFO [train.py:898] (0/4) Epoch 10, batch 3600, loss[loss=0.2957, simple_loss=0.3401, pruned_loss=0.1257, over 11743.00 frames. ], tot_loss[loss=0.1932, simple_loss=0.2773, pruned_loss=0.05452, over 3587116.75 frames. ], batch size: 130, lr: 1.11e-02, grad_scale: 8.0 +2023-03-09 01:34:43,320 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-09 01:35:11,625 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-10.pt +2023-03-09 01:35:41,693 INFO [train.py:898] (0/4) Epoch 11, batch 0, loss[loss=0.1931, simple_loss=0.2771, pruned_loss=0.05454, over 18289.00 frames. ], tot_loss[loss=0.1931, simple_loss=0.2771, pruned_loss=0.05454, over 18289.00 frames. ], batch size: 49, lr: 1.06e-02, grad_scale: 8.0 +2023-03-09 01:35:41,695 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 01:35:53,392 INFO [train.py:932] (0/4) Epoch 11, validation: loss=0.1597, simple_loss=0.2625, pruned_loss=0.0284, over 944034.00 frames. +2023-03-09 01:35:53,393 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-09 01:35:56,695 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.034e+02 3.291e+02 3.800e+02 4.653e+02 8.329e+02, threshold=7.601e+02, percent-clipped=2.0 +2023-03-09 01:36:26,614 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7917, 4.6822, 4.8569, 4.5470, 4.5909, 4.6388, 5.0216, 4.8952], + device='cuda:0'), covar=tensor([0.0067, 0.0072, 0.0074, 0.0101, 0.0073, 0.0109, 0.0067, 0.0105], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0056, 0.0058, 0.0074, 0.0061, 0.0086, 0.0071, 0.0071], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 01:36:51,732 INFO [train.py:898] (0/4) Epoch 11, batch 50, loss[loss=0.2012, simple_loss=0.2885, pruned_loss=0.05692, over 17965.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2758, pruned_loss=0.05336, over 822333.42 frames. ], batch size: 65, lr: 1.06e-02, grad_scale: 16.0 +2023-03-09 01:37:09,954 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.42 vs. limit=5.0 +2023-03-09 01:37:35,822 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36427.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:37:51,442 INFO [train.py:898] (0/4) Epoch 11, batch 100, loss[loss=0.2146, simple_loss=0.3007, pruned_loss=0.0643, over 18184.00 frames. ], tot_loss[loss=0.1888, simple_loss=0.2724, pruned_loss=0.05261, over 1439201.17 frames. ], batch size: 60, lr: 1.06e-02, grad_scale: 16.0 +2023-03-09 01:37:54,910 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.346e+02 3.421e+02 4.363e+02 5.513e+02 1.312e+03, threshold=8.726e+02, percent-clipped=4.0 +2023-03-09 01:38:01,110 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6324, 3.5158, 1.9586, 4.4269, 3.1018, 4.4710, 2.3109, 3.9611], + device='cuda:0'), covar=tensor([0.0462, 0.0671, 0.1448, 0.0326, 0.0757, 0.0222, 0.1163, 0.0355], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0210, 0.0175, 0.0232, 0.0180, 0.0235, 0.0190, 0.0184], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 01:38:22,618 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36467.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:38:32,142 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=36475.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:38:50,306 INFO [train.py:898] (0/4) Epoch 11, batch 150, loss[loss=0.1668, simple_loss=0.246, pruned_loss=0.04376, over 18409.00 frames. ], tot_loss[loss=0.1887, simple_loss=0.2724, pruned_loss=0.05248, over 1922607.95 frames. ], batch size: 42, lr: 1.06e-02, grad_scale: 16.0 +2023-03-09 01:39:04,597 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.79 vs. limit=5.0 +2023-03-09 01:39:18,199 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=36515.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:39:44,289 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36537.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:39:44,449 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0421, 5.2555, 2.5146, 5.0208, 4.8538, 5.2393, 5.0761, 2.6760], + device='cuda:0'), covar=tensor([0.0169, 0.0042, 0.0820, 0.0073, 0.0074, 0.0058, 0.0082, 0.1030], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0064, 0.0087, 0.0079, 0.0074, 0.0063, 0.0074, 0.0090], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0003, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 01:39:48,739 INFO [train.py:898] (0/4) Epoch 11, batch 200, loss[loss=0.1606, simple_loss=0.2415, pruned_loss=0.03981, over 17276.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2744, pruned_loss=0.0531, over 2299610.39 frames. ], batch size: 38, lr: 1.06e-02, grad_scale: 16.0 +2023-03-09 01:39:52,129 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.119e+02 3.228e+02 3.660e+02 4.259e+02 9.099e+02, threshold=7.320e+02, percent-clipped=1.0 +2023-03-09 01:40:47,360 INFO [train.py:898] (0/4) Epoch 11, batch 250, loss[loss=0.2043, simple_loss=0.2923, pruned_loss=0.05814, over 18498.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2746, pruned_loss=0.0533, over 2579493.58 frames. ], batch size: 53, lr: 1.06e-02, grad_scale: 8.0 +2023-03-09 01:41:04,131 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36605.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:41:07,662 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36608.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:41:12,315 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36612.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:41:32,567 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9031, 3.9836, 2.3907, 3.9923, 4.8840, 2.4304, 3.3856, 3.5108], + device='cuda:0'), covar=tensor([0.0086, 0.1178, 0.1599, 0.0593, 0.0064, 0.1337, 0.0829, 0.0902], + device='cuda:0'), in_proj_covar=tensor([0.0109, 0.0221, 0.0186, 0.0182, 0.0085, 0.0169, 0.0195, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 01:41:39,745 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.68 vs. limit=5.0 +2023-03-09 01:41:41,638 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6503, 4.8395, 4.7748, 4.6661, 4.6913, 5.3261, 4.9900, 4.7581], + device='cuda:0'), covar=tensor([0.0946, 0.0795, 0.0817, 0.0762, 0.1155, 0.0828, 0.0618, 0.1433], + device='cuda:0'), in_proj_covar=tensor([0.0303, 0.0235, 0.0251, 0.0249, 0.0281, 0.0349, 0.0231, 0.0338], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 01:41:47,067 INFO [train.py:898] (0/4) Epoch 11, batch 300, loss[loss=0.2191, simple_loss=0.2989, pruned_loss=0.0697, over 18405.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2748, pruned_loss=0.05324, over 2801442.50 frames. ], batch size: 48, lr: 1.06e-02, grad_scale: 8.0 +2023-03-09 01:41:51,508 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.170e+02 3.364e+02 4.244e+02 4.969e+02 8.450e+02, threshold=8.489e+02, percent-clipped=1.0 +2023-03-09 01:42:15,721 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36666.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:42:18,992 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36669.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:42:23,484 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36673.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:42:45,383 INFO [train.py:898] (0/4) Epoch 11, batch 350, loss[loss=0.1752, simple_loss=0.2517, pruned_loss=0.04935, over 18169.00 frames. ], tot_loss[loss=0.1923, simple_loss=0.2762, pruned_loss=0.05415, over 2975796.56 frames. ], batch size: 44, lr: 1.06e-02, grad_scale: 8.0 +2023-03-09 01:42:55,947 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36700.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 01:43:32,401 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-03-09 01:43:44,247 INFO [train.py:898] (0/4) Epoch 11, batch 400, loss[loss=0.1683, simple_loss=0.2541, pruned_loss=0.04124, over 18358.00 frames. ], tot_loss[loss=0.1922, simple_loss=0.2765, pruned_loss=0.05392, over 3108738.16 frames. ], batch size: 46, lr: 1.06e-02, grad_scale: 8.0 +2023-03-09 01:43:48,756 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.208e+02 3.230e+02 3.792e+02 4.617e+02 9.263e+02, threshold=7.584e+02, percent-clipped=1.0 +2023-03-09 01:43:49,055 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4464, 6.0430, 5.4566, 5.8378, 5.5493, 5.5576, 6.1105, 6.0469], + device='cuda:0'), covar=tensor([0.0953, 0.0608, 0.0381, 0.0578, 0.1204, 0.0626, 0.0477, 0.0544], + device='cuda:0'), in_proj_covar=tensor([0.0520, 0.0424, 0.0332, 0.0473, 0.0642, 0.0472, 0.0605, 0.0451], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 01:44:07,364 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36761.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 01:44:42,724 INFO [train.py:898] (0/4) Epoch 11, batch 450, loss[loss=0.1675, simple_loss=0.2473, pruned_loss=0.04386, over 18412.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2756, pruned_loss=0.05327, over 3231228.75 frames. ], batch size: 42, lr: 1.05e-02, grad_scale: 8.0 +2023-03-09 01:44:59,905 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36805.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 01:45:03,707 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.28 vs. limit=5.0 +2023-03-09 01:45:06,666 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.4464, 3.3679, 4.7581, 4.2108, 3.2230, 2.9361, 4.1432, 4.7239], + device='cuda:0'), covar=tensor([0.0818, 0.1408, 0.0089, 0.0285, 0.0816, 0.1049, 0.0332, 0.0257], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0238, 0.0100, 0.0158, 0.0175, 0.0174, 0.0169, 0.0141], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 01:45:36,365 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=36837.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:45:41,938 INFO [train.py:898] (0/4) Epoch 11, batch 500, loss[loss=0.1795, simple_loss=0.2708, pruned_loss=0.04416, over 18282.00 frames. ], tot_loss[loss=0.1908, simple_loss=0.2754, pruned_loss=0.05307, over 3318050.45 frames. ], batch size: 49, lr: 1.05e-02, grad_scale: 8.0 +2023-03-09 01:45:47,161 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.052e+02 3.251e+02 4.103e+02 5.001e+02 1.385e+03, threshold=8.205e+02, percent-clipped=3.0 +2023-03-09 01:46:11,996 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36866.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 01:46:33,288 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=36885.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:46:39,966 INFO [train.py:898] (0/4) Epoch 11, batch 550, loss[loss=0.2079, simple_loss=0.2917, pruned_loss=0.06203, over 18089.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.275, pruned_loss=0.05284, over 3368035.77 frames. ], batch size: 62, lr: 1.05e-02, grad_scale: 8.0 +2023-03-09 01:47:04,658 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36911.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:47:38,614 INFO [train.py:898] (0/4) Epoch 11, batch 600, loss[loss=0.2363, simple_loss=0.3223, pruned_loss=0.07514, over 18494.00 frames. ], tot_loss[loss=0.1895, simple_loss=0.2745, pruned_loss=0.05231, over 3423015.95 frames. ], batch size: 51, lr: 1.05e-02, grad_scale: 8.0 +2023-03-09 01:47:42,788 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.67 vs. limit=5.0 +2023-03-09 01:47:43,327 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.417e+02 3.303e+02 3.773e+02 4.556e+02 8.624e+02, threshold=7.545e+02, percent-clipped=1.0 +2023-03-09 01:47:46,535 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7829, 3.7015, 3.4586, 3.0460, 3.5117, 2.8289, 2.7126, 3.7042], + device='cuda:0'), covar=tensor([0.0028, 0.0064, 0.0066, 0.0126, 0.0070, 0.0153, 0.0167, 0.0057], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0118, 0.0104, 0.0153, 0.0103, 0.0147, 0.0155, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 01:47:47,946 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-09 01:47:48,224 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-09 01:48:00,232 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=36958.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:48:03,430 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36961.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:48:06,902 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36964.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:48:11,417 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=36968.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:48:16,128 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=36972.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:48:16,185 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.4229, 3.6119, 5.1008, 4.1598, 3.1771, 2.8007, 4.2840, 5.2036], + device='cuda:0'), covar=tensor([0.0851, 0.1710, 0.0110, 0.0374, 0.0916, 0.1186, 0.0339, 0.0153], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0239, 0.0101, 0.0158, 0.0176, 0.0175, 0.0171, 0.0141], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 01:48:17,215 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4075, 5.3946, 4.9311, 5.2920, 5.3886, 4.7052, 5.2530, 5.0167], + device='cuda:0'), covar=tensor([0.0399, 0.0380, 0.1445, 0.0771, 0.0478, 0.0418, 0.0399, 0.0963], + device='cuda:0'), in_proj_covar=tensor([0.0415, 0.0460, 0.0608, 0.0363, 0.0350, 0.0425, 0.0457, 0.0575], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 01:48:37,473 INFO [train.py:898] (0/4) Epoch 11, batch 650, loss[loss=0.1743, simple_loss=0.2561, pruned_loss=0.0463, over 18552.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2741, pruned_loss=0.05206, over 3469920.92 frames. ], batch size: 45, lr: 1.05e-02, grad_scale: 8.0 +2023-03-09 01:48:54,528 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37004.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:49:11,929 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37019.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:49:36,604 INFO [train.py:898] (0/4) Epoch 11, batch 700, loss[loss=0.1888, simple_loss=0.2789, pruned_loss=0.04935, over 17870.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2748, pruned_loss=0.05258, over 3490716.19 frames. ], batch size: 70, lr: 1.05e-02, grad_scale: 8.0 +2023-03-09 01:49:40,958 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.338e+02 3.308e+02 3.875e+02 4.751e+02 1.116e+03, threshold=7.751e+02, percent-clipped=5.0 +2023-03-09 01:49:54,642 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37056.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 01:50:05,325 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37065.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:50:34,406 INFO [train.py:898] (0/4) Epoch 11, batch 750, loss[loss=0.1921, simple_loss=0.2917, pruned_loss=0.04625, over 18399.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2747, pruned_loss=0.05272, over 3512310.50 frames. ], batch size: 52, lr: 1.05e-02, grad_scale: 8.0 +2023-03-09 01:51:33,947 INFO [train.py:898] (0/4) Epoch 11, batch 800, loss[loss=0.2016, simple_loss=0.2952, pruned_loss=0.05393, over 18315.00 frames. ], tot_loss[loss=0.19, simple_loss=0.2752, pruned_loss=0.05244, over 3529756.02 frames. ], batch size: 54, lr: 1.05e-02, grad_scale: 8.0 +2023-03-09 01:51:38,506 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.114e+02 3.217e+02 3.576e+02 4.437e+02 1.024e+03, threshold=7.151e+02, percent-clipped=5.0 +2023-03-09 01:51:58,760 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37161.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 01:52:33,384 INFO [train.py:898] (0/4) Epoch 11, batch 850, loss[loss=0.1539, simple_loss=0.2303, pruned_loss=0.03875, over 18432.00 frames. ], tot_loss[loss=0.1897, simple_loss=0.2745, pruned_loss=0.05241, over 3546101.51 frames. ], batch size: 43, lr: 1.05e-02, grad_scale: 8.0 +2023-03-09 01:53:32,674 INFO [train.py:898] (0/4) Epoch 11, batch 900, loss[loss=0.1946, simple_loss=0.2783, pruned_loss=0.05547, over 17219.00 frames. ], tot_loss[loss=0.1897, simple_loss=0.2745, pruned_loss=0.05248, over 3562761.26 frames. ], batch size: 78, lr: 1.05e-02, grad_scale: 4.0 +2023-03-09 01:53:38,422 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.869e+02 3.048e+02 3.504e+02 4.549e+02 1.028e+03, threshold=7.008e+02, percent-clipped=4.0 +2023-03-09 01:53:56,438 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37261.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:54:00,412 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37264.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:54:01,625 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5952, 2.6442, 3.6284, 3.6395, 2.4720, 3.9281, 3.7121, 2.5464], + device='cuda:0'), covar=tensor([0.0273, 0.1261, 0.0253, 0.0199, 0.1439, 0.0208, 0.0366, 0.1091], + device='cuda:0'), in_proj_covar=tensor([0.0184, 0.0222, 0.0150, 0.0144, 0.0212, 0.0183, 0.0209, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 01:54:04,548 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37267.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:54:05,896 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37268.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:54:05,940 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37268.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:54:22,664 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9144, 3.9339, 5.1054, 2.7443, 4.2858, 2.6073, 2.9377, 1.8762], + device='cuda:0'), covar=tensor([0.0912, 0.0657, 0.0071, 0.0747, 0.0540, 0.2208, 0.2407, 0.1776], + device='cuda:0'), in_proj_covar=tensor([0.0195, 0.0214, 0.0116, 0.0168, 0.0227, 0.0245, 0.0279, 0.0208], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 01:54:32,494 INFO [train.py:898] (0/4) Epoch 11, batch 950, loss[loss=0.1774, simple_loss=0.2652, pruned_loss=0.04481, over 18389.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2732, pruned_loss=0.05229, over 3566968.16 frames. ], batch size: 50, lr: 1.05e-02, grad_scale: 4.0 +2023-03-09 01:54:52,582 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=37309.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:54:56,037 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=37312.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:54:58,325 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37314.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:55:01,076 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=37316.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:55:17,988 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37329.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:55:31,163 INFO [train.py:898] (0/4) Epoch 11, batch 1000, loss[loss=0.2051, simple_loss=0.294, pruned_loss=0.05807, over 17725.00 frames. ], tot_loss[loss=0.1898, simple_loss=0.2743, pruned_loss=0.05258, over 3579742.32 frames. ], batch size: 70, lr: 1.05e-02, grad_scale: 4.0 +2023-03-09 01:55:36,614 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.267e+02 3.387e+02 4.010e+02 5.026e+02 9.863e+02, threshold=8.020e+02, percent-clipped=4.0 +2023-03-09 01:55:48,198 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37356.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 01:55:52,754 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37360.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:55:56,296 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3323, 5.4920, 2.9516, 5.3474, 5.2754, 5.6355, 5.3801, 2.8360], + device='cuda:0'), covar=tensor([0.0138, 0.0064, 0.0655, 0.0058, 0.0052, 0.0046, 0.0078, 0.0858], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0066, 0.0087, 0.0079, 0.0074, 0.0064, 0.0076, 0.0090], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 01:55:57,546 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37364.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:56:29,968 INFO [train.py:898] (0/4) Epoch 11, batch 1050, loss[loss=0.169, simple_loss=0.262, pruned_loss=0.03801, over 18585.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2751, pruned_loss=0.05309, over 3577580.08 frames. ], batch size: 54, lr: 1.05e-02, grad_scale: 4.0 +2023-03-09 01:56:44,663 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=37404.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 01:57:06,834 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6580, 2.0795, 2.6640, 2.9456, 3.4775, 5.2445, 4.6817, 4.0247], + device='cuda:0'), covar=tensor([0.1192, 0.2029, 0.2381, 0.1253, 0.1643, 0.0083, 0.0351, 0.0452], + device='cuda:0'), in_proj_covar=tensor([0.0238, 0.0296, 0.0310, 0.0245, 0.0359, 0.0187, 0.0260, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 01:57:09,730 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=37425.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 01:57:26,769 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-09 01:57:28,533 INFO [train.py:898] (0/4) Epoch 11, batch 1100, loss[loss=0.205, simple_loss=0.2792, pruned_loss=0.06546, over 13340.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2749, pruned_loss=0.05286, over 3578550.76 frames. ], batch size: 130, lr: 1.05e-02, grad_scale: 4.0 +2023-03-09 01:57:34,021 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.186e+02 3.484e+02 3.927e+02 4.853e+02 9.182e+02, threshold=7.853e+02, percent-clipped=3.0 +2023-03-09 01:57:51,274 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37461.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 01:57:56,037 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6651, 3.6429, 4.9696, 4.3400, 3.1289, 3.0353, 4.3132, 5.1252], + device='cuda:0'), covar=tensor([0.0775, 0.1509, 0.0100, 0.0298, 0.0893, 0.0995, 0.0318, 0.0149], + device='cuda:0'), in_proj_covar=tensor([0.0137, 0.0243, 0.0102, 0.0159, 0.0179, 0.0176, 0.0172, 0.0144], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 01:58:27,590 INFO [train.py:898] (0/4) Epoch 11, batch 1150, loss[loss=0.1636, simple_loss=0.2514, pruned_loss=0.0379, over 18491.00 frames. ], tot_loss[loss=0.1899, simple_loss=0.2745, pruned_loss=0.05265, over 3584506.99 frames. ], batch size: 53, lr: 1.04e-02, grad_scale: 4.0 +2023-03-09 01:58:48,330 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=37509.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 01:59:02,362 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.94 vs. limit=5.0 +2023-03-09 01:59:20,585 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0480, 3.7583, 5.0793, 2.9608, 4.3319, 2.5394, 3.0098, 1.9782], + device='cuda:0'), covar=tensor([0.0869, 0.0791, 0.0082, 0.0673, 0.0578, 0.2308, 0.2583, 0.1749], + device='cuda:0'), in_proj_covar=tensor([0.0198, 0.0217, 0.0118, 0.0170, 0.0229, 0.0249, 0.0283, 0.0210], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 01:59:26,243 INFO [train.py:898] (0/4) Epoch 11, batch 1200, loss[loss=0.1841, simple_loss=0.273, pruned_loss=0.04761, over 18382.00 frames. ], tot_loss[loss=0.1899, simple_loss=0.2748, pruned_loss=0.05254, over 3595840.10 frames. ], batch size: 50, lr: 1.04e-02, grad_scale: 8.0 +2023-03-09 01:59:31,793 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.211e+02 3.034e+02 3.620e+02 4.493e+02 1.296e+03, threshold=7.239e+02, percent-clipped=3.0 +2023-03-09 01:59:55,936 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37567.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:00:24,872 INFO [train.py:898] (0/4) Epoch 11, batch 1250, loss[loss=0.1779, simple_loss=0.2636, pruned_loss=0.04609, over 18421.00 frames. ], tot_loss[loss=0.1904, simple_loss=0.2755, pruned_loss=0.0527, over 3585345.77 frames. ], batch size: 48, lr: 1.04e-02, grad_scale: 8.0 +2023-03-09 02:00:52,616 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37614.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:00:53,513 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=37615.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:00:54,948 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8469, 3.8514, 3.6429, 3.2036, 3.6064, 2.9406, 2.8926, 3.7926], + device='cuda:0'), covar=tensor([0.0029, 0.0063, 0.0057, 0.0100, 0.0057, 0.0128, 0.0150, 0.0045], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0118, 0.0104, 0.0152, 0.0103, 0.0148, 0.0157, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 02:00:57,421 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-09 02:01:04,003 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37624.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:01:11,231 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-09 02:01:24,586 INFO [train.py:898] (0/4) Epoch 11, batch 1300, loss[loss=0.1606, simple_loss=0.2435, pruned_loss=0.0388, over 18373.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.275, pruned_loss=0.0528, over 3573053.81 frames. ], batch size: 46, lr: 1.04e-02, grad_scale: 8.0 +2023-03-09 02:01:31,769 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.066e+02 3.446e+02 4.010e+02 4.726e+02 9.288e+02, threshold=8.020e+02, percent-clipped=3.0 +2023-03-09 02:01:46,688 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37660.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:01:48,852 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=37662.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:01:49,336 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-09 02:02:23,199 INFO [train.py:898] (0/4) Epoch 11, batch 1350, loss[loss=0.2135, simple_loss=0.2945, pruned_loss=0.06627, over 13144.00 frames. ], tot_loss[loss=0.1907, simple_loss=0.2752, pruned_loss=0.05303, over 3580416.38 frames. ], batch size: 129, lr: 1.04e-02, grad_scale: 4.0 +2023-03-09 02:02:24,846 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5394, 1.9867, 2.6240, 2.7965, 3.2897, 5.2148, 4.5937, 3.9268], + device='cuda:0'), covar=tensor([0.1366, 0.2199, 0.2593, 0.1402, 0.2025, 0.0099, 0.0398, 0.0530], + device='cuda:0'), in_proj_covar=tensor([0.0243, 0.0300, 0.0313, 0.0250, 0.0363, 0.0188, 0.0263, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 02:02:43,281 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=37708.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:02:56,754 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=37720.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:03:21,772 INFO [train.py:898] (0/4) Epoch 11, batch 1400, loss[loss=0.188, simple_loss=0.2782, pruned_loss=0.04896, over 15800.00 frames. ], tot_loss[loss=0.191, simple_loss=0.2756, pruned_loss=0.0532, over 3577683.90 frames. ], batch size: 94, lr: 1.04e-02, grad_scale: 4.0 +2023-03-09 02:03:29,183 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.922e+02 3.017e+02 3.586e+02 4.269e+02 9.001e+02, threshold=7.171e+02, percent-clipped=1.0 +2023-03-09 02:03:51,025 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-09 02:04:20,147 INFO [train.py:898] (0/4) Epoch 11, batch 1450, loss[loss=0.1837, simple_loss=0.2613, pruned_loss=0.05303, over 18505.00 frames. ], tot_loss[loss=0.1912, simple_loss=0.2757, pruned_loss=0.05332, over 3577590.85 frames. ], batch size: 44, lr: 1.04e-02, grad_scale: 4.0 +2023-03-09 02:04:26,292 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8591, 4.0718, 2.4394, 4.0596, 4.8405, 2.3870, 3.7133, 3.8416], + device='cuda:0'), covar=tensor([0.0064, 0.0727, 0.1342, 0.0448, 0.0052, 0.1232, 0.0566, 0.0590], + device='cuda:0'), in_proj_covar=tensor([0.0111, 0.0226, 0.0186, 0.0184, 0.0088, 0.0170, 0.0195, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 02:05:06,817 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-09 02:05:19,807 INFO [train.py:898] (0/4) Epoch 11, batch 1500, loss[loss=0.151, simple_loss=0.2342, pruned_loss=0.03384, over 18488.00 frames. ], tot_loss[loss=0.1905, simple_loss=0.2751, pruned_loss=0.05295, over 3579312.58 frames. ], batch size: 44, lr: 1.04e-02, grad_scale: 4.0 +2023-03-09 02:05:27,763 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.812e+02 3.086e+02 3.620e+02 4.300e+02 1.406e+03, threshold=7.239e+02, percent-clipped=4.0 +2023-03-09 02:06:18,399 INFO [train.py:898] (0/4) Epoch 11, batch 1550, loss[loss=0.1979, simple_loss=0.2808, pruned_loss=0.05754, over 18310.00 frames. ], tot_loss[loss=0.1899, simple_loss=0.2746, pruned_loss=0.05255, over 3584564.10 frames. ], batch size: 54, lr: 1.04e-02, grad_scale: 4.0 +2023-03-09 02:06:43,805 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.4775, 4.4483, 4.6375, 4.3827, 4.4296, 4.4562, 4.6974, 4.6727], + device='cuda:0'), covar=tensor([0.0104, 0.0104, 0.0087, 0.0117, 0.0082, 0.0145, 0.0091, 0.0093], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0055, 0.0058, 0.0074, 0.0061, 0.0085, 0.0071, 0.0070], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 02:06:45,246 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-09 02:06:58,443 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=37924.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:07:17,439 INFO [train.py:898] (0/4) Epoch 11, batch 1600, loss[loss=0.202, simple_loss=0.2908, pruned_loss=0.05663, over 18348.00 frames. ], tot_loss[loss=0.1898, simple_loss=0.2745, pruned_loss=0.0525, over 3576157.61 frames. ], batch size: 56, lr: 1.04e-02, grad_scale: 8.0 +2023-03-09 02:07:22,329 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6810, 2.8693, 2.8136, 2.7703, 3.6152, 3.5556, 3.2159, 2.9367], + device='cuda:0'), covar=tensor([0.0165, 0.0265, 0.0497, 0.0339, 0.0207, 0.0150, 0.0315, 0.0345], + device='cuda:0'), in_proj_covar=tensor([0.0116, 0.0104, 0.0147, 0.0136, 0.0101, 0.0088, 0.0133, 0.0127], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 02:07:24,292 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.933e+02 3.097e+02 3.749e+02 4.631e+02 9.709e+02, threshold=7.497e+02, percent-clipped=4.0 +2023-03-09 02:07:35,943 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2387, 5.2061, 4.7526, 5.1543, 5.1592, 4.4912, 5.0318, 4.8175], + device='cuda:0'), covar=tensor([0.0378, 0.0429, 0.1408, 0.0663, 0.0585, 0.0522, 0.0464, 0.1007], + device='cuda:0'), in_proj_covar=tensor([0.0412, 0.0462, 0.0617, 0.0365, 0.0354, 0.0425, 0.0458, 0.0585], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 02:07:38,201 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9314, 5.5171, 5.4969, 5.4455, 5.1253, 5.4329, 4.7988, 5.3741], + device='cuda:0'), covar=tensor([0.0210, 0.0212, 0.0148, 0.0307, 0.0260, 0.0162, 0.1029, 0.0216], + device='cuda:0'), in_proj_covar=tensor([0.0175, 0.0217, 0.0205, 0.0246, 0.0220, 0.0227, 0.0283, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0005, 0.0006, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-09 02:07:54,472 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=37972.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:08:02,490 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=37979.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:08:15,894 INFO [train.py:898] (0/4) Epoch 11, batch 1650, loss[loss=0.2125, simple_loss=0.2936, pruned_loss=0.06568, over 17971.00 frames. ], tot_loss[loss=0.1891, simple_loss=0.2736, pruned_loss=0.05227, over 3581892.37 frames. ], batch size: 65, lr: 1.04e-02, grad_scale: 8.0 +2023-03-09 02:08:26,391 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-38000.pt +2023-03-09 02:08:55,115 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=38020.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:09:18,061 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=38040.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:09:18,736 INFO [train.py:898] (0/4) Epoch 11, batch 1700, loss[loss=0.1913, simple_loss=0.2743, pruned_loss=0.05416, over 15862.00 frames. ], tot_loss[loss=0.1903, simple_loss=0.2746, pruned_loss=0.05303, over 3573378.25 frames. ], batch size: 94, lr: 1.04e-02, grad_scale: 8.0 +2023-03-09 02:09:25,206 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.262e+02 3.409e+02 3.922e+02 5.492e+02 2.210e+03, threshold=7.843e+02, percent-clipped=9.0 +2023-03-09 02:09:50,082 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=38068.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:10:05,218 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.56 vs. limit=5.0 +2023-03-09 02:10:16,084 INFO [train.py:898] (0/4) Epoch 11, batch 1750, loss[loss=0.2363, simple_loss=0.3101, pruned_loss=0.08125, over 12244.00 frames. ], tot_loss[loss=0.1911, simple_loss=0.2754, pruned_loss=0.05344, over 3563710.16 frames. ], batch size: 130, lr: 1.04e-02, grad_scale: 8.0 +2023-03-09 02:10:44,328 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1940, 3.9992, 5.0103, 2.7760, 4.4148, 2.7631, 3.1038, 2.0603], + device='cuda:0'), covar=tensor([0.0792, 0.0696, 0.0093, 0.0761, 0.0492, 0.2127, 0.2377, 0.1729], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0214, 0.0117, 0.0167, 0.0227, 0.0246, 0.0278, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 02:11:12,519 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-09 02:11:15,130 INFO [train.py:898] (0/4) Epoch 11, batch 1800, loss[loss=0.1826, simple_loss=0.2637, pruned_loss=0.05071, over 18363.00 frames. ], tot_loss[loss=0.1913, simple_loss=0.2756, pruned_loss=0.05353, over 3571642.47 frames. ], batch size: 46, lr: 1.04e-02, grad_scale: 8.0 +2023-03-09 02:11:21,537 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.087e+02 3.079e+02 3.713e+02 4.653e+02 8.656e+02, threshold=7.427e+02, percent-clipped=3.0 +2023-03-09 02:12:12,800 INFO [train.py:898] (0/4) Epoch 11, batch 1850, loss[loss=0.1716, simple_loss=0.2557, pruned_loss=0.04371, over 18496.00 frames. ], tot_loss[loss=0.1915, simple_loss=0.2758, pruned_loss=0.05356, over 3577032.14 frames. ], batch size: 47, lr: 1.04e-02, grad_scale: 8.0 +2023-03-09 02:12:33,285 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8133, 5.3153, 5.2808, 5.3395, 4.8302, 5.1867, 4.6458, 5.1404], + device='cuda:0'), covar=tensor([0.0206, 0.0267, 0.0196, 0.0308, 0.0334, 0.0198, 0.1089, 0.0264], + device='cuda:0'), in_proj_covar=tensor([0.0174, 0.0219, 0.0206, 0.0247, 0.0221, 0.0224, 0.0282, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0005, 0.0006, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-09 02:12:38,903 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.66 vs. limit=5.0 +2023-03-09 02:13:12,755 INFO [train.py:898] (0/4) Epoch 11, batch 1900, loss[loss=0.1608, simple_loss=0.2402, pruned_loss=0.04072, over 18244.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2753, pruned_loss=0.05294, over 3585850.91 frames. ], batch size: 45, lr: 1.03e-02, grad_scale: 8.0 +2023-03-09 02:13:19,699 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.086e+02 3.362e+02 3.965e+02 4.724e+02 1.180e+03, threshold=7.931e+02, percent-clipped=5.0 +2023-03-09 02:13:43,773 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2428, 5.3332, 4.3435, 5.1264, 5.2951, 4.8979, 5.1295, 4.8351], + device='cuda:0'), covar=tensor([0.0783, 0.0635, 0.2894, 0.1362, 0.0756, 0.0473, 0.0785, 0.1195], + device='cuda:0'), in_proj_covar=tensor([0.0409, 0.0461, 0.0616, 0.0361, 0.0351, 0.0419, 0.0459, 0.0578], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 02:14:11,579 INFO [train.py:898] (0/4) Epoch 11, batch 1950, loss[loss=0.2043, simple_loss=0.2922, pruned_loss=0.05824, over 16010.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2751, pruned_loss=0.05256, over 3588475.42 frames. ], batch size: 94, lr: 1.03e-02, grad_scale: 8.0 +2023-03-09 02:14:11,800 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9344, 4.9754, 5.1799, 5.0421, 4.9987, 5.6194, 5.2209, 5.0431], + device='cuda:0'), covar=tensor([0.1014, 0.0753, 0.0716, 0.0754, 0.1470, 0.0954, 0.0647, 0.1670], + device='cuda:0'), in_proj_covar=tensor([0.0306, 0.0235, 0.0251, 0.0254, 0.0287, 0.0358, 0.0232, 0.0348], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0004, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 02:15:04,205 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=38335.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:15:10,798 INFO [train.py:898] (0/4) Epoch 11, batch 2000, loss[loss=0.1986, simple_loss=0.288, pruned_loss=0.05455, over 18187.00 frames. ], tot_loss[loss=0.1901, simple_loss=0.2752, pruned_loss=0.05252, over 3591167.44 frames. ], batch size: 60, lr: 1.03e-02, grad_scale: 8.0 +2023-03-09 02:15:11,529 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-09 02:15:17,660 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.188e+02 3.163e+02 3.706e+02 4.529e+02 9.366e+02, threshold=7.411e+02, percent-clipped=1.0 +2023-03-09 02:15:32,772 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=38360.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:16:08,896 INFO [train.py:898] (0/4) Epoch 11, batch 2050, loss[loss=0.1732, simple_loss=0.2655, pruned_loss=0.04049, over 18386.00 frames. ], tot_loss[loss=0.1896, simple_loss=0.2748, pruned_loss=0.05223, over 3592495.44 frames. ], batch size: 50, lr: 1.03e-02, grad_scale: 8.0 +2023-03-09 02:16:44,565 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=38421.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:17:08,307 INFO [train.py:898] (0/4) Epoch 11, batch 2100, loss[loss=0.1766, simple_loss=0.2564, pruned_loss=0.04845, over 18253.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2738, pruned_loss=0.0517, over 3604169.07 frames. ], batch size: 47, lr: 1.03e-02, grad_scale: 4.0 +2023-03-09 02:17:15,906 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.369e+02 3.250e+02 4.019e+02 4.989e+02 1.105e+03, threshold=8.037e+02, percent-clipped=2.0 +2023-03-09 02:18:03,340 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0313, 5.0160, 4.5397, 5.0012, 4.9533, 4.3341, 4.8276, 4.6089], + device='cuda:0'), covar=tensor([0.0467, 0.0532, 0.1682, 0.0741, 0.0583, 0.0491, 0.0544, 0.1078], + device='cuda:0'), in_proj_covar=tensor([0.0412, 0.0459, 0.0619, 0.0363, 0.0350, 0.0425, 0.0456, 0.0585], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 02:18:07,058 INFO [train.py:898] (0/4) Epoch 11, batch 2150, loss[loss=0.2445, simple_loss=0.3138, pruned_loss=0.08761, over 12773.00 frames. ], tot_loss[loss=0.189, simple_loss=0.274, pruned_loss=0.05203, over 3588113.72 frames. ], batch size: 129, lr: 1.03e-02, grad_scale: 4.0 +2023-03-09 02:18:53,004 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-09 02:19:05,991 INFO [train.py:898] (0/4) Epoch 11, batch 2200, loss[loss=0.1594, simple_loss=0.2379, pruned_loss=0.04042, over 18430.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2737, pruned_loss=0.05206, over 3580824.54 frames. ], batch size: 43, lr: 1.03e-02, grad_scale: 4.0 +2023-03-09 02:19:13,836 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.988e+02 3.259e+02 3.995e+02 5.001e+02 1.029e+03, threshold=7.990e+02, percent-clipped=4.0 +2023-03-09 02:19:54,564 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5000, 2.0846, 2.6996, 2.7048, 3.2976, 5.1525, 4.6115, 3.9218], + device='cuda:0'), covar=tensor([0.1323, 0.2116, 0.2393, 0.1463, 0.1899, 0.0087, 0.0371, 0.0504], + device='cuda:0'), in_proj_covar=tensor([0.0244, 0.0300, 0.0313, 0.0249, 0.0364, 0.0187, 0.0259, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 02:20:04,920 INFO [train.py:898] (0/4) Epoch 11, batch 2250, loss[loss=0.1989, simple_loss=0.297, pruned_loss=0.05041, over 18281.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.273, pruned_loss=0.05157, over 3582085.98 frames. ], batch size: 57, lr: 1.03e-02, grad_scale: 4.0 +2023-03-09 02:20:57,220 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=38635.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:21:04,276 INFO [train.py:898] (0/4) Epoch 11, batch 2300, loss[loss=0.231, simple_loss=0.3148, pruned_loss=0.0736, over 16073.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2731, pruned_loss=0.05173, over 3587342.58 frames. ], batch size: 94, lr: 1.03e-02, grad_scale: 4.0 +2023-03-09 02:21:12,484 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.254e+02 3.045e+02 3.798e+02 4.329e+02 8.065e+02, threshold=7.597e+02, percent-clipped=1.0 +2023-03-09 02:21:13,208 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-03-09 02:21:22,180 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5551, 3.3170, 2.2785, 4.3761, 2.7614, 4.4374, 2.4597, 4.0032], + device='cuda:0'), covar=tensor([0.0585, 0.0855, 0.1353, 0.0440, 0.0950, 0.0229, 0.1174, 0.0357], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0212, 0.0177, 0.0238, 0.0181, 0.0240, 0.0193, 0.0187], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 02:21:53,528 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=38683.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:22:03,152 INFO [train.py:898] (0/4) Epoch 11, batch 2350, loss[loss=0.2232, simple_loss=0.3155, pruned_loss=0.06548, over 18134.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2731, pruned_loss=0.05196, over 3583057.02 frames. ], batch size: 62, lr: 1.03e-02, grad_scale: 4.0 +2023-03-09 02:22:32,549 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=38716.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:23:01,669 INFO [train.py:898] (0/4) Epoch 11, batch 2400, loss[loss=0.1638, simple_loss=0.2428, pruned_loss=0.04236, over 18457.00 frames. ], tot_loss[loss=0.1889, simple_loss=0.2736, pruned_loss=0.05214, over 3582545.62 frames. ], batch size: 43, lr: 1.03e-02, grad_scale: 8.0 +2023-03-09 02:23:10,154 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.250e+02 3.115e+02 4.067e+02 4.907e+02 9.173e+02, threshold=8.134e+02, percent-clipped=4.0 +2023-03-09 02:23:22,676 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4865, 5.4260, 5.0807, 5.3708, 5.4140, 4.7915, 5.3280, 5.0043], + device='cuda:0'), covar=tensor([0.0404, 0.0414, 0.1403, 0.0848, 0.0508, 0.0425, 0.0400, 0.1053], + device='cuda:0'), in_proj_covar=tensor([0.0407, 0.0460, 0.0610, 0.0363, 0.0350, 0.0421, 0.0451, 0.0582], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 02:24:00,849 INFO [train.py:898] (0/4) Epoch 11, batch 2450, loss[loss=0.1838, simple_loss=0.2739, pruned_loss=0.04683, over 18305.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2732, pruned_loss=0.05196, over 3577493.61 frames. ], batch size: 54, lr: 1.03e-02, grad_scale: 8.0 +2023-03-09 02:24:02,911 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.88 vs. limit=5.0 +2023-03-09 02:24:38,836 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=38823.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:24:59,957 INFO [train.py:898] (0/4) Epoch 11, batch 2500, loss[loss=0.2139, simple_loss=0.3008, pruned_loss=0.06351, over 18473.00 frames. ], tot_loss[loss=0.1894, simple_loss=0.274, pruned_loss=0.05234, over 3574574.61 frames. ], batch size: 59, lr: 1.03e-02, grad_scale: 8.0 +2023-03-09 02:25:08,484 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.178e+02 3.118e+02 3.887e+02 4.654e+02 1.248e+03, threshold=7.775e+02, percent-clipped=2.0 +2023-03-09 02:25:40,538 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-09 02:25:43,800 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3752, 4.3180, 2.8347, 4.3386, 5.3344, 2.6004, 4.0912, 4.1809], + device='cuda:0'), covar=tensor([0.0080, 0.1215, 0.1320, 0.0498, 0.0060, 0.1215, 0.0591, 0.0660], + device='cuda:0'), in_proj_covar=tensor([0.0117, 0.0234, 0.0191, 0.0187, 0.0091, 0.0174, 0.0202, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 02:25:47,160 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5096, 3.4156, 2.2108, 4.3852, 2.8202, 4.3752, 2.2778, 4.0771], + device='cuda:0'), covar=tensor([0.0555, 0.0752, 0.1321, 0.0393, 0.0887, 0.0277, 0.1195, 0.0275], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0211, 0.0178, 0.0238, 0.0179, 0.0238, 0.0191, 0.0186], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 02:25:50,727 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=38884.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:25:58,817 INFO [train.py:898] (0/4) Epoch 11, batch 2550, loss[loss=0.192, simple_loss=0.2788, pruned_loss=0.05261, over 17158.00 frames. ], tot_loss[loss=0.1886, simple_loss=0.2736, pruned_loss=0.05181, over 3590568.31 frames. ], batch size: 78, lr: 1.03e-02, grad_scale: 8.0 +2023-03-09 02:26:57,534 INFO [train.py:898] (0/4) Epoch 11, batch 2600, loss[loss=0.1809, simple_loss=0.2735, pruned_loss=0.04416, over 18569.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.2736, pruned_loss=0.05152, over 3595101.44 frames. ], batch size: 54, lr: 1.03e-02, grad_scale: 8.0 +2023-03-09 02:27:06,475 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.982e+02 2.998e+02 3.498e+02 4.234e+02 9.480e+02, threshold=6.995e+02, percent-clipped=2.0 +2023-03-09 02:27:56,947 INFO [train.py:898] (0/4) Epoch 11, batch 2650, loss[loss=0.2089, simple_loss=0.3009, pruned_loss=0.05847, over 18026.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2732, pruned_loss=0.05153, over 3601681.39 frames. ], batch size: 65, lr: 1.02e-02, grad_scale: 8.0 +2023-03-09 02:28:27,525 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39016.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:28:56,249 INFO [train.py:898] (0/4) Epoch 11, batch 2700, loss[loss=0.1968, simple_loss=0.2886, pruned_loss=0.05248, over 18318.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2731, pruned_loss=0.0516, over 3589608.03 frames. ], batch size: 54, lr: 1.02e-02, grad_scale: 8.0 +2023-03-09 02:29:04,935 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.335e+02 3.292e+02 3.968e+02 4.768e+02 1.831e+03, threshold=7.936e+02, percent-clipped=8.0 +2023-03-09 02:29:24,689 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=39064.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:29:47,316 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8706, 3.5823, 4.8701, 2.8135, 4.2363, 2.5651, 3.0032, 1.7913], + device='cuda:0'), covar=tensor([0.0914, 0.0851, 0.0089, 0.0700, 0.0572, 0.2210, 0.2428, 0.1859], + device='cuda:0'), in_proj_covar=tensor([0.0195, 0.0215, 0.0117, 0.0169, 0.0229, 0.0247, 0.0282, 0.0210], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 02:29:52,241 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-09 02:29:55,895 INFO [train.py:898] (0/4) Epoch 11, batch 2750, loss[loss=0.1598, simple_loss=0.2469, pruned_loss=0.03639, over 18370.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.2728, pruned_loss=0.0514, over 3592727.94 frames. ], batch size: 50, lr: 1.02e-02, grad_scale: 8.0 +2023-03-09 02:30:55,601 INFO [train.py:898] (0/4) Epoch 11, batch 2800, loss[loss=0.1765, simple_loss=0.2621, pruned_loss=0.0455, over 18361.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2726, pruned_loss=0.05135, over 3583612.30 frames. ], batch size: 46, lr: 1.02e-02, grad_scale: 8.0 +2023-03-09 02:31:04,053 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.053e+02 3.386e+02 4.032e+02 4.876e+02 1.472e+03, threshold=8.064e+02, percent-clipped=5.0 +2023-03-09 02:31:42,038 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39179.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:31:48,934 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39185.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:31:55,217 INFO [train.py:898] (0/4) Epoch 11, batch 2850, loss[loss=0.2146, simple_loss=0.2933, pruned_loss=0.06794, over 18269.00 frames. ], tot_loss[loss=0.1881, simple_loss=0.2726, pruned_loss=0.0518, over 3584491.89 frames. ], batch size: 57, lr: 1.02e-02, grad_scale: 8.0 +2023-03-09 02:32:54,405 INFO [train.py:898] (0/4) Epoch 11, batch 2900, loss[loss=0.1578, simple_loss=0.2409, pruned_loss=0.03736, over 17576.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.2722, pruned_loss=0.05138, over 3591282.20 frames. ], batch size: 39, lr: 1.02e-02, grad_scale: 8.0 +2023-03-09 02:33:00,454 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39246.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:33:02,142 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.400e+02 3.149e+02 3.663e+02 4.555e+02 1.238e+03, threshold=7.326e+02, percent-clipped=2.0 +2023-03-09 02:33:05,762 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6524, 2.9079, 4.1974, 3.5873, 2.5579, 4.5460, 3.8312, 2.8475], + device='cuda:0'), covar=tensor([0.0467, 0.1217, 0.0192, 0.0386, 0.1358, 0.0209, 0.0503, 0.0867], + device='cuda:0'), in_proj_covar=tensor([0.0184, 0.0216, 0.0148, 0.0139, 0.0202, 0.0179, 0.0205, 0.0185], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 02:33:45,289 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39284.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:33:53,054 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-09 02:33:53,614 INFO [train.py:898] (0/4) Epoch 11, batch 2950, loss[loss=0.1912, simple_loss=0.283, pruned_loss=0.04972, over 18393.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2725, pruned_loss=0.05149, over 3591627.04 frames. ], batch size: 52, lr: 1.02e-02, grad_scale: 8.0 +2023-03-09 02:34:53,169 INFO [train.py:898] (0/4) Epoch 11, batch 3000, loss[loss=0.1834, simple_loss=0.2743, pruned_loss=0.04629, over 18473.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2736, pruned_loss=0.05175, over 3602108.18 frames. ], batch size: 51, lr: 1.02e-02, grad_scale: 8.0 +2023-03-09 02:34:53,171 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 02:35:05,606 INFO [train.py:932] (0/4) Epoch 11, validation: loss=0.1587, simple_loss=0.2603, pruned_loss=0.02852, over 944034.00 frames. +2023-03-09 02:35:05,607 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-09 02:35:10,782 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39345.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:35:13,862 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.978e+02 3.242e+02 3.927e+02 4.658e+02 9.416e+02, threshold=7.854e+02, percent-clipped=4.0 +2023-03-09 02:35:38,601 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-09 02:36:04,358 INFO [train.py:898] (0/4) Epoch 11, batch 3050, loss[loss=0.1739, simple_loss=0.255, pruned_loss=0.04644, over 18379.00 frames. ], tot_loss[loss=0.189, simple_loss=0.2741, pruned_loss=0.05191, over 3599323.64 frames. ], batch size: 50, lr: 1.02e-02, grad_scale: 8.0 +2023-03-09 02:37:04,061 INFO [train.py:898] (0/4) Epoch 11, batch 3100, loss[loss=0.1693, simple_loss=0.2502, pruned_loss=0.04422, over 18454.00 frames. ], tot_loss[loss=0.1885, simple_loss=0.2734, pruned_loss=0.05179, over 3597491.86 frames. ], batch size: 43, lr: 1.02e-02, grad_scale: 8.0 +2023-03-09 02:37:12,130 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.117e+02 3.287e+02 3.708e+02 4.469e+02 1.141e+03, threshold=7.415e+02, percent-clipped=2.0 +2023-03-09 02:37:23,512 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6850, 2.1294, 2.7483, 2.6956, 3.4829, 5.1524, 4.5977, 3.9932], + device='cuda:0'), covar=tensor([0.1250, 0.2010, 0.2408, 0.1402, 0.1687, 0.0106, 0.0373, 0.0506], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0301, 0.0317, 0.0248, 0.0361, 0.0189, 0.0260, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 02:37:24,077 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-09 02:37:30,747 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39463.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:37:49,480 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39479.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:38:01,373 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=5.04 vs. limit=5.0 +2023-03-09 02:38:02,613 INFO [train.py:898] (0/4) Epoch 11, batch 3150, loss[loss=0.1737, simple_loss=0.2581, pruned_loss=0.04468, over 18538.00 frames. ], tot_loss[loss=0.1883, simple_loss=0.273, pruned_loss=0.05182, over 3587976.54 frames. ], batch size: 49, lr: 1.02e-02, grad_scale: 8.0 +2023-03-09 02:38:42,100 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39524.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:38:45,194 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=39527.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:39:01,995 INFO [train.py:898] (0/4) Epoch 11, batch 3200, loss[loss=0.1918, simple_loss=0.2762, pruned_loss=0.05368, over 18532.00 frames. ], tot_loss[loss=0.1879, simple_loss=0.2727, pruned_loss=0.05149, over 3582222.30 frames. ], batch size: 49, lr: 1.02e-02, grad_scale: 8.0 +2023-03-09 02:39:02,147 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39541.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:39:09,610 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.046e+02 3.172e+02 3.769e+02 4.644e+02 9.591e+02, threshold=7.537e+02, percent-clipped=4.0 +2023-03-09 02:40:01,221 INFO [train.py:898] (0/4) Epoch 11, batch 3250, loss[loss=0.1738, simple_loss=0.255, pruned_loss=0.0463, over 18434.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.2726, pruned_loss=0.05138, over 3586932.71 frames. ], batch size: 43, lr: 1.02e-02, grad_scale: 8.0 +2023-03-09 02:40:20,789 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9048, 3.8134, 4.9126, 2.6432, 4.2407, 2.5990, 3.0029, 1.7779], + device='cuda:0'), covar=tensor([0.0903, 0.0703, 0.0118, 0.0768, 0.0573, 0.2229, 0.2427, 0.1744], + device='cuda:0'), in_proj_covar=tensor([0.0196, 0.0215, 0.0118, 0.0170, 0.0229, 0.0247, 0.0282, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 02:40:22,105 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-09 02:40:30,184 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9535, 4.9183, 5.0271, 4.7928, 4.6934, 4.7944, 5.1235, 5.1571], + device='cuda:0'), covar=tensor([0.0054, 0.0061, 0.0058, 0.0093, 0.0059, 0.0112, 0.0066, 0.0088], + device='cuda:0'), in_proj_covar=tensor([0.0081, 0.0057, 0.0061, 0.0076, 0.0063, 0.0088, 0.0074, 0.0073], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 02:40:58,328 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39639.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 02:40:59,374 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39640.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:41:00,287 INFO [train.py:898] (0/4) Epoch 11, batch 3300, loss[loss=0.1767, simple_loss=0.2705, pruned_loss=0.04149, over 18342.00 frames. ], tot_loss[loss=0.1878, simple_loss=0.273, pruned_loss=0.0513, over 3586718.58 frames. ], batch size: 55, lr: 1.02e-02, grad_scale: 8.0 +2023-03-09 02:41:02,807 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.26 vs. limit=5.0 +2023-03-09 02:41:08,714 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.189e+02 3.111e+02 3.675e+02 4.353e+02 7.934e+02, threshold=7.351e+02, percent-clipped=2.0 +2023-03-09 02:41:13,707 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-09 02:41:20,551 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-03-09 02:41:34,204 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-09 02:41:53,927 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1038, 5.4728, 2.9797, 5.2034, 5.1353, 5.4275, 5.2124, 2.7042], + device='cuda:0'), covar=tensor([0.0155, 0.0041, 0.0667, 0.0073, 0.0061, 0.0065, 0.0076, 0.0925], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0066, 0.0087, 0.0081, 0.0076, 0.0066, 0.0076, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 02:41:59,410 INFO [train.py:898] (0/4) Epoch 11, batch 3350, loss[loss=0.1977, simple_loss=0.2892, pruned_loss=0.05312, over 17871.00 frames. ], tot_loss[loss=0.1882, simple_loss=0.2736, pruned_loss=0.0514, over 3590873.38 frames. ], batch size: 70, lr: 1.02e-02, grad_scale: 8.0 +2023-03-09 02:42:10,638 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39700.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 02:42:18,620 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5397, 2.7754, 2.5297, 2.7491, 3.5596, 3.5158, 3.0987, 3.0058], + device='cuda:0'), covar=tensor([0.0172, 0.0301, 0.0540, 0.0307, 0.0161, 0.0138, 0.0345, 0.0277], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0110, 0.0148, 0.0137, 0.0105, 0.0092, 0.0134, 0.0129], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 02:42:58,361 INFO [train.py:898] (0/4) Epoch 11, batch 3400, loss[loss=0.186, simple_loss=0.272, pruned_loss=0.05002, over 18300.00 frames. ], tot_loss[loss=0.1877, simple_loss=0.273, pruned_loss=0.05122, over 3597859.90 frames. ], batch size: 49, lr: 1.02e-02, grad_scale: 8.0 +2023-03-09 02:43:06,473 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.021e+02 3.201e+02 3.760e+02 4.727e+02 8.419e+02, threshold=7.521e+02, percent-clipped=1.0 +2023-03-09 02:43:29,956 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39767.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:43:57,996 INFO [train.py:898] (0/4) Epoch 11, batch 3450, loss[loss=0.1877, simple_loss=0.2745, pruned_loss=0.05045, over 18130.00 frames. ], tot_loss[loss=0.1874, simple_loss=0.2726, pruned_loss=0.05105, over 3590792.38 frames. ], batch size: 62, lr: 1.01e-02, grad_scale: 8.0 +2023-03-09 02:44:26,827 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=39815.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:44:31,324 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39819.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:44:42,935 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39828.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:44:53,129 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5242, 2.9319, 2.5628, 2.6811, 3.5081, 3.4430, 3.1358, 2.9552], + device='cuda:0'), covar=tensor([0.0138, 0.0263, 0.0596, 0.0350, 0.0184, 0.0176, 0.0328, 0.0317], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0110, 0.0149, 0.0136, 0.0105, 0.0093, 0.0134, 0.0130], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 02:44:57,104 INFO [train.py:898] (0/4) Epoch 11, batch 3500, loss[loss=0.1884, simple_loss=0.2562, pruned_loss=0.06035, over 18497.00 frames. ], tot_loss[loss=0.187, simple_loss=0.2723, pruned_loss=0.05083, over 3599579.44 frames. ], batch size: 44, lr: 1.01e-02, grad_scale: 8.0 +2023-03-09 02:44:57,336 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39841.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:45:05,096 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.228e+02 3.240e+02 3.890e+02 4.468e+02 8.251e+02, threshold=7.780e+02, percent-clipped=2.0 +2023-03-09 02:45:28,175 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.64 vs. limit=5.0 +2023-03-09 02:45:37,711 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=39876.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:45:52,194 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=39889.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:45:54,247 INFO [train.py:898] (0/4) Epoch 11, batch 3550, loss[loss=0.1842, simple_loss=0.2713, pruned_loss=0.04855, over 18545.00 frames. ], tot_loss[loss=0.1872, simple_loss=0.2724, pruned_loss=0.05095, over 3585484.53 frames. ], batch size: 54, lr: 1.01e-02, grad_scale: 8.0 +2023-03-09 02:45:57,170 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-03-09 02:46:48,079 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=39940.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:46:48,884 INFO [train.py:898] (0/4) Epoch 11, batch 3600, loss[loss=0.1995, simple_loss=0.287, pruned_loss=0.05596, over 18461.00 frames. ], tot_loss[loss=0.1866, simple_loss=0.2716, pruned_loss=0.05085, over 3590168.67 frames. ], batch size: 59, lr: 1.01e-02, grad_scale: 8.0 +2023-03-09 02:46:55,871 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.247e+02 3.210e+02 3.696e+02 4.808e+02 8.251e+02, threshold=7.392e+02, percent-clipped=2.0 +2023-03-09 02:47:23,983 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-11.pt +2023-03-09 02:47:53,935 INFO [train.py:898] (0/4) Epoch 12, batch 0, loss[loss=0.1961, simple_loss=0.2823, pruned_loss=0.05499, over 18408.00 frames. ], tot_loss[loss=0.1961, simple_loss=0.2823, pruned_loss=0.05499, over 18408.00 frames. ], batch size: 52, lr: 9.70e-03, grad_scale: 8.0 +2023-03-09 02:47:53,937 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 02:48:01,182 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7767, 5.9081, 4.1986, 5.7992, 5.5280, 5.9636, 5.7990, 4.0118], + device='cuda:0'), covar=tensor([0.0124, 0.0035, 0.0456, 0.0038, 0.0058, 0.0031, 0.0056, 0.0619], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0067, 0.0088, 0.0082, 0.0077, 0.0066, 0.0076, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0004, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 02:48:05,008 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([6.1257, 6.5247, 5.7445, 6.4360, 6.1211, 6.1756, 6.5952, 6.5131], + device='cuda:0'), covar=tensor([0.0873, 0.0473, 0.0236, 0.0392, 0.1003, 0.0492, 0.0322, 0.0444], + device='cuda:0'), in_proj_covar=tensor([0.0517, 0.0429, 0.0324, 0.0455, 0.0630, 0.0461, 0.0605, 0.0452], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 02:48:05,855 INFO [train.py:932] (0/4) Epoch 12, validation: loss=0.1577, simple_loss=0.2601, pruned_loss=0.02771, over 944034.00 frames. +2023-03-09 02:48:05,855 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-09 02:48:21,414 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=39988.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:48:29,611 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=39995.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 02:48:36,669 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-40000.pt +2023-03-09 02:49:09,610 INFO [train.py:898] (0/4) Epoch 12, batch 50, loss[loss=0.2059, simple_loss=0.29, pruned_loss=0.06087, over 18294.00 frames. ], tot_loss[loss=0.1936, simple_loss=0.2787, pruned_loss=0.05427, over 800269.58 frames. ], batch size: 57, lr: 9.69e-03, grad_scale: 8.0 +2023-03-09 02:49:35,920 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.813e+02 3.496e+02 4.076e+02 5.396e+02 1.029e+03, threshold=8.152e+02, percent-clipped=4.0 +2023-03-09 02:50:08,398 INFO [train.py:898] (0/4) Epoch 12, batch 100, loss[loss=0.2024, simple_loss=0.2916, pruned_loss=0.05657, over 18570.00 frames. ], tot_loss[loss=0.1906, simple_loss=0.2756, pruned_loss=0.05282, over 1413556.62 frames. ], batch size: 54, lr: 9.69e-03, grad_scale: 4.0 +2023-03-09 02:50:44,871 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1958, 3.9359, 5.3866, 3.2501, 4.5426, 2.6650, 3.0544, 2.0794], + device='cuda:0'), covar=tensor([0.0821, 0.0706, 0.0065, 0.0677, 0.0466, 0.2203, 0.2582, 0.1714], + device='cuda:0'), in_proj_covar=tensor([0.0198, 0.0217, 0.0119, 0.0171, 0.0231, 0.0248, 0.0285, 0.0211], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 02:50:55,096 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.3720, 2.1659, 2.1126, 2.0497, 2.5063, 2.3117, 2.3210, 2.1024], + device='cuda:0'), covar=tensor([0.0207, 0.0212, 0.0360, 0.0296, 0.0166, 0.0138, 0.0282, 0.0243], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0108, 0.0146, 0.0135, 0.0103, 0.0092, 0.0133, 0.0128], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 02:51:00,687 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40119.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:51:05,210 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40123.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:51:07,322 INFO [train.py:898] (0/4) Epoch 12, batch 150, loss[loss=0.1796, simple_loss=0.2652, pruned_loss=0.04701, over 18489.00 frames. ], tot_loss[loss=0.1875, simple_loss=0.272, pruned_loss=0.05149, over 1896802.64 frames. ], batch size: 51, lr: 9.68e-03, grad_scale: 4.0 +2023-03-09 02:51:08,903 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1130, 3.3590, 4.5324, 4.0603, 3.1297, 4.9518, 4.3025, 3.0292], + device='cuda:0'), covar=tensor([0.0356, 0.1116, 0.0193, 0.0321, 0.1242, 0.0122, 0.0327, 0.0963], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0226, 0.0155, 0.0148, 0.0215, 0.0188, 0.0211, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 02:51:31,330 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-09 02:51:36,219 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.055e+02 3.093e+02 3.775e+02 4.453e+02 9.107e+02, threshold=7.551e+02, percent-clipped=1.0 +2023-03-09 02:51:39,183 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-09 02:51:46,083 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-09 02:51:54,573 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.76 vs. limit=2.0 +2023-03-09 02:51:57,395 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=40167.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:52:01,936 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40171.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:52:06,355 INFO [train.py:898] (0/4) Epoch 12, batch 200, loss[loss=0.1922, simple_loss=0.2832, pruned_loss=0.05061, over 17959.00 frames. ], tot_loss[loss=0.1869, simple_loss=0.2714, pruned_loss=0.0512, over 2267560.39 frames. ], batch size: 65, lr: 9.68e-03, grad_scale: 4.0 +2023-03-09 02:52:21,027 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5912, 5.5062, 5.0902, 5.4988, 5.4384, 4.8802, 5.3828, 5.0987], + device='cuda:0'), covar=tensor([0.0352, 0.0348, 0.1430, 0.0730, 0.0532, 0.0400, 0.0392, 0.0989], + device='cuda:0'), in_proj_covar=tensor([0.0399, 0.0463, 0.0614, 0.0363, 0.0348, 0.0420, 0.0450, 0.0578], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 02:52:29,878 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40195.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:52:58,306 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40218.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:53:06,000 INFO [train.py:898] (0/4) Epoch 12, batch 250, loss[loss=0.1829, simple_loss=0.2687, pruned_loss=0.04851, over 18511.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2711, pruned_loss=0.05074, over 2573191.18 frames. ], batch size: 59, lr: 9.67e-03, grad_scale: 4.0 +2023-03-09 02:53:33,645 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.098e+02 3.173e+02 3.973e+02 4.861e+02 1.364e+03, threshold=7.946e+02, percent-clipped=3.0 +2023-03-09 02:53:42,346 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40256.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:54:01,818 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40272.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:54:05,013 INFO [train.py:898] (0/4) Epoch 12, batch 300, loss[loss=0.187, simple_loss=0.2764, pruned_loss=0.04883, over 18563.00 frames. ], tot_loss[loss=0.1848, simple_loss=0.2698, pruned_loss=0.04991, over 2813132.27 frames. ], batch size: 54, lr: 9.66e-03, grad_scale: 4.0 +2023-03-09 02:54:09,948 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40279.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:54:28,450 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40295.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 02:55:04,169 INFO [train.py:898] (0/4) Epoch 12, batch 350, loss[loss=0.1611, simple_loss=0.2404, pruned_loss=0.04097, over 17616.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2694, pruned_loss=0.04981, over 2994972.96 frames. ], batch size: 39, lr: 9.66e-03, grad_scale: 4.0 +2023-03-09 02:55:10,089 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40330.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:55:13,523 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40333.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:55:24,772 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=40343.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 02:55:31,752 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.235e+02 3.021e+02 3.795e+02 4.582e+02 1.168e+03, threshold=7.590e+02, percent-clipped=5.0 +2023-03-09 02:56:02,389 INFO [train.py:898] (0/4) Epoch 12, batch 400, loss[loss=0.1894, simple_loss=0.2797, pruned_loss=0.0495, over 18319.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2694, pruned_loss=0.04969, over 3133627.61 frames. ], batch size: 54, lr: 9.65e-03, grad_scale: 8.0 +2023-03-09 02:56:11,357 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8569, 2.8396, 4.2161, 3.7677, 2.7462, 4.6175, 4.0428, 2.9355], + device='cuda:0'), covar=tensor([0.0333, 0.1268, 0.0209, 0.0322, 0.1320, 0.0144, 0.0388, 0.0899], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0224, 0.0155, 0.0146, 0.0214, 0.0188, 0.0212, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 02:56:21,547 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40391.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:56:59,491 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40423.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:57:01,660 INFO [train.py:898] (0/4) Epoch 12, batch 450, loss[loss=0.1821, simple_loss=0.2737, pruned_loss=0.04522, over 18575.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.269, pruned_loss=0.04939, over 3239392.58 frames. ], batch size: 54, lr: 9.65e-03, grad_scale: 8.0 +2023-03-09 02:57:02,100 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8140, 3.6964, 3.4508, 3.1631, 3.4310, 2.7603, 2.8002, 3.6355], + device='cuda:0'), covar=tensor([0.0036, 0.0063, 0.0059, 0.0107, 0.0079, 0.0152, 0.0160, 0.0057], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0122, 0.0106, 0.0155, 0.0108, 0.0152, 0.0157, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 02:57:17,285 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.50 vs. limit=5.0 +2023-03-09 02:57:29,713 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.841e+02 3.105e+02 3.603e+02 4.102e+02 7.155e+02, threshold=7.205e+02, percent-clipped=0.0 +2023-03-09 02:57:55,782 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=40471.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:57:55,848 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40471.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:58:00,091 INFO [train.py:898] (0/4) Epoch 12, batch 500, loss[loss=0.1684, simple_loss=0.2446, pruned_loss=0.04609, over 18405.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2702, pruned_loss=0.04987, over 3315078.99 frames. ], batch size: 42, lr: 9.64e-03, grad_scale: 8.0 +2023-03-09 02:58:04,005 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2321, 4.3688, 2.4050, 4.1872, 5.3144, 2.8232, 3.6266, 3.6832], + device='cuda:0'), covar=tensor([0.0101, 0.1122, 0.1698, 0.0581, 0.0051, 0.1289, 0.0859, 0.0991], + device='cuda:0'), in_proj_covar=tensor([0.0120, 0.0237, 0.0191, 0.0190, 0.0092, 0.0179, 0.0203, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 02:58:20,972 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40492.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:58:51,812 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=40519.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:58:59,237 INFO [train.py:898] (0/4) Epoch 12, batch 550, loss[loss=0.1801, simple_loss=0.2725, pruned_loss=0.04383, over 17798.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2708, pruned_loss=0.05011, over 3373063.45 frames. ], batch size: 70, lr: 9.63e-03, grad_scale: 8.0 +2023-03-09 02:59:26,876 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.116e+02 3.269e+02 4.036e+02 4.725e+02 9.923e+02, threshold=8.073e+02, percent-clipped=3.0 +2023-03-09 02:59:29,406 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40551.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:59:31,797 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40553.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:59:40,999 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5692, 2.8127, 2.3909, 2.7498, 3.5912, 3.6069, 2.9917, 2.8779], + device='cuda:0'), covar=tensor([0.0173, 0.0291, 0.0623, 0.0378, 0.0145, 0.0114, 0.0358, 0.0371], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0110, 0.0151, 0.0140, 0.0105, 0.0093, 0.0136, 0.0132], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 02:59:56,942 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40574.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 02:59:57,875 INFO [train.py:898] (0/4) Epoch 12, batch 600, loss[loss=0.2068, simple_loss=0.2923, pruned_loss=0.06069, over 18229.00 frames. ], tot_loss[loss=0.1862, simple_loss=0.2716, pruned_loss=0.05043, over 3412666.59 frames. ], batch size: 60, lr: 9.63e-03, grad_scale: 4.0 +2023-03-09 03:00:21,439 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-09 03:00:32,697 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0153, 4.5287, 4.6572, 3.2545, 3.6690, 3.5429, 2.6190, 2.1893], + device='cuda:0'), covar=tensor([0.0189, 0.0116, 0.0063, 0.0296, 0.0302, 0.0201, 0.0702, 0.0928], + device='cuda:0'), in_proj_covar=tensor([0.0059, 0.0047, 0.0048, 0.0060, 0.0079, 0.0057, 0.0071, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 03:00:56,611 INFO [train.py:898] (0/4) Epoch 12, batch 650, loss[loss=0.2075, simple_loss=0.2938, pruned_loss=0.06057, over 18642.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2713, pruned_loss=0.04996, over 3454366.28 frames. ], batch size: 52, lr: 9.62e-03, grad_scale: 4.0 +2023-03-09 03:01:00,797 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40628.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:01:26,540 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.880e+02 2.745e+02 3.552e+02 4.267e+02 1.309e+03, threshold=7.104e+02, percent-clipped=1.0 +2023-03-09 03:01:28,126 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40651.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:01:51,226 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0880, 5.0725, 4.6572, 5.0625, 4.9882, 4.4514, 4.9338, 4.7172], + device='cuda:0'), covar=tensor([0.0461, 0.0433, 0.1475, 0.0577, 0.0533, 0.0461, 0.0429, 0.0965], + device='cuda:0'), in_proj_covar=tensor([0.0417, 0.0478, 0.0636, 0.0375, 0.0356, 0.0433, 0.0467, 0.0600], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 03:01:55,471 INFO [train.py:898] (0/4) Epoch 12, batch 700, loss[loss=0.1887, simple_loss=0.2671, pruned_loss=0.05515, over 18492.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2714, pruned_loss=0.04996, over 3485276.93 frames. ], batch size: 47, lr: 9.62e-03, grad_scale: 4.0 +2023-03-09 03:02:08,732 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40686.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:02:39,447 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40712.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:02:54,261 INFO [train.py:898] (0/4) Epoch 12, batch 750, loss[loss=0.1831, simple_loss=0.2618, pruned_loss=0.05224, over 18269.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.271, pruned_loss=0.0501, over 3508254.38 frames. ], batch size: 47, lr: 9.61e-03, grad_scale: 4.0 +2023-03-09 03:03:25,289 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.144e+02 3.250e+02 3.776e+02 4.415e+02 7.567e+02, threshold=7.551e+02, percent-clipped=1.0 +2023-03-09 03:03:26,692 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0640, 5.0618, 5.1526, 5.1344, 5.0511, 5.7503, 5.3767, 5.1345], + device='cuda:0'), covar=tensor([0.1081, 0.0712, 0.0743, 0.0749, 0.1390, 0.0723, 0.0650, 0.1516], + device='cuda:0'), in_proj_covar=tensor([0.0303, 0.0238, 0.0249, 0.0252, 0.0287, 0.0357, 0.0236, 0.0343], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 03:03:35,700 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1009, 5.1664, 5.3179, 5.2679, 5.2540, 5.9469, 5.6157, 5.4017], + device='cuda:0'), covar=tensor([0.1109, 0.0601, 0.0679, 0.0654, 0.1311, 0.0695, 0.0656, 0.1389], + device='cuda:0'), in_proj_covar=tensor([0.0303, 0.0237, 0.0249, 0.0252, 0.0287, 0.0356, 0.0236, 0.0343], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 03:03:52,298 INFO [train.py:898] (0/4) Epoch 12, batch 800, loss[loss=0.1809, simple_loss=0.2696, pruned_loss=0.04609, over 18354.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2716, pruned_loss=0.0501, over 3530040.96 frames. ], batch size: 50, lr: 9.61e-03, grad_scale: 4.0 +2023-03-09 03:04:14,001 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7491, 5.2644, 5.2116, 5.2633, 4.8171, 5.1190, 4.5330, 5.1867], + device='cuda:0'), covar=tensor([0.0249, 0.0270, 0.0232, 0.0350, 0.0355, 0.0211, 0.1158, 0.0244], + device='cuda:0'), in_proj_covar=tensor([0.0181, 0.0225, 0.0218, 0.0259, 0.0229, 0.0232, 0.0289, 0.0215], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-09 03:04:15,128 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5152, 5.2750, 5.6454, 5.6390, 5.4225, 6.2222, 5.9030, 5.5939], + device='cuda:0'), covar=tensor([0.1117, 0.0696, 0.0731, 0.0708, 0.1774, 0.0873, 0.0626, 0.1914], + device='cuda:0'), in_proj_covar=tensor([0.0305, 0.0238, 0.0251, 0.0252, 0.0290, 0.0358, 0.0238, 0.0344], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0004, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 03:04:23,307 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.32 vs. limit=5.0 +2023-03-09 03:04:51,191 INFO [train.py:898] (0/4) Epoch 12, batch 850, loss[loss=0.1738, simple_loss=0.2505, pruned_loss=0.04849, over 18590.00 frames. ], tot_loss[loss=0.1865, simple_loss=0.2722, pruned_loss=0.05037, over 3544408.72 frames. ], batch size: 45, lr: 9.60e-03, grad_scale: 4.0 +2023-03-09 03:05:17,983 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=40848.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:05:21,110 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.184e+02 3.104e+02 3.641e+02 4.538e+02 1.053e+03, threshold=7.281e+02, percent-clipped=3.0 +2023-03-09 03:05:21,341 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40851.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:05:26,258 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5745, 2.2415, 2.6466, 2.6539, 3.5166, 5.2235, 4.7985, 4.1113], + device='cuda:0'), covar=tensor([0.1327, 0.1955, 0.2408, 0.1429, 0.1657, 0.0095, 0.0314, 0.0456], + device='cuda:0'), in_proj_covar=tensor([0.0246, 0.0302, 0.0321, 0.0250, 0.0365, 0.0189, 0.0262, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 03:05:38,510 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40865.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:05:40,985 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5372, 1.9694, 2.3790, 2.4502, 3.3430, 5.1237, 4.7212, 4.0820], + device='cuda:0'), covar=tensor([0.1301, 0.2201, 0.2647, 0.1525, 0.1817, 0.0101, 0.0322, 0.0457], + device='cuda:0'), in_proj_covar=tensor([0.0248, 0.0305, 0.0324, 0.0252, 0.0368, 0.0190, 0.0264, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 03:05:48,701 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40874.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:05:49,596 INFO [train.py:898] (0/4) Epoch 12, batch 900, loss[loss=0.1873, simple_loss=0.2789, pruned_loss=0.04779, over 18284.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.272, pruned_loss=0.05026, over 3558544.32 frames. ], batch size: 57, lr: 9.59e-03, grad_scale: 4.0 +2023-03-09 03:06:18,119 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=40899.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:06:19,586 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40900.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:06:45,190 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=40922.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:06:48,428 INFO [train.py:898] (0/4) Epoch 12, batch 950, loss[loss=0.1883, simple_loss=0.2817, pruned_loss=0.04739, over 18343.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2717, pruned_loss=0.05002, over 3569833.01 frames. ], batch size: 55, lr: 9.59e-03, grad_scale: 4.0 +2023-03-09 03:06:49,940 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40926.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:06:52,129 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40928.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:07:18,899 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.424e+02 3.344e+02 4.191e+02 5.324e+02 1.167e+03, threshold=8.382e+02, percent-clipped=6.0 +2023-03-09 03:07:31,642 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=40961.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 03:07:46,841 INFO [train.py:898] (0/4) Epoch 12, batch 1000, loss[loss=0.1866, simple_loss=0.2841, pruned_loss=0.04458, over 18214.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.271, pruned_loss=0.04967, over 3572775.90 frames. ], batch size: 60, lr: 9.58e-03, grad_scale: 4.0 +2023-03-09 03:07:47,969 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=40976.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:07:59,461 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=40986.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:08:11,881 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=40997.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:08:24,339 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41007.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:08:37,309 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-09 03:08:45,143 INFO [train.py:898] (0/4) Epoch 12, batch 1050, loss[loss=0.1942, simple_loss=0.2858, pruned_loss=0.05127, over 18480.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2716, pruned_loss=0.05005, over 3571604.61 frames. ], batch size: 53, lr: 9.58e-03, grad_scale: 4.0 +2023-03-09 03:08:55,723 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=41034.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:09:08,614 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41045.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:09:14,993 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.197e+02 3.189e+02 3.617e+02 4.210e+02 9.012e+02, threshold=7.234e+02, percent-clipped=1.0 +2023-03-09 03:09:18,882 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1867, 4.3587, 2.5127, 4.2194, 5.2837, 2.4756, 3.7046, 3.8167], + device='cuda:0'), covar=tensor([0.0092, 0.1030, 0.1583, 0.0552, 0.0053, 0.1450, 0.0751, 0.0766], + device='cuda:0'), in_proj_covar=tensor([0.0119, 0.0234, 0.0191, 0.0189, 0.0092, 0.0177, 0.0203, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 03:09:24,025 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41058.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:09:43,919 INFO [train.py:898] (0/4) Epoch 12, batch 1100, loss[loss=0.1934, simple_loss=0.2828, pruned_loss=0.05201, over 16537.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2714, pruned_loss=0.04986, over 3570065.80 frames. ], batch size: 94, lr: 9.57e-03, grad_scale: 4.0 +2023-03-09 03:10:15,216 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41102.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:10:19,422 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41106.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:10:41,772 INFO [train.py:898] (0/4) Epoch 12, batch 1150, loss[loss=0.1684, simple_loss=0.2527, pruned_loss=0.04205, over 18494.00 frames. ], tot_loss[loss=0.1859, simple_loss=0.2718, pruned_loss=0.04997, over 3577419.87 frames. ], batch size: 47, lr: 9.56e-03, grad_scale: 4.0 +2023-03-09 03:11:07,795 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41148.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:11:11,014 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.262e+02 3.020e+02 3.568e+02 4.440e+02 1.142e+03, threshold=7.137e+02, percent-clipped=4.0 +2023-03-09 03:11:14,227 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5235, 2.0399, 2.6752, 2.5645, 3.3484, 4.9521, 4.5097, 3.9047], + device='cuda:0'), covar=tensor([0.1358, 0.2127, 0.2350, 0.1485, 0.1758, 0.0120, 0.0399, 0.0513], + device='cuda:0'), in_proj_covar=tensor([0.0249, 0.0307, 0.0328, 0.0253, 0.0370, 0.0191, 0.0266, 0.0210], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 03:11:25,766 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41163.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:11:39,968 INFO [train.py:898] (0/4) Epoch 12, batch 1200, loss[loss=0.2023, simple_loss=0.2791, pruned_loss=0.06279, over 18509.00 frames. ], tot_loss[loss=0.1857, simple_loss=0.2713, pruned_loss=0.05004, over 3588724.15 frames. ], batch size: 47, lr: 9.56e-03, grad_scale: 8.0 +2023-03-09 03:12:03,587 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=41196.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:12:04,930 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6549, 3.2519, 4.7538, 3.0099, 4.1237, 2.4338, 2.8125, 1.9992], + device='cuda:0'), covar=tensor([0.1026, 0.0946, 0.0101, 0.0597, 0.0551, 0.2407, 0.2413, 0.1806], + device='cuda:0'), in_proj_covar=tensor([0.0200, 0.0219, 0.0123, 0.0173, 0.0232, 0.0250, 0.0289, 0.0213], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 03:12:34,319 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41221.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:12:38,721 INFO [train.py:898] (0/4) Epoch 12, batch 1250, loss[loss=0.1585, simple_loss=0.2466, pruned_loss=0.03518, over 18358.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2719, pruned_loss=0.05038, over 3582303.46 frames. ], batch size: 46, lr: 9.55e-03, grad_scale: 8.0 +2023-03-09 03:13:08,569 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.122e+02 3.147e+02 3.624e+02 4.404e+02 8.408e+02, threshold=7.247e+02, percent-clipped=2.0 +2023-03-09 03:13:14,381 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41256.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 03:13:37,228 INFO [train.py:898] (0/4) Epoch 12, batch 1300, loss[loss=0.2045, simple_loss=0.2751, pruned_loss=0.06698, over 18264.00 frames. ], tot_loss[loss=0.1868, simple_loss=0.2723, pruned_loss=0.0507, over 3573544.58 frames. ], batch size: 45, lr: 9.55e-03, grad_scale: 8.0 +2023-03-09 03:14:14,021 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41307.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:14:35,259 INFO [train.py:898] (0/4) Epoch 12, batch 1350, loss[loss=0.1546, simple_loss=0.2429, pruned_loss=0.03317, over 18414.00 frames. ], tot_loss[loss=0.1868, simple_loss=0.2722, pruned_loss=0.05067, over 3572158.20 frames. ], batch size: 48, lr: 9.54e-03, grad_scale: 8.0 +2023-03-09 03:14:57,228 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7366, 3.3394, 5.2227, 3.0487, 4.4801, 2.6307, 3.0424, 1.9518], + device='cuda:0'), covar=tensor([0.1046, 0.0956, 0.0069, 0.0656, 0.0470, 0.2261, 0.2317, 0.1813], + device='cuda:0'), in_proj_covar=tensor([0.0199, 0.0219, 0.0121, 0.0173, 0.0231, 0.0249, 0.0287, 0.0213], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 03:15:05,737 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.782e+02 3.011e+02 3.906e+02 4.787e+02 1.227e+03, threshold=7.812e+02, percent-clipped=6.0 +2023-03-09 03:15:08,149 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41353.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:15:10,463 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=41355.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:15:34,028 INFO [train.py:898] (0/4) Epoch 12, batch 1400, loss[loss=0.2013, simple_loss=0.2885, pruned_loss=0.05712, over 17835.00 frames. ], tot_loss[loss=0.1861, simple_loss=0.2714, pruned_loss=0.05042, over 3577784.99 frames. ], batch size: 70, lr: 9.54e-03, grad_scale: 8.0 +2023-03-09 03:15:36,563 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9397, 4.9278, 5.0985, 5.0560, 4.9516, 5.6921, 5.2689, 5.0033], + device='cuda:0'), covar=tensor([0.1110, 0.0664, 0.0690, 0.0728, 0.1341, 0.0758, 0.0647, 0.1559], + device='cuda:0'), in_proj_covar=tensor([0.0314, 0.0239, 0.0256, 0.0257, 0.0296, 0.0363, 0.0242, 0.0354], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0004, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 03:16:04,719 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41401.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:16:32,786 INFO [train.py:898] (0/4) Epoch 12, batch 1450, loss[loss=0.1912, simple_loss=0.2798, pruned_loss=0.05131, over 18037.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2709, pruned_loss=0.05016, over 3588923.34 frames. ], batch size: 65, lr: 9.53e-03, grad_scale: 8.0 +2023-03-09 03:16:38,804 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-09 03:17:03,286 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.105e+02 2.980e+02 3.633e+02 4.455e+02 8.287e+02, threshold=7.266e+02, percent-clipped=1.0 +2023-03-09 03:17:11,395 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=41458.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:17:31,244 INFO [train.py:898] (0/4) Epoch 12, batch 1500, loss[loss=0.1987, simple_loss=0.2813, pruned_loss=0.05802, over 17981.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2704, pruned_loss=0.05001, over 3585204.47 frames. ], batch size: 65, lr: 9.52e-03, grad_scale: 8.0 +2023-03-09 03:18:22,214 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8431, 2.9008, 2.0905, 3.2593, 2.4419, 2.9947, 2.2661, 2.8514], + device='cuda:0'), covar=tensor([0.0527, 0.0731, 0.1129, 0.0537, 0.0781, 0.0323, 0.0951, 0.0377], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0213, 0.0179, 0.0243, 0.0180, 0.0244, 0.0191, 0.0186], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 03:18:24,459 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41521.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:18:28,667 INFO [train.py:898] (0/4) Epoch 12, batch 1550, loss[loss=0.1989, simple_loss=0.2824, pruned_loss=0.05767, over 16977.00 frames. ], tot_loss[loss=0.1855, simple_loss=0.2705, pruned_loss=0.05024, over 3585274.76 frames. ], batch size: 78, lr: 9.52e-03, grad_scale: 8.0 +2023-03-09 03:19:00,403 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.006e+02 3.193e+02 3.688e+02 4.708e+02 1.427e+03, threshold=7.376e+02, percent-clipped=3.0 +2023-03-09 03:19:06,203 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41556.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 03:19:20,679 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=41569.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:19:27,333 INFO [train.py:898] (0/4) Epoch 12, batch 1600, loss[loss=0.1766, simple_loss=0.2696, pruned_loss=0.04175, over 17760.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.27, pruned_loss=0.05012, over 3584641.99 frames. ], batch size: 70, lr: 9.51e-03, grad_scale: 8.0 +2023-03-09 03:20:02,985 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=41604.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:20:18,011 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8551, 3.6369, 4.9535, 3.0464, 4.1447, 2.6355, 3.0214, 1.7081], + device='cuda:0'), covar=tensor([0.0953, 0.0831, 0.0095, 0.0702, 0.0650, 0.2173, 0.2490, 0.1923], + device='cuda:0'), in_proj_covar=tensor([0.0198, 0.0221, 0.0123, 0.0175, 0.0233, 0.0248, 0.0292, 0.0213], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 03:20:26,595 INFO [train.py:898] (0/4) Epoch 12, batch 1650, loss[loss=0.2043, simple_loss=0.2934, pruned_loss=0.0576, over 18346.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2703, pruned_loss=0.04987, over 3598953.27 frames. ], batch size: 56, lr: 9.51e-03, grad_scale: 8.0 +2023-03-09 03:20:58,271 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.922e+02 3.190e+02 3.689e+02 4.449e+02 1.002e+03, threshold=7.378e+02, percent-clipped=1.0 +2023-03-09 03:21:00,802 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3264, 5.3017, 4.9241, 5.2071, 5.2589, 4.6243, 5.1744, 4.9228], + device='cuda:0'), covar=tensor([0.0392, 0.0410, 0.1244, 0.0766, 0.0480, 0.0445, 0.0386, 0.1047], + device='cuda:0'), in_proj_covar=tensor([0.0421, 0.0489, 0.0635, 0.0378, 0.0362, 0.0439, 0.0471, 0.0606], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 03:21:00,832 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41653.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:21:01,987 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41654.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:21:16,274 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-09 03:21:25,771 INFO [train.py:898] (0/4) Epoch 12, batch 1700, loss[loss=0.1974, simple_loss=0.2824, pruned_loss=0.05626, over 18347.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.2697, pruned_loss=0.04939, over 3595889.28 frames. ], batch size: 56, lr: 9.50e-03, grad_scale: 8.0 +2023-03-09 03:21:47,245 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.4628, 3.4776, 5.0327, 4.3433, 3.1939, 3.0073, 4.5225, 5.1479], + device='cuda:0'), covar=tensor([0.0822, 0.1808, 0.0115, 0.0330, 0.0900, 0.1124, 0.0294, 0.0222], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0248, 0.0109, 0.0162, 0.0179, 0.0176, 0.0174, 0.0152], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 03:21:57,720 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=41701.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:21:57,919 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41701.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:22:10,361 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8315, 4.4316, 4.4428, 3.3887, 3.6713, 3.6011, 2.5431, 2.3581], + device='cuda:0'), covar=tensor([0.0209, 0.0158, 0.0079, 0.0270, 0.0293, 0.0191, 0.0693, 0.0826], + device='cuda:0'), in_proj_covar=tensor([0.0059, 0.0047, 0.0048, 0.0060, 0.0079, 0.0058, 0.0071, 0.0078], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 03:22:13,777 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41715.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 03:22:24,613 INFO [train.py:898] (0/4) Epoch 12, batch 1750, loss[loss=0.2434, simple_loss=0.313, pruned_loss=0.08692, over 12450.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.2696, pruned_loss=0.04944, over 3597302.53 frames. ], batch size: 130, lr: 9.50e-03, grad_scale: 8.0 +2023-03-09 03:22:53,564 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=41749.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:22:56,197 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.910e+02 3.102e+02 3.582e+02 4.165e+02 6.416e+02, threshold=7.165e+02, percent-clipped=1.0 +2023-03-09 03:23:04,301 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=41758.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:23:07,735 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1213, 4.1915, 2.2741, 4.3712, 5.1695, 2.6338, 3.9392, 3.8903], + device='cuda:0'), covar=tensor([0.0097, 0.0993, 0.1672, 0.0421, 0.0061, 0.1205, 0.0584, 0.0712], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0234, 0.0190, 0.0188, 0.0093, 0.0176, 0.0203, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 03:23:22,852 INFO [train.py:898] (0/4) Epoch 12, batch 1800, loss[loss=0.2101, simple_loss=0.2927, pruned_loss=0.06372, over 17764.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2696, pruned_loss=0.04946, over 3598537.62 frames. ], batch size: 70, lr: 9.49e-03, grad_scale: 8.0 +2023-03-09 03:23:59,579 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=41806.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:24:00,894 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8735, 4.5030, 4.6404, 3.3295, 3.6160, 3.7329, 2.7019, 2.3480], + device='cuda:0'), covar=tensor([0.0211, 0.0186, 0.0063, 0.0298, 0.0320, 0.0176, 0.0688, 0.0913], + device='cuda:0'), in_proj_covar=tensor([0.0060, 0.0048, 0.0048, 0.0060, 0.0080, 0.0057, 0.0071, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 03:24:20,774 INFO [train.py:898] (0/4) Epoch 12, batch 1850, loss[loss=0.1875, simple_loss=0.2771, pruned_loss=0.04888, over 18389.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2693, pruned_loss=0.04963, over 3590855.57 frames. ], batch size: 52, lr: 9.49e-03, grad_scale: 8.0 +2023-03-09 03:24:51,662 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.068e+02 3.165e+02 3.584e+02 4.367e+02 1.390e+03, threshold=7.168e+02, percent-clipped=5.0 +2023-03-09 03:25:09,494 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1644, 3.4387, 3.4192, 2.8500, 3.0590, 2.9536, 2.3591, 2.1285], + device='cuda:0'), covar=tensor([0.0229, 0.0144, 0.0095, 0.0256, 0.0290, 0.0210, 0.0596, 0.0742], + device='cuda:0'), in_proj_covar=tensor([0.0060, 0.0048, 0.0048, 0.0061, 0.0080, 0.0058, 0.0071, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 03:25:19,163 INFO [train.py:898] (0/4) Epoch 12, batch 1900, loss[loss=0.2011, simple_loss=0.2864, pruned_loss=0.05791, over 17828.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2701, pruned_loss=0.04986, over 3595782.63 frames. ], batch size: 70, lr: 9.48e-03, grad_scale: 8.0 +2023-03-09 03:25:27,532 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8007, 3.2489, 4.4514, 3.9467, 2.6286, 4.7959, 4.2150, 3.0536], + device='cuda:0'), covar=tensor([0.0463, 0.1224, 0.0207, 0.0371, 0.1599, 0.0181, 0.0441, 0.0908], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0225, 0.0159, 0.0144, 0.0211, 0.0186, 0.0211, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 03:25:42,512 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5004, 5.4584, 5.0527, 5.4528, 5.4316, 4.7452, 5.3470, 5.0869], + device='cuda:0'), covar=tensor([0.0347, 0.0372, 0.1084, 0.0534, 0.0444, 0.0414, 0.0322, 0.0979], + device='cuda:0'), in_proj_covar=tensor([0.0416, 0.0487, 0.0631, 0.0370, 0.0359, 0.0433, 0.0464, 0.0597], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 03:25:45,888 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.62 vs. limit=2.0 +2023-03-09 03:25:47,115 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.88 vs. limit=2.0 +2023-03-09 03:25:56,942 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41906.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:26:17,870 INFO [train.py:898] (0/4) Epoch 12, batch 1950, loss[loss=0.1842, simple_loss=0.273, pruned_loss=0.04766, over 18379.00 frames. ], tot_loss[loss=0.1863, simple_loss=0.2712, pruned_loss=0.05066, over 3578623.79 frames. ], batch size: 56, lr: 9.47e-03, grad_scale: 8.0 +2023-03-09 03:26:44,507 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=41948.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 03:26:47,526 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.082e+02 3.043e+02 3.852e+02 4.648e+02 1.107e+03, threshold=7.704e+02, percent-clipped=2.0 +2023-03-09 03:27:07,330 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=41967.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:27:16,288 INFO [train.py:898] (0/4) Epoch 12, batch 2000, loss[loss=0.2039, simple_loss=0.289, pruned_loss=0.05936, over 18092.00 frames. ], tot_loss[loss=0.1856, simple_loss=0.2706, pruned_loss=0.05024, over 3589802.74 frames. ], batch size: 62, lr: 9.47e-03, grad_scale: 8.0 +2023-03-09 03:27:22,195 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8976, 4.5187, 4.6353, 3.4194, 3.7392, 3.6615, 2.3896, 2.1105], + device='cuda:0'), covar=tensor([0.0227, 0.0136, 0.0060, 0.0258, 0.0295, 0.0178, 0.0747, 0.0903], + device='cuda:0'), in_proj_covar=tensor([0.0062, 0.0049, 0.0049, 0.0061, 0.0082, 0.0059, 0.0072, 0.0081], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0006, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 03:27:45,341 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-42000.pt +2023-03-09 03:28:02,108 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42009.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 03:28:03,174 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42010.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 03:28:20,369 INFO [train.py:898] (0/4) Epoch 12, batch 2050, loss[loss=0.1741, simple_loss=0.2566, pruned_loss=0.04584, over 18492.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2702, pruned_loss=0.05004, over 3581643.58 frames. ], batch size: 51, lr: 9.46e-03, grad_scale: 8.0 +2023-03-09 03:28:50,578 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.053e+02 3.014e+02 3.595e+02 4.305e+02 8.922e+02, threshold=7.191e+02, percent-clipped=2.0 +2023-03-09 03:29:19,680 INFO [train.py:898] (0/4) Epoch 12, batch 2100, loss[loss=0.1753, simple_loss=0.2654, pruned_loss=0.04257, over 18626.00 frames. ], tot_loss[loss=0.1854, simple_loss=0.2703, pruned_loss=0.05026, over 3568007.10 frames. ], batch size: 52, lr: 9.46e-03, grad_scale: 4.0 +2023-03-09 03:30:18,700 INFO [train.py:898] (0/4) Epoch 12, batch 2150, loss[loss=0.2026, simple_loss=0.2867, pruned_loss=0.05929, over 18612.00 frames. ], tot_loss[loss=0.1845, simple_loss=0.2698, pruned_loss=0.04957, over 3582624.05 frames. ], batch size: 52, lr: 9.45e-03, grad_scale: 4.0 +2023-03-09 03:30:49,400 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.910e+02 3.158e+02 3.739e+02 4.529e+02 7.533e+02, threshold=7.477e+02, percent-clipped=1.0 +2023-03-09 03:31:17,039 INFO [train.py:898] (0/4) Epoch 12, batch 2200, loss[loss=0.1576, simple_loss=0.2427, pruned_loss=0.03621, over 18256.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2702, pruned_loss=0.04979, over 3580420.85 frames. ], batch size: 47, lr: 9.45e-03, grad_scale: 4.0 +2023-03-09 03:31:28,606 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4774, 6.0651, 5.4599, 5.7899, 5.5667, 5.4679, 6.0355, 6.0503], + device='cuda:0'), covar=tensor([0.1085, 0.0600, 0.0436, 0.0633, 0.1318, 0.0692, 0.0540, 0.0597], + device='cuda:0'), in_proj_covar=tensor([0.0539, 0.0445, 0.0341, 0.0475, 0.0658, 0.0484, 0.0630, 0.0471], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 03:32:15,669 INFO [train.py:898] (0/4) Epoch 12, batch 2250, loss[loss=0.1587, simple_loss=0.2449, pruned_loss=0.03627, over 18367.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.2696, pruned_loss=0.04936, over 3584054.22 frames. ], batch size: 46, lr: 9.44e-03, grad_scale: 4.0 +2023-03-09 03:32:46,342 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.218e+02 3.002e+02 3.427e+02 4.398e+02 7.070e+02, threshold=6.854e+02, percent-clipped=0.0 +2023-03-09 03:32:58,288 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42262.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:33:14,289 INFO [train.py:898] (0/4) Epoch 12, batch 2300, loss[loss=0.1799, simple_loss=0.2618, pruned_loss=0.04902, over 18560.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2702, pruned_loss=0.04962, over 3582035.66 frames. ], batch size: 49, lr: 9.44e-03, grad_scale: 4.0 +2023-03-09 03:33:24,664 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8719, 5.3218, 4.9589, 5.0968, 4.9226, 4.8536, 5.3475, 5.3229], + device='cuda:0'), covar=tensor([0.1138, 0.0689, 0.0819, 0.0755, 0.1443, 0.0658, 0.0625, 0.0686], + device='cuda:0'), in_proj_covar=tensor([0.0539, 0.0444, 0.0344, 0.0476, 0.0657, 0.0482, 0.0632, 0.0470], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 03:33:33,669 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1051, 5.0192, 5.2633, 5.3064, 5.0128, 5.7929, 5.4421, 5.1325], + device='cuda:0'), covar=tensor([0.1030, 0.0670, 0.0700, 0.0626, 0.1444, 0.0717, 0.0583, 0.1669], + device='cuda:0'), in_proj_covar=tensor([0.0313, 0.0238, 0.0253, 0.0255, 0.0298, 0.0361, 0.0240, 0.0349], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0004, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 03:33:47,406 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42304.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 03:33:54,936 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=42310.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 03:34:13,011 INFO [train.py:898] (0/4) Epoch 12, batch 2350, loss[loss=0.2009, simple_loss=0.2922, pruned_loss=0.05478, over 18260.00 frames. ], tot_loss[loss=0.1847, simple_loss=0.2705, pruned_loss=0.04942, over 3585653.33 frames. ], batch size: 60, lr: 9.43e-03, grad_scale: 4.0 +2023-03-09 03:34:29,199 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-09 03:34:43,300 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.195e+02 3.349e+02 3.901e+02 4.944e+02 8.097e+02, threshold=7.803e+02, percent-clipped=6.0 +2023-03-09 03:34:50,330 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=42358.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:35:10,383 INFO [train.py:898] (0/4) Epoch 12, batch 2400, loss[loss=0.1638, simple_loss=0.2446, pruned_loss=0.04147, over 18418.00 frames. ], tot_loss[loss=0.1837, simple_loss=0.2694, pruned_loss=0.04906, over 3586308.72 frames. ], batch size: 42, lr: 9.42e-03, grad_scale: 8.0 +2023-03-09 03:35:15,933 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42379.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:36:06,529 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6911, 3.0153, 2.6381, 2.8858, 3.7484, 3.6854, 3.2494, 3.0719], + device='cuda:0'), covar=tensor([0.0213, 0.0257, 0.0551, 0.0323, 0.0154, 0.0166, 0.0297, 0.0301], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0111, 0.0154, 0.0139, 0.0108, 0.0093, 0.0137, 0.0133], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 03:36:06,895 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.41 vs. limit=5.0 +2023-03-09 03:36:08,428 INFO [train.py:898] (0/4) Epoch 12, batch 2450, loss[loss=0.2012, simple_loss=0.2857, pruned_loss=0.05839, over 18240.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.2696, pruned_loss=0.04914, over 3573240.77 frames. ], batch size: 60, lr: 9.42e-03, grad_scale: 8.0 +2023-03-09 03:36:27,161 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42440.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:36:40,614 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.166e+02 3.219e+02 3.774e+02 4.354e+02 8.177e+02, threshold=7.548e+02, percent-clipped=1.0 +2023-03-09 03:37:07,550 INFO [train.py:898] (0/4) Epoch 12, batch 2500, loss[loss=0.174, simple_loss=0.2657, pruned_loss=0.04117, over 18375.00 frames. ], tot_loss[loss=0.1851, simple_loss=0.2706, pruned_loss=0.04978, over 3562809.57 frames. ], batch size: 50, lr: 9.41e-03, grad_scale: 8.0 +2023-03-09 03:37:44,310 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42506.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:37:49,989 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5559, 3.3214, 2.2717, 4.4011, 2.8729, 4.3092, 1.9785, 3.8851], + device='cuda:0'), covar=tensor([0.0520, 0.0748, 0.1151, 0.0370, 0.0802, 0.0241, 0.1278, 0.0304], + device='cuda:0'), in_proj_covar=tensor([0.0196, 0.0214, 0.0180, 0.0244, 0.0180, 0.0244, 0.0192, 0.0185], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 03:38:06,088 INFO [train.py:898] (0/4) Epoch 12, batch 2550, loss[loss=0.1635, simple_loss=0.2506, pruned_loss=0.03818, over 18370.00 frames. ], tot_loss[loss=0.185, simple_loss=0.2705, pruned_loss=0.04978, over 3568894.95 frames. ], batch size: 46, lr: 9.41e-03, grad_scale: 8.0 +2023-03-09 03:38:34,574 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5968, 3.4499, 4.6180, 2.8987, 3.9785, 2.4988, 2.8371, 1.9045], + device='cuda:0'), covar=tensor([0.1066, 0.0812, 0.0165, 0.0697, 0.0597, 0.2338, 0.2557, 0.1814], + device='cuda:0'), in_proj_covar=tensor([0.0197, 0.0218, 0.0122, 0.0172, 0.0230, 0.0247, 0.0287, 0.0210], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 03:38:37,552 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.199e+02 3.182e+02 3.754e+02 4.508e+02 9.914e+02, threshold=7.507e+02, percent-clipped=4.0 +2023-03-09 03:38:39,726 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.68 vs. limit=5.0 +2023-03-09 03:38:49,103 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=42562.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:38:54,924 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42567.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:38:56,175 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7076, 4.2497, 4.3452, 3.2625, 3.5395, 3.3510, 2.4437, 2.2674], + device='cuda:0'), covar=tensor([0.0200, 0.0141, 0.0066, 0.0317, 0.0356, 0.0238, 0.0758, 0.0879], + device='cuda:0'), in_proj_covar=tensor([0.0061, 0.0048, 0.0049, 0.0061, 0.0082, 0.0059, 0.0072, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0006, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 03:39:04,473 INFO [train.py:898] (0/4) Epoch 12, batch 2600, loss[loss=0.206, simple_loss=0.2932, pruned_loss=0.05942, over 18248.00 frames. ], tot_loss[loss=0.1852, simple_loss=0.2708, pruned_loss=0.04981, over 3579639.51 frames. ], batch size: 60, lr: 9.40e-03, grad_scale: 8.0 +2023-03-09 03:39:28,030 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6476, 3.6932, 5.0312, 4.3216, 3.3368, 2.8609, 4.5015, 5.2610], + device='cuda:0'), covar=tensor([0.0791, 0.1533, 0.0131, 0.0360, 0.0849, 0.1196, 0.0362, 0.0212], + device='cuda:0'), in_proj_covar=tensor([0.0137, 0.0248, 0.0110, 0.0164, 0.0178, 0.0174, 0.0175, 0.0153], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 03:39:35,428 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.74 vs. limit=2.0 +2023-03-09 03:39:39,329 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=42604.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 03:39:46,073 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=42610.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:40:03,350 INFO [train.py:898] (0/4) Epoch 12, batch 2650, loss[loss=0.1733, simple_loss=0.2628, pruned_loss=0.04186, over 18410.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2704, pruned_loss=0.04968, over 3584977.23 frames. ], batch size: 52, lr: 9.40e-03, grad_scale: 8.0 +2023-03-09 03:40:08,267 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0741, 5.6178, 5.2572, 5.3308, 5.1539, 5.0960, 5.6320, 5.6457], + device='cuda:0'), covar=tensor([0.1164, 0.0743, 0.0579, 0.0782, 0.1574, 0.0793, 0.0662, 0.0730], + device='cuda:0'), in_proj_covar=tensor([0.0548, 0.0448, 0.0346, 0.0485, 0.0668, 0.0493, 0.0641, 0.0478], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 03:40:31,565 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42649.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:40:34,773 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.962e+02 3.110e+02 3.787e+02 4.423e+02 7.417e+02, threshold=7.574e+02, percent-clipped=0.0 +2023-03-09 03:40:35,007 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=42652.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 03:40:54,572 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7361, 4.4831, 4.6224, 4.4642, 4.4427, 4.6142, 4.8994, 4.7341], + device='cuda:0'), covar=tensor([0.0091, 0.0134, 0.0118, 0.0129, 0.0106, 0.0138, 0.0107, 0.0170], + device='cuda:0'), in_proj_covar=tensor([0.0082, 0.0058, 0.0061, 0.0079, 0.0065, 0.0090, 0.0076, 0.0075], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 03:41:00,848 INFO [train.py:898] (0/4) Epoch 12, batch 2700, loss[loss=0.2163, simple_loss=0.295, pruned_loss=0.06883, over 18291.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2703, pruned_loss=0.04946, over 3589755.02 frames. ], batch size: 60, lr: 9.39e-03, grad_scale: 8.0 +2023-03-09 03:41:08,227 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8642, 4.5375, 4.7260, 3.4456, 3.7217, 3.5193, 2.6278, 2.6213], + device='cuda:0'), covar=tensor([0.0192, 0.0143, 0.0054, 0.0256, 0.0286, 0.0200, 0.0647, 0.0762], + device='cuda:0'), in_proj_covar=tensor([0.0059, 0.0046, 0.0048, 0.0059, 0.0080, 0.0058, 0.0070, 0.0077], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 03:41:35,300 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0398, 5.0939, 5.0413, 4.8883, 4.7879, 4.9316, 5.2593, 5.2172], + device='cuda:0'), covar=tensor([0.0066, 0.0069, 0.0051, 0.0098, 0.0063, 0.0100, 0.0109, 0.0118], + device='cuda:0'), in_proj_covar=tensor([0.0082, 0.0058, 0.0061, 0.0078, 0.0064, 0.0090, 0.0075, 0.0074], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 03:41:42,301 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42710.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:41:59,026 INFO [train.py:898] (0/4) Epoch 12, batch 2750, loss[loss=0.201, simple_loss=0.2848, pruned_loss=0.05862, over 18348.00 frames. ], tot_loss[loss=0.1846, simple_loss=0.2702, pruned_loss=0.04945, over 3590094.72 frames. ], batch size: 56, lr: 9.39e-03, grad_scale: 8.0 +2023-03-09 03:42:11,291 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42735.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:42:13,153 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-09 03:42:32,192 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.860e+02 3.123e+02 3.590e+02 4.311e+02 1.461e+03, threshold=7.180e+02, percent-clipped=1.0 +2023-03-09 03:42:58,258 INFO [train.py:898] (0/4) Epoch 12, batch 2800, loss[loss=0.1959, simple_loss=0.2826, pruned_loss=0.05458, over 18117.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.2701, pruned_loss=0.04921, over 3597774.53 frames. ], batch size: 62, lr: 9.38e-03, grad_scale: 8.0 +2023-03-09 03:43:21,643 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8819, 4.7041, 2.8763, 4.5224, 4.4352, 4.7546, 4.5122, 2.5610], + device='cuda:0'), covar=tensor([0.0173, 0.0065, 0.0628, 0.0117, 0.0076, 0.0061, 0.0103, 0.0987], + device='cuda:0'), in_proj_covar=tensor([0.0080, 0.0069, 0.0089, 0.0085, 0.0078, 0.0067, 0.0078, 0.0093], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 03:43:51,246 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.48 vs. limit=5.0 +2023-03-09 03:43:57,219 INFO [train.py:898] (0/4) Epoch 12, batch 2850, loss[loss=0.1773, simple_loss=0.2694, pruned_loss=0.04259, over 18625.00 frames. ], tot_loss[loss=0.183, simple_loss=0.2689, pruned_loss=0.04859, over 3606606.24 frames. ], batch size: 52, lr: 9.38e-03, grad_scale: 8.0 +2023-03-09 03:44:18,332 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2361, 4.4470, 2.8306, 4.3744, 5.3907, 2.8103, 3.8601, 3.9539], + device='cuda:0'), covar=tensor([0.0133, 0.1002, 0.1500, 0.0542, 0.0066, 0.1268, 0.0727, 0.0848], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0238, 0.0193, 0.0190, 0.0094, 0.0176, 0.0205, 0.0208], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 03:44:24,930 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3303, 5.9007, 5.4188, 5.6377, 5.4373, 5.3408, 5.9355, 5.9010], + device='cuda:0'), covar=tensor([0.1261, 0.0650, 0.0471, 0.0731, 0.1600, 0.0683, 0.0560, 0.0636], + device='cuda:0'), in_proj_covar=tensor([0.0550, 0.0448, 0.0340, 0.0482, 0.0667, 0.0488, 0.0636, 0.0476], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 03:44:27,910 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.023e+02 3.014e+02 3.707e+02 4.689e+02 1.677e+03, threshold=7.413e+02, percent-clipped=4.0 +2023-03-09 03:44:35,899 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3439, 5.2278, 5.4485, 5.3655, 5.2021, 6.0715, 5.6657, 5.4164], + device='cuda:0'), covar=tensor([0.1084, 0.0581, 0.0758, 0.0668, 0.1498, 0.0729, 0.0605, 0.1532], + device='cuda:0'), in_proj_covar=tensor([0.0318, 0.0239, 0.0258, 0.0259, 0.0303, 0.0365, 0.0247, 0.0357], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0004, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 03:44:37,112 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42859.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:44:40,343 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=42862.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:44:54,953 INFO [train.py:898] (0/4) Epoch 12, batch 2900, loss[loss=0.1898, simple_loss=0.2805, pruned_loss=0.04949, over 17061.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2695, pruned_loss=0.04884, over 3606578.56 frames. ], batch size: 78, lr: 9.37e-03, grad_scale: 8.0 +2023-03-09 03:44:59,708 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=42879.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:45:33,447 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4975, 5.9482, 5.4454, 5.7689, 5.5064, 5.4358, 6.0233, 5.9578], + device='cuda:0'), covar=tensor([0.1083, 0.0696, 0.0453, 0.0729, 0.1508, 0.0719, 0.0579, 0.0701], + device='cuda:0'), in_proj_covar=tensor([0.0546, 0.0448, 0.0339, 0.0483, 0.0667, 0.0487, 0.0635, 0.0475], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 03:45:47,652 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42920.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:45:53,049 INFO [train.py:898] (0/4) Epoch 12, batch 2950, loss[loss=0.1725, simple_loss=0.2532, pruned_loss=0.04591, over 18495.00 frames. ], tot_loss[loss=0.1833, simple_loss=0.269, pruned_loss=0.04881, over 3603413.23 frames. ], batch size: 47, lr: 9.36e-03, grad_scale: 8.0 +2023-03-09 03:46:11,096 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=42940.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:46:12,301 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-09 03:46:24,424 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.040e+02 2.922e+02 3.740e+02 4.612e+02 1.225e+03, threshold=7.481e+02, percent-clipped=6.0 +2023-03-09 03:46:51,979 INFO [train.py:898] (0/4) Epoch 12, batch 3000, loss[loss=0.2041, simple_loss=0.2928, pruned_loss=0.05767, over 17795.00 frames. ], tot_loss[loss=0.1839, simple_loss=0.2693, pruned_loss=0.04919, over 3593699.59 frames. ], batch size: 70, lr: 9.36e-03, grad_scale: 8.0 +2023-03-09 03:46:51,981 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 03:47:04,093 INFO [train.py:932] (0/4) Epoch 12, validation: loss=0.1557, simple_loss=0.2578, pruned_loss=0.02677, over 944034.00 frames. +2023-03-09 03:47:04,094 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-09 03:47:40,271 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43005.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:48:03,059 INFO [train.py:898] (0/4) Epoch 12, batch 3050, loss[loss=0.1669, simple_loss=0.2493, pruned_loss=0.0422, over 18563.00 frames. ], tot_loss[loss=0.1832, simple_loss=0.2688, pruned_loss=0.04882, over 3591566.37 frames. ], batch size: 45, lr: 9.35e-03, grad_scale: 8.0 +2023-03-09 03:48:14,790 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43035.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:48:35,555 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.270e+02 3.222e+02 3.719e+02 4.518e+02 9.955e+02, threshold=7.438e+02, percent-clipped=1.0 +2023-03-09 03:48:53,791 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6521, 2.7999, 4.1727, 3.7629, 2.4213, 4.4914, 3.9758, 2.7273], + device='cuda:0'), covar=tensor([0.0445, 0.1462, 0.0238, 0.0305, 0.1769, 0.0187, 0.0388, 0.1055], + device='cuda:0'), in_proj_covar=tensor([0.0191, 0.0222, 0.0162, 0.0146, 0.0214, 0.0186, 0.0210, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 03:48:54,432 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-09 03:49:01,462 INFO [train.py:898] (0/4) Epoch 12, batch 3100, loss[loss=0.2012, simple_loss=0.2863, pruned_loss=0.05803, over 18410.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2698, pruned_loss=0.04913, over 3593218.88 frames. ], batch size: 52, lr: 9.35e-03, grad_scale: 8.0 +2023-03-09 03:49:11,131 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=43083.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:49:30,078 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9413, 4.4468, 4.6086, 3.3161, 3.5639, 3.4838, 2.7102, 2.2923], + device='cuda:0'), covar=tensor([0.0178, 0.0142, 0.0064, 0.0293, 0.0375, 0.0301, 0.0709, 0.0950], + device='cuda:0'), in_proj_covar=tensor([0.0061, 0.0048, 0.0050, 0.0061, 0.0083, 0.0060, 0.0073, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0006, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 03:50:00,222 INFO [train.py:898] (0/4) Epoch 12, batch 3150, loss[loss=0.1468, simple_loss=0.2315, pruned_loss=0.031, over 18274.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2699, pruned_loss=0.04904, over 3601023.19 frames. ], batch size: 45, lr: 9.34e-03, grad_scale: 8.0 +2023-03-09 03:50:31,050 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.182e+02 3.274e+02 3.785e+02 4.716e+02 1.243e+03, threshold=7.571e+02, percent-clipped=5.0 +2023-03-09 03:50:43,691 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43162.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:50:58,079 INFO [train.py:898] (0/4) Epoch 12, batch 3200, loss[loss=0.1919, simple_loss=0.2749, pruned_loss=0.05444, over 18385.00 frames. ], tot_loss[loss=0.1843, simple_loss=0.27, pruned_loss=0.0493, over 3602611.83 frames. ], batch size: 50, lr: 9.34e-03, grad_scale: 8.0 +2023-03-09 03:51:39,586 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=43210.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:51:45,880 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43215.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:51:57,219 INFO [train.py:898] (0/4) Epoch 12, batch 3250, loss[loss=0.1849, simple_loss=0.2775, pruned_loss=0.04609, over 18386.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2695, pruned_loss=0.04891, over 3604049.20 frames. ], batch size: 50, lr: 9.33e-03, grad_scale: 8.0 +2023-03-09 03:52:08,860 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43235.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:52:09,228 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-09 03:52:28,373 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.179e+02 3.099e+02 3.708e+02 4.167e+02 9.547e+02, threshold=7.415e+02, percent-clipped=3.0 +2023-03-09 03:52:44,945 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.83 vs. limit=5.0 +2023-03-09 03:52:55,946 INFO [train.py:898] (0/4) Epoch 12, batch 3300, loss[loss=0.2071, simple_loss=0.2914, pruned_loss=0.06135, over 18335.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2696, pruned_loss=0.04922, over 3599845.88 frames. ], batch size: 56, lr: 9.33e-03, grad_scale: 8.0 +2023-03-09 03:53:30,842 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43305.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:53:54,818 INFO [train.py:898] (0/4) Epoch 12, batch 3350, loss[loss=0.1924, simple_loss=0.2773, pruned_loss=0.05372, over 18416.00 frames. ], tot_loss[loss=0.1836, simple_loss=0.2691, pruned_loss=0.04898, over 3598982.29 frames. ], batch size: 48, lr: 9.32e-03, grad_scale: 8.0 +2023-03-09 03:54:25,612 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.041e+02 3.064e+02 3.610e+02 4.297e+02 1.025e+03, threshold=7.219e+02, percent-clipped=3.0 +2023-03-09 03:54:26,030 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0641, 5.0673, 5.1689, 4.9651, 4.8739, 4.9679, 5.3612, 5.2181], + device='cuda:0'), covar=tensor([0.0053, 0.0062, 0.0048, 0.0075, 0.0055, 0.0086, 0.0054, 0.0087], + device='cuda:0'), in_proj_covar=tensor([0.0081, 0.0057, 0.0060, 0.0077, 0.0064, 0.0088, 0.0074, 0.0074], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 03:54:26,911 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=43353.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:54:36,316 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-09 03:54:53,265 INFO [train.py:898] (0/4) Epoch 12, batch 3400, loss[loss=0.1682, simple_loss=0.2469, pruned_loss=0.0447, over 18346.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2696, pruned_loss=0.04917, over 3598950.80 frames. ], batch size: 46, lr: 9.32e-03, grad_scale: 8.0 +2023-03-09 03:55:18,872 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7416, 3.6966, 3.6141, 3.3007, 3.4611, 2.8786, 2.8409, 3.7369], + device='cuda:0'), covar=tensor([0.0043, 0.0064, 0.0058, 0.0095, 0.0067, 0.0151, 0.0173, 0.0048], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0126, 0.0108, 0.0158, 0.0111, 0.0156, 0.0161, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 03:55:51,948 INFO [train.py:898] (0/4) Epoch 12, batch 3450, loss[loss=0.2023, simple_loss=0.2939, pruned_loss=0.05532, over 18003.00 frames. ], tot_loss[loss=0.1842, simple_loss=0.27, pruned_loss=0.04923, over 3576538.09 frames. ], batch size: 65, lr: 9.31e-03, grad_scale: 8.0 +2023-03-09 03:55:54,619 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5959, 3.4399, 2.0625, 4.4596, 3.1903, 4.4743, 2.2197, 4.0312], + device='cuda:0'), covar=tensor([0.0464, 0.0741, 0.1320, 0.0372, 0.0695, 0.0277, 0.1182, 0.0296], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0212, 0.0179, 0.0245, 0.0179, 0.0241, 0.0189, 0.0184], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 03:55:59,020 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4664, 6.0297, 5.4716, 5.7966, 5.5402, 5.4989, 6.0848, 6.0456], + device='cuda:0'), covar=tensor([0.1231, 0.0668, 0.0418, 0.0647, 0.1504, 0.0672, 0.0515, 0.0630], + device='cuda:0'), in_proj_covar=tensor([0.0539, 0.0445, 0.0338, 0.0480, 0.0656, 0.0483, 0.0631, 0.0469], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 03:56:23,235 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.156e+02 3.064e+02 3.567e+02 4.230e+02 7.375e+02, threshold=7.135e+02, percent-clipped=1.0 +2023-03-09 03:56:51,194 INFO [train.py:898] (0/4) Epoch 12, batch 3500, loss[loss=0.1597, simple_loss=0.2512, pruned_loss=0.03415, over 18531.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2701, pruned_loss=0.04903, over 3576593.88 frames. ], batch size: 49, lr: 9.31e-03, grad_scale: 8.0 +2023-03-09 03:57:35,058 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=43514.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:57:36,026 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43515.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:57:46,261 INFO [train.py:898] (0/4) Epoch 12, batch 3550, loss[loss=0.2565, simple_loss=0.3207, pruned_loss=0.09615, over 12116.00 frames. ], tot_loss[loss=0.184, simple_loss=0.2699, pruned_loss=0.0491, over 3577699.54 frames. ], batch size: 130, lr: 9.30e-03, grad_scale: 8.0 +2023-03-09 03:57:57,029 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=43535.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:58:15,322 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.095e+02 3.008e+02 3.651e+02 4.331e+02 1.115e+03, threshold=7.301e+02, percent-clipped=5.0 +2023-03-09 03:58:27,333 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=43563.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:58:34,953 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9213, 2.3320, 2.7781, 3.0669, 3.5831, 5.3382, 4.9168, 4.2416], + device='cuda:0'), covar=tensor([0.1236, 0.2050, 0.2481, 0.1305, 0.1839, 0.0104, 0.0328, 0.0477], + device='cuda:0'), in_proj_covar=tensor([0.0254, 0.0310, 0.0330, 0.0252, 0.0367, 0.0199, 0.0267, 0.0214], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 03:58:40,413 INFO [train.py:898] (0/4) Epoch 12, batch 3600, loss[loss=0.2056, simple_loss=0.2932, pruned_loss=0.05897, over 16177.00 frames. ], tot_loss[loss=0.1841, simple_loss=0.2696, pruned_loss=0.04926, over 3566329.77 frames. ], batch size: 94, lr: 9.30e-03, grad_scale: 8.0 +2023-03-09 03:58:40,712 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=43575.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:58:49,319 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=43583.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 03:58:59,666 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6420, 2.9304, 4.3779, 3.8234, 2.4787, 4.5312, 4.0065, 2.9116], + device='cuda:0'), covar=tensor([0.0489, 0.1454, 0.0204, 0.0343, 0.1752, 0.0207, 0.0438, 0.1020], + device='cuda:0'), in_proj_covar=tensor([0.0193, 0.0223, 0.0161, 0.0146, 0.0215, 0.0188, 0.0210, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 03:59:17,022 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-12.pt +2023-03-09 03:59:46,565 INFO [train.py:898] (0/4) Epoch 13, batch 0, loss[loss=0.1785, simple_loss=0.2695, pruned_loss=0.04378, over 18605.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2695, pruned_loss=0.04378, over 18605.00 frames. ], batch size: 52, lr: 8.93e-03, grad_scale: 8.0 +2023-03-09 03:59:46,566 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 03:59:58,390 INFO [train.py:932] (0/4) Epoch 13, validation: loss=0.1568, simple_loss=0.2587, pruned_loss=0.02742, over 944034.00 frames. +2023-03-09 03:59:58,391 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-09 04:00:01,918 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4498, 5.2569, 5.5782, 5.4634, 5.4440, 6.1989, 5.8285, 5.6767], + device='cuda:0'), covar=tensor([0.1001, 0.0591, 0.0729, 0.0874, 0.1631, 0.0826, 0.0672, 0.1498], + device='cuda:0'), in_proj_covar=tensor([0.0315, 0.0243, 0.0258, 0.0257, 0.0302, 0.0365, 0.0245, 0.0355], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0004, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 04:00:07,036 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9347, 5.8400, 5.3068, 5.6996, 5.2116, 5.7035, 6.0145, 5.7915], + device='cuda:0'), covar=tensor([0.2480, 0.1148, 0.0815, 0.1276, 0.2444, 0.1071, 0.0984, 0.1185], + device='cuda:0'), in_proj_covar=tensor([0.0532, 0.0444, 0.0337, 0.0475, 0.0645, 0.0477, 0.0626, 0.0469], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 04:00:49,458 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.977e+02 3.410e+02 4.106e+02 5.043e+02 1.786e+03, threshold=8.212e+02, percent-clipped=7.0 +2023-03-09 04:00:57,372 INFO [train.py:898] (0/4) Epoch 13, batch 50, loss[loss=0.1756, simple_loss=0.2606, pruned_loss=0.04529, over 18421.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2707, pruned_loss=0.04955, over 805686.73 frames. ], batch size: 48, lr: 8.92e-03, grad_scale: 8.0 +2023-03-09 04:01:53,892 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-09 04:01:56,183 INFO [train.py:898] (0/4) Epoch 13, batch 100, loss[loss=0.1874, simple_loss=0.2822, pruned_loss=0.04631, over 18370.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2665, pruned_loss=0.0477, over 1427839.26 frames. ], batch size: 55, lr: 8.92e-03, grad_scale: 8.0 +2023-03-09 04:02:46,860 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.73 vs. limit=2.0 +2023-03-09 04:02:47,115 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.935e+02 2.941e+02 3.301e+02 3.834e+02 7.601e+02, threshold=6.602e+02, percent-clipped=0.0 +2023-03-09 04:02:55,217 INFO [train.py:898] (0/4) Epoch 13, batch 150, loss[loss=0.1913, simple_loss=0.2717, pruned_loss=0.0554, over 17915.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.266, pruned_loss=0.04792, over 1902432.70 frames. ], batch size: 65, lr: 8.91e-03, grad_scale: 8.0 +2023-03-09 04:03:17,298 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.00 vs. limit=2.0 +2023-03-09 04:03:28,262 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=43787.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:03:53,891 INFO [train.py:898] (0/4) Epoch 13, batch 200, loss[loss=0.208, simple_loss=0.3036, pruned_loss=0.0562, over 18491.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2675, pruned_loss=0.04837, over 2282336.66 frames. ], batch size: 59, lr: 8.91e-03, grad_scale: 8.0 +2023-03-09 04:04:40,652 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=43848.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:04:44,847 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.130e+02 3.045e+02 3.524e+02 4.456e+02 7.991e+02, threshold=7.049e+02, percent-clipped=5.0 +2023-03-09 04:04:53,610 INFO [train.py:898] (0/4) Epoch 13, batch 250, loss[loss=0.1827, simple_loss=0.2736, pruned_loss=0.04587, over 18392.00 frames. ], tot_loss[loss=0.182, simple_loss=0.2676, pruned_loss=0.04818, over 2573524.32 frames. ], batch size: 52, lr: 8.90e-03, grad_scale: 8.0 +2023-03-09 04:05:06,431 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=43870.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:05:52,782 INFO [train.py:898] (0/4) Epoch 13, batch 300, loss[loss=0.169, simple_loss=0.2554, pruned_loss=0.04131, over 18533.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2666, pruned_loss=0.04767, over 2802295.43 frames. ], batch size: 49, lr: 8.90e-03, grad_scale: 8.0 +2023-03-09 04:05:53,195 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=43909.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:06:22,399 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-09 04:06:43,697 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.829e+02 2.907e+02 3.635e+02 4.242e+02 7.161e+02, threshold=7.270e+02, percent-clipped=1.0 +2023-03-09 04:06:52,914 INFO [train.py:898] (0/4) Epoch 13, batch 350, loss[loss=0.1706, simple_loss=0.2473, pruned_loss=0.047, over 18261.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2665, pruned_loss=0.04725, over 2975249.98 frames. ], batch size: 45, lr: 8.89e-03, grad_scale: 8.0 +2023-03-09 04:06:53,314 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=43959.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:07:06,112 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=43970.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:07:24,099 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-09 04:07:42,004 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-44000.pt +2023-03-09 04:07:56,747 INFO [train.py:898] (0/4) Epoch 13, batch 400, loss[loss=0.1553, simple_loss=0.2411, pruned_loss=0.0348, over 18416.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2673, pruned_loss=0.04754, over 3113273.28 frames. ], batch size: 42, lr: 8.89e-03, grad_scale: 8.0 +2023-03-09 04:08:10,348 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=44020.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:08:42,088 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.92 vs. limit=5.0 +2023-03-09 04:08:47,397 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.657e+02 2.874e+02 3.424e+02 4.227e+02 8.593e+02, threshold=6.849e+02, percent-clipped=4.0 +2023-03-09 04:08:55,696 INFO [train.py:898] (0/4) Epoch 13, batch 450, loss[loss=0.1698, simple_loss=0.2555, pruned_loss=0.04206, over 18235.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.2663, pruned_loss=0.04704, over 3225355.83 frames. ], batch size: 45, lr: 8.88e-03, grad_scale: 16.0 +2023-03-09 04:09:54,527 INFO [train.py:898] (0/4) Epoch 13, batch 500, loss[loss=0.1836, simple_loss=0.2711, pruned_loss=0.0481, over 18497.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2671, pruned_loss=0.04729, over 3304821.93 frames. ], batch size: 51, lr: 8.88e-03, grad_scale: 16.0 +2023-03-09 04:10:34,490 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44143.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:10:44,862 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.103e+02 3.159e+02 3.633e+02 4.577e+02 9.760e+02, threshold=7.265e+02, percent-clipped=1.0 +2023-03-09 04:10:53,407 INFO [train.py:898] (0/4) Epoch 13, batch 550, loss[loss=0.1921, simple_loss=0.2724, pruned_loss=0.05586, over 18414.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2675, pruned_loss=0.04754, over 3364120.09 frames. ], batch size: 48, lr: 8.87e-03, grad_scale: 16.0 +2023-03-09 04:11:07,405 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44170.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:11:48,519 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4456, 5.2079, 5.6388, 5.5275, 5.3142, 6.1822, 5.9081, 5.5790], + device='cuda:0'), covar=tensor([0.0962, 0.0640, 0.0664, 0.0698, 0.1506, 0.0733, 0.0547, 0.1420], + device='cuda:0'), in_proj_covar=tensor([0.0316, 0.0244, 0.0258, 0.0258, 0.0300, 0.0362, 0.0240, 0.0356], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0004, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 04:11:53,539 INFO [train.py:898] (0/4) Epoch 13, batch 600, loss[loss=0.1914, simple_loss=0.2784, pruned_loss=0.05217, over 18299.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.2681, pruned_loss=0.04779, over 3411301.78 frames. ], batch size: 54, lr: 8.87e-03, grad_scale: 16.0 +2023-03-09 04:12:04,387 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=44218.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:12:42,718 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.069e+02 3.135e+02 3.711e+02 4.524e+02 8.017e+02, threshold=7.422e+02, percent-clipped=1.0 +2023-03-09 04:12:51,315 INFO [train.py:898] (0/4) Epoch 13, batch 650, loss[loss=0.2044, simple_loss=0.2913, pruned_loss=0.05879, over 18356.00 frames. ], tot_loss[loss=0.1819, simple_loss=0.2685, pruned_loss=0.04764, over 3457640.34 frames. ], batch size: 56, lr: 8.86e-03, grad_scale: 16.0 +2023-03-09 04:12:58,885 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44265.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:13:49,386 INFO [train.py:898] (0/4) Epoch 13, batch 700, loss[loss=0.2044, simple_loss=0.2913, pruned_loss=0.05878, over 18566.00 frames. ], tot_loss[loss=0.1823, simple_loss=0.2689, pruned_loss=0.04781, over 3484829.26 frames. ], batch size: 54, lr: 8.86e-03, grad_scale: 8.0 +2023-03-09 04:13:57,531 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44315.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:14:33,786 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.74 vs. limit=2.0 +2023-03-09 04:14:41,969 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.372e+02 3.099e+02 3.698e+02 4.790e+02 1.213e+03, threshold=7.395e+02, percent-clipped=5.0 +2023-03-09 04:14:48,749 INFO [train.py:898] (0/4) Epoch 13, batch 750, loss[loss=0.1543, simple_loss=0.2408, pruned_loss=0.03389, over 18281.00 frames. ], tot_loss[loss=0.1818, simple_loss=0.268, pruned_loss=0.04779, over 3501459.44 frames. ], batch size: 47, lr: 8.85e-03, grad_scale: 8.0 +2023-03-09 04:15:08,995 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.37 vs. limit=5.0 +2023-03-09 04:15:48,578 INFO [train.py:898] (0/4) Epoch 13, batch 800, loss[loss=0.1882, simple_loss=0.2726, pruned_loss=0.05187, over 18362.00 frames. ], tot_loss[loss=0.1821, simple_loss=0.2682, pruned_loss=0.04797, over 3530229.40 frames. ], batch size: 56, lr: 8.85e-03, grad_scale: 8.0 +2023-03-09 04:15:57,388 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.55 vs. limit=2.0 +2023-03-09 04:16:29,533 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44443.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:16:40,517 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.201e+02 3.366e+02 4.098e+02 5.276e+02 1.273e+03, threshold=8.196e+02, percent-clipped=10.0 +2023-03-09 04:16:44,469 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6771, 2.7482, 4.1603, 3.8826, 2.7269, 4.6552, 3.9898, 2.7735], + device='cuda:0'), covar=tensor([0.0433, 0.1406, 0.0258, 0.0292, 0.1437, 0.0168, 0.0437, 0.0982], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0223, 0.0165, 0.0146, 0.0213, 0.0188, 0.0213, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 04:16:47,524 INFO [train.py:898] (0/4) Epoch 13, batch 850, loss[loss=0.1912, simple_loss=0.2774, pruned_loss=0.05251, over 17822.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.268, pruned_loss=0.04818, over 3540690.34 frames. ], batch size: 70, lr: 8.84e-03, grad_scale: 8.0 +2023-03-09 04:17:14,109 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-09 04:17:22,305 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.4799, 4.5326, 4.5986, 4.3144, 4.3890, 4.3679, 4.6903, 4.6693], + device='cuda:0'), covar=tensor([0.0079, 0.0074, 0.0068, 0.0100, 0.0069, 0.0124, 0.0080, 0.0102], + device='cuda:0'), in_proj_covar=tensor([0.0081, 0.0058, 0.0061, 0.0077, 0.0065, 0.0088, 0.0074, 0.0075], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 04:17:25,092 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4971, 5.9754, 5.5685, 5.8022, 5.5715, 5.5053, 6.0818, 6.0561], + device='cuda:0'), covar=tensor([0.1074, 0.0672, 0.0407, 0.0641, 0.1334, 0.0723, 0.0484, 0.0617], + device='cuda:0'), in_proj_covar=tensor([0.0545, 0.0454, 0.0345, 0.0486, 0.0670, 0.0491, 0.0636, 0.0482], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 04:17:26,099 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=44491.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:17:45,919 INFO [train.py:898] (0/4) Epoch 13, batch 900, loss[loss=0.1682, simple_loss=0.256, pruned_loss=0.0402, over 18488.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2667, pruned_loss=0.04754, over 3555941.84 frames. ], batch size: 47, lr: 8.84e-03, grad_scale: 8.0 +2023-03-09 04:18:06,171 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3613, 5.3679, 4.9563, 5.2765, 5.2624, 4.6964, 5.1961, 4.9797], + device='cuda:0'), covar=tensor([0.0398, 0.0378, 0.1304, 0.0728, 0.0554, 0.0394, 0.0379, 0.0878], + device='cuda:0'), in_proj_covar=tensor([0.0411, 0.0479, 0.0628, 0.0374, 0.0368, 0.0432, 0.0463, 0.0601], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 04:18:07,290 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=44527.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:18:34,341 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=44550.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:18:37,465 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.936e+02 2.996e+02 3.521e+02 4.536e+02 7.966e+02, threshold=7.042e+02, percent-clipped=0.0 +2023-03-09 04:18:43,961 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.72 vs. limit=2.0 +2023-03-09 04:18:44,370 INFO [train.py:898] (0/4) Epoch 13, batch 950, loss[loss=0.1855, simple_loss=0.2775, pruned_loss=0.04677, over 18454.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2669, pruned_loss=0.0473, over 3560772.63 frames. ], batch size: 53, lr: 8.84e-03, grad_scale: 8.0 +2023-03-09 04:18:51,511 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44565.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:19:18,485 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=44588.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:19:27,575 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7130, 5.3071, 5.2830, 5.2750, 4.7953, 5.1389, 4.5156, 5.1823], + device='cuda:0'), covar=tensor([0.0236, 0.0289, 0.0197, 0.0365, 0.0391, 0.0239, 0.1215, 0.0269], + device='cuda:0'), in_proj_covar=tensor([0.0186, 0.0233, 0.0222, 0.0265, 0.0234, 0.0233, 0.0292, 0.0225], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0005, 0.0007, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-09 04:19:43,367 INFO [train.py:898] (0/4) Epoch 13, batch 1000, loss[loss=0.2176, simple_loss=0.2898, pruned_loss=0.07269, over 12561.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.267, pruned_loss=0.04754, over 3547784.23 frames. ], batch size: 130, lr: 8.83e-03, grad_scale: 8.0 +2023-03-09 04:19:46,102 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=44611.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:19:48,240 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=44613.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:19:50,571 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=44615.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:20:20,147 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7717, 4.1281, 4.1514, 4.1560, 3.8269, 4.0734, 3.6919, 4.1059], + device='cuda:0'), covar=tensor([0.0291, 0.0329, 0.0266, 0.0488, 0.0361, 0.0268, 0.0988, 0.0329], + device='cuda:0'), in_proj_covar=tensor([0.0187, 0.0234, 0.0223, 0.0267, 0.0236, 0.0233, 0.0294, 0.0226], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0005, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 04:20:36,781 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.889e+02 3.069e+02 3.543e+02 4.304e+02 1.212e+03, threshold=7.086e+02, percent-clipped=7.0 +2023-03-09 04:20:42,642 INFO [train.py:898] (0/4) Epoch 13, batch 1050, loss[loss=0.2048, simple_loss=0.2849, pruned_loss=0.0623, over 16013.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2674, pruned_loss=0.0477, over 3548743.66 frames. ], batch size: 94, lr: 8.83e-03, grad_scale: 4.0 +2023-03-09 04:20:47,322 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=44663.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:21:41,976 INFO [train.py:898] (0/4) Epoch 13, batch 1100, loss[loss=0.1634, simple_loss=0.239, pruned_loss=0.04393, over 17641.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2674, pruned_loss=0.04767, over 3558659.39 frames. ], batch size: 39, lr: 8.82e-03, grad_scale: 4.0 +2023-03-09 04:22:20,644 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-03-09 04:22:28,381 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2623, 5.2643, 4.8427, 5.1907, 5.1816, 4.5813, 5.0992, 4.8589], + device='cuda:0'), covar=tensor([0.0417, 0.0387, 0.1414, 0.0712, 0.0500, 0.0400, 0.0406, 0.0975], + device='cuda:0'), in_proj_covar=tensor([0.0413, 0.0485, 0.0640, 0.0380, 0.0373, 0.0437, 0.0466, 0.0607], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 04:22:35,539 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.712e+02 2.982e+02 3.567e+02 4.035e+02 7.842e+02, threshold=7.134e+02, percent-clipped=1.0 +2023-03-09 04:22:41,054 INFO [train.py:898] (0/4) Epoch 13, batch 1150, loss[loss=0.1698, simple_loss=0.2516, pruned_loss=0.04398, over 18166.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.267, pruned_loss=0.04759, over 3561721.58 frames. ], batch size: 44, lr: 8.82e-03, grad_scale: 4.0 +2023-03-09 04:23:40,892 INFO [train.py:898] (0/4) Epoch 13, batch 1200, loss[loss=0.168, simple_loss=0.2589, pruned_loss=0.03857, over 18389.00 frames. ], tot_loss[loss=0.1817, simple_loss=0.2676, pruned_loss=0.04785, over 3565785.22 frames. ], batch size: 52, lr: 8.81e-03, grad_scale: 8.0 +2023-03-09 04:24:33,794 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.745e+02 2.967e+02 3.663e+02 4.375e+02 1.162e+03, threshold=7.327e+02, percent-clipped=1.0 +2023-03-09 04:24:40,015 INFO [train.py:898] (0/4) Epoch 13, batch 1250, loss[loss=0.1557, simple_loss=0.2371, pruned_loss=0.03713, over 18439.00 frames. ], tot_loss[loss=0.1802, simple_loss=0.266, pruned_loss=0.04716, over 3577616.90 frames. ], batch size: 43, lr: 8.81e-03, grad_scale: 8.0 +2023-03-09 04:25:07,167 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44883.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:25:35,325 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=44906.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:25:38,476 INFO [train.py:898] (0/4) Epoch 13, batch 1300, loss[loss=0.1851, simple_loss=0.2719, pruned_loss=0.04909, over 18575.00 frames. ], tot_loss[loss=0.1813, simple_loss=0.2667, pruned_loss=0.0479, over 3561244.17 frames. ], batch size: 54, lr: 8.80e-03, grad_scale: 8.0 +2023-03-09 04:25:41,491 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8414, 4.9039, 4.9073, 4.6828, 4.6777, 4.6684, 5.0512, 5.0539], + device='cuda:0'), covar=tensor([0.0073, 0.0076, 0.0067, 0.0100, 0.0066, 0.0130, 0.0085, 0.0099], + device='cuda:0'), in_proj_covar=tensor([0.0081, 0.0057, 0.0060, 0.0076, 0.0065, 0.0088, 0.0074, 0.0074], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 04:26:31,152 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.969e+02 2.987e+02 3.642e+02 4.868e+02 9.952e+02, threshold=7.283e+02, percent-clipped=3.0 +2023-03-09 04:26:36,919 INFO [train.py:898] (0/4) Epoch 13, batch 1350, loss[loss=0.1847, simple_loss=0.2732, pruned_loss=0.04817, over 18382.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2668, pruned_loss=0.04775, over 3573640.78 frames. ], batch size: 50, lr: 8.80e-03, grad_scale: 8.0 +2023-03-09 04:27:11,381 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.80 vs. limit=2.0 +2023-03-09 04:27:36,020 INFO [train.py:898] (0/4) Epoch 13, batch 1400, loss[loss=0.2399, simple_loss=0.309, pruned_loss=0.08545, over 12475.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.267, pruned_loss=0.0481, over 3563747.84 frames. ], batch size: 130, lr: 8.79e-03, grad_scale: 8.0 +2023-03-09 04:28:28,831 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.199e+02 3.129e+02 3.917e+02 4.727e+02 1.364e+03, threshold=7.834e+02, percent-clipped=4.0 +2023-03-09 04:28:32,214 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45056.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:28:35,367 INFO [train.py:898] (0/4) Epoch 13, batch 1450, loss[loss=0.2005, simple_loss=0.2865, pruned_loss=0.05726, over 18369.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2669, pruned_loss=0.04812, over 3555855.46 frames. ], batch size: 56, lr: 8.79e-03, grad_scale: 8.0 +2023-03-09 04:29:01,582 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45081.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:29:34,107 INFO [train.py:898] (0/4) Epoch 13, batch 1500, loss[loss=0.1763, simple_loss=0.2724, pruned_loss=0.04013, over 18478.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2663, pruned_loss=0.04766, over 3574757.27 frames. ], batch size: 51, lr: 8.78e-03, grad_scale: 8.0 +2023-03-09 04:29:44,346 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45117.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 04:29:50,376 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9346, 3.1852, 4.3271, 3.9951, 3.0581, 4.7666, 4.0873, 3.1568], + device='cuda:0'), covar=tensor([0.0411, 0.1287, 0.0226, 0.0309, 0.1322, 0.0163, 0.0418, 0.0861], + device='cuda:0'), in_proj_covar=tensor([0.0196, 0.0226, 0.0166, 0.0146, 0.0216, 0.0194, 0.0216, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 04:30:13,274 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45142.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:30:26,572 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.379e+02 3.129e+02 3.724e+02 4.200e+02 9.663e+02, threshold=7.448e+02, percent-clipped=2.0 +2023-03-09 04:30:33,406 INFO [train.py:898] (0/4) Epoch 13, batch 1550, loss[loss=0.1801, simple_loss=0.2674, pruned_loss=0.0464, over 16136.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2666, pruned_loss=0.0475, over 3584950.79 frames. ], batch size: 94, lr: 8.78e-03, grad_scale: 8.0 +2023-03-09 04:31:01,952 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45183.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:31:28,191 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45206.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:31:32,575 INFO [train.py:898] (0/4) Epoch 13, batch 1600, loss[loss=0.1923, simple_loss=0.2862, pruned_loss=0.04915, over 18563.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2662, pruned_loss=0.04718, over 3589073.94 frames. ], batch size: 54, lr: 8.77e-03, grad_scale: 8.0 +2023-03-09 04:31:51,295 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0821, 5.2376, 5.3868, 5.3176, 5.1851, 5.9347, 5.5552, 5.3876], + device='cuda:0'), covar=tensor([0.1166, 0.0693, 0.0712, 0.0722, 0.1302, 0.0783, 0.0657, 0.1503], + device='cuda:0'), in_proj_covar=tensor([0.0322, 0.0252, 0.0266, 0.0265, 0.0306, 0.0379, 0.0248, 0.0365], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0004, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 04:31:58,493 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=45231.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:32:24,702 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.718e+02 3.180e+02 3.814e+02 4.660e+02 1.002e+03, threshold=7.628e+02, percent-clipped=5.0 +2023-03-09 04:32:24,867 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=45254.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:32:30,353 INFO [train.py:898] (0/4) Epoch 13, batch 1650, loss[loss=0.1981, simple_loss=0.2883, pruned_loss=0.05397, over 18143.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2669, pruned_loss=0.04716, over 3593666.82 frames. ], batch size: 62, lr: 8.77e-03, grad_scale: 8.0 +2023-03-09 04:33:09,012 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6092, 3.7635, 5.1512, 4.5434, 3.3185, 2.8010, 4.4072, 5.3433], + device='cuda:0'), covar=tensor([0.0831, 0.1552, 0.0125, 0.0307, 0.0934, 0.1190, 0.0381, 0.0220], + device='cuda:0'), in_proj_covar=tensor([0.0138, 0.0251, 0.0113, 0.0165, 0.0182, 0.0177, 0.0178, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 04:33:29,092 INFO [train.py:898] (0/4) Epoch 13, batch 1700, loss[loss=0.1514, simple_loss=0.2363, pruned_loss=0.03327, over 18260.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2655, pruned_loss=0.04664, over 3598836.29 frames. ], batch size: 45, lr: 8.76e-03, grad_scale: 8.0 +2023-03-09 04:34:15,793 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45348.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 04:34:22,328 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.903e+02 2.834e+02 3.670e+02 4.509e+02 1.027e+03, threshold=7.340e+02, percent-clipped=3.0 +2023-03-09 04:34:28,087 INFO [train.py:898] (0/4) Epoch 13, batch 1750, loss[loss=0.156, simple_loss=0.2389, pruned_loss=0.0366, over 18343.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2668, pruned_loss=0.04722, over 3591483.49 frames. ], batch size: 46, lr: 8.76e-03, grad_scale: 8.0 +2023-03-09 04:34:33,223 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.58 vs. limit=2.0 +2023-03-09 04:35:11,053 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-09 04:35:27,717 INFO [train.py:898] (0/4) Epoch 13, batch 1800, loss[loss=0.1549, simple_loss=0.2318, pruned_loss=0.039, over 18400.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2658, pruned_loss=0.04702, over 3590459.00 frames. ], batch size: 42, lr: 8.75e-03, grad_scale: 8.0 +2023-03-09 04:35:28,219 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45409.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 04:35:31,462 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=45412.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 04:36:01,089 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=45437.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:36:21,057 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.197e+02 3.089e+02 3.615e+02 4.391e+02 8.521e+02, threshold=7.230e+02, percent-clipped=5.0 +2023-03-09 04:36:26,729 INFO [train.py:898] (0/4) Epoch 13, batch 1850, loss[loss=0.1595, simple_loss=0.2381, pruned_loss=0.04047, over 18424.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2669, pruned_loss=0.04758, over 3575915.51 frames. ], batch size: 42, lr: 8.75e-03, grad_scale: 8.0 +2023-03-09 04:37:18,236 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7453, 2.2196, 2.7570, 2.8655, 3.3859, 5.1372, 4.8314, 4.0134], + device='cuda:0'), covar=tensor([0.1318, 0.2196, 0.2322, 0.1361, 0.1878, 0.0124, 0.0315, 0.0549], + device='cuda:0'), in_proj_covar=tensor([0.0257, 0.0312, 0.0334, 0.0255, 0.0368, 0.0201, 0.0269, 0.0217], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 04:37:25,259 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.58 vs. limit=2.0 +2023-03-09 04:37:25,625 INFO [train.py:898] (0/4) Epoch 13, batch 1900, loss[loss=0.1852, simple_loss=0.2691, pruned_loss=0.0507, over 18482.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2665, pruned_loss=0.04745, over 3580106.85 frames. ], batch size: 51, lr: 8.74e-03, grad_scale: 8.0 +2023-03-09 04:37:50,467 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1377, 5.0946, 5.3389, 5.3315, 5.0827, 5.9742, 5.5828, 5.3239], + device='cuda:0'), covar=tensor([0.1140, 0.0743, 0.0747, 0.0751, 0.1561, 0.0710, 0.0561, 0.1578], + device='cuda:0'), in_proj_covar=tensor([0.0326, 0.0257, 0.0270, 0.0270, 0.0312, 0.0377, 0.0251, 0.0370], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0004, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 04:38:18,704 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.016e+02 2.946e+02 3.593e+02 4.543e+02 1.018e+03, threshold=7.187e+02, percent-clipped=4.0 +2023-03-09 04:38:24,488 INFO [train.py:898] (0/4) Epoch 13, batch 1950, loss[loss=0.1637, simple_loss=0.2515, pruned_loss=0.03797, over 18152.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2673, pruned_loss=0.0475, over 3579084.50 frames. ], batch size: 44, lr: 8.74e-03, grad_scale: 8.0 +2023-03-09 04:38:59,304 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2181, 5.5997, 2.8957, 5.3633, 5.2815, 5.6153, 5.4119, 2.8952], + device='cuda:0'), covar=tensor([0.0165, 0.0057, 0.0714, 0.0069, 0.0063, 0.0055, 0.0071, 0.0870], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0070, 0.0089, 0.0086, 0.0079, 0.0067, 0.0078, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 04:39:18,537 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2821, 4.3433, 2.5923, 4.4063, 5.4091, 2.4711, 3.9826, 4.1852], + device='cuda:0'), covar=tensor([0.0088, 0.1104, 0.1511, 0.0496, 0.0040, 0.1243, 0.0591, 0.0649], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0241, 0.0195, 0.0191, 0.0095, 0.0175, 0.0205, 0.0208], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 04:39:23,850 INFO [train.py:898] (0/4) Epoch 13, batch 2000, loss[loss=0.1632, simple_loss=0.2398, pruned_loss=0.04335, over 18494.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2671, pruned_loss=0.04739, over 3579178.24 frames. ], batch size: 44, lr: 8.73e-03, grad_scale: 8.0 +2023-03-09 04:39:25,300 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6747, 4.5294, 4.6757, 4.4362, 4.4463, 4.4952, 4.7956, 4.7864], + device='cuda:0'), covar=tensor([0.0073, 0.0084, 0.0085, 0.0109, 0.0076, 0.0124, 0.0085, 0.0096], + device='cuda:0'), in_proj_covar=tensor([0.0084, 0.0059, 0.0062, 0.0079, 0.0065, 0.0091, 0.0076, 0.0076], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 04:40:03,486 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2747, 4.2942, 2.7224, 4.5075, 5.3432, 2.6966, 4.0138, 4.1706], + device='cuda:0'), covar=tensor([0.0086, 0.1200, 0.1402, 0.0467, 0.0050, 0.1136, 0.0562, 0.0661], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0242, 0.0196, 0.0191, 0.0095, 0.0176, 0.0205, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 04:40:17,736 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.026e+02 2.791e+02 3.256e+02 3.955e+02 1.111e+03, threshold=6.512e+02, percent-clipped=4.0 +2023-03-09 04:40:23,301 INFO [train.py:898] (0/4) Epoch 13, batch 2050, loss[loss=0.1912, simple_loss=0.2746, pruned_loss=0.05393, over 18566.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2673, pruned_loss=0.04739, over 3586970.99 frames. ], batch size: 54, lr: 8.73e-03, grad_scale: 8.0 +2023-03-09 04:40:58,624 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-09 04:41:10,592 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3186, 5.1703, 5.4367, 5.4830, 5.2181, 6.0752, 5.7733, 5.4074], + device='cuda:0'), covar=tensor([0.1006, 0.0637, 0.0721, 0.0744, 0.1592, 0.0771, 0.0562, 0.1606], + device='cuda:0'), in_proj_covar=tensor([0.0317, 0.0251, 0.0264, 0.0265, 0.0305, 0.0371, 0.0245, 0.0363], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0002, 0.0002, 0.0003, 0.0004, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 04:41:14,551 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45702.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:41:17,225 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=45704.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 04:41:22,660 INFO [train.py:898] (0/4) Epoch 13, batch 2100, loss[loss=0.1586, simple_loss=0.2352, pruned_loss=0.04095, over 18265.00 frames. ], tot_loss[loss=0.181, simple_loss=0.267, pruned_loss=0.04747, over 3589967.15 frames. ], batch size: 45, lr: 8.72e-03, grad_scale: 8.0 +2023-03-09 04:41:26,323 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45712.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 04:41:55,638 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=45737.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:42:14,745 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.079e+02 3.037e+02 3.571e+02 4.287e+02 1.143e+03, threshold=7.142e+02, percent-clipped=3.0 +2023-03-09 04:42:21,832 INFO [train.py:898] (0/4) Epoch 13, batch 2150, loss[loss=0.2067, simple_loss=0.3013, pruned_loss=0.05603, over 17926.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2676, pruned_loss=0.04777, over 3588035.35 frames. ], batch size: 65, lr: 8.72e-03, grad_scale: 8.0 +2023-03-09 04:42:23,153 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=45760.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:42:26,693 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45763.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:42:38,904 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45774.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:42:51,019 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=45785.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:43:20,423 INFO [train.py:898] (0/4) Epoch 13, batch 2200, loss[loss=0.201, simple_loss=0.2831, pruned_loss=0.05949, over 18352.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2672, pruned_loss=0.04753, over 3597206.56 frames. ], batch size: 56, lr: 8.72e-03, grad_scale: 8.0 +2023-03-09 04:43:50,832 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45835.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:44:12,838 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.979e+02 3.106e+02 3.775e+02 4.323e+02 7.775e+02, threshold=7.551e+02, percent-clipped=4.0 +2023-03-09 04:44:18,536 INFO [train.py:898] (0/4) Epoch 13, batch 2250, loss[loss=0.1697, simple_loss=0.2494, pruned_loss=0.04496, over 18394.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2671, pruned_loss=0.04751, over 3595932.83 frames. ], batch size: 42, lr: 8.71e-03, grad_scale: 8.0 +2023-03-09 04:44:18,886 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=45859.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:45:16,357 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8683, 3.4425, 2.5640, 3.3162, 4.0212, 2.3759, 3.2356, 3.4613], + device='cuda:0'), covar=tensor([0.0199, 0.1004, 0.1367, 0.0672, 0.0087, 0.1412, 0.0727, 0.0622], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0240, 0.0194, 0.0190, 0.0095, 0.0175, 0.0203, 0.0208], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 04:45:17,119 INFO [train.py:898] (0/4) Epoch 13, batch 2300, loss[loss=0.2333, simple_loss=0.306, pruned_loss=0.08034, over 12517.00 frames. ], tot_loss[loss=0.1815, simple_loss=0.2674, pruned_loss=0.04783, over 3581667.84 frames. ], batch size: 129, lr: 8.71e-03, grad_scale: 8.0 +2023-03-09 04:45:31,361 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=45920.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:45:55,274 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-03-09 04:46:10,649 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.145e+02 3.100e+02 3.461e+02 4.036e+02 9.492e+02, threshold=6.921e+02, percent-clipped=1.0 +2023-03-09 04:46:16,251 INFO [train.py:898] (0/4) Epoch 13, batch 2350, loss[loss=0.1828, simple_loss=0.2734, pruned_loss=0.04607, over 18357.00 frames. ], tot_loss[loss=0.1812, simple_loss=0.2671, pruned_loss=0.04764, over 3585252.23 frames. ], batch size: 55, lr: 8.70e-03, grad_scale: 8.0 +2023-03-09 04:47:05,321 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-46000.pt +2023-03-09 04:47:15,002 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46004.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 04:47:20,470 INFO [train.py:898] (0/4) Epoch 13, batch 2400, loss[loss=0.162, simple_loss=0.2457, pruned_loss=0.03911, over 18354.00 frames. ], tot_loss[loss=0.1804, simple_loss=0.2664, pruned_loss=0.04718, over 3587174.66 frames. ], batch size: 46, lr: 8.70e-03, grad_scale: 8.0 +2023-03-09 04:47:34,358 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5217, 3.5522, 3.3587, 3.0655, 3.3822, 2.6858, 2.6680, 3.6089], + device='cuda:0'), covar=tensor([0.0046, 0.0070, 0.0072, 0.0112, 0.0062, 0.0159, 0.0167, 0.0048], + device='cuda:0'), in_proj_covar=tensor([0.0107, 0.0129, 0.0111, 0.0161, 0.0112, 0.0158, 0.0161, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 04:47:56,632 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8889, 4.8978, 5.0765, 4.7095, 4.7853, 4.7501, 5.0829, 5.1279], + device='cuda:0'), covar=tensor([0.0075, 0.0076, 0.0065, 0.0092, 0.0064, 0.0136, 0.0085, 0.0084], + device='cuda:0'), in_proj_covar=tensor([0.0083, 0.0059, 0.0061, 0.0078, 0.0065, 0.0090, 0.0076, 0.0075], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 04:48:10,971 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=46052.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 04:48:14,681 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.917e+02 3.063e+02 3.864e+02 4.591e+02 1.180e+03, threshold=7.727e+02, percent-clipped=5.0 +2023-03-09 04:48:18,311 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46058.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:48:19,203 INFO [train.py:898] (0/4) Epoch 13, batch 2450, loss[loss=0.1909, simple_loss=0.2706, pruned_loss=0.05558, over 18307.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2664, pruned_loss=0.0475, over 3584263.61 frames. ], batch size: 49, lr: 8.69e-03, grad_scale: 8.0 +2023-03-09 04:49:01,252 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8076, 4.8199, 4.9464, 4.6619, 4.6311, 4.7233, 5.0324, 4.9972], + device='cuda:0'), covar=tensor([0.0073, 0.0073, 0.0058, 0.0090, 0.0070, 0.0117, 0.0066, 0.0112], + device='cuda:0'), in_proj_covar=tensor([0.0083, 0.0059, 0.0061, 0.0079, 0.0066, 0.0090, 0.0076, 0.0076], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 04:49:18,289 INFO [train.py:898] (0/4) Epoch 13, batch 2500, loss[loss=0.1853, simple_loss=0.2777, pruned_loss=0.04641, over 18619.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.267, pruned_loss=0.04738, over 3595535.79 frames. ], batch size: 52, lr: 8.69e-03, grad_scale: 8.0 +2023-03-09 04:49:42,830 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46130.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:50:11,946 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.884e+02 3.026e+02 3.575e+02 4.443e+02 8.459e+02, threshold=7.149e+02, percent-clipped=1.0 +2023-03-09 04:50:16,987 INFO [train.py:898] (0/4) Epoch 13, batch 2550, loss[loss=0.1692, simple_loss=0.2529, pruned_loss=0.04277, over 18375.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2675, pruned_loss=0.04765, over 3586790.87 frames. ], batch size: 46, lr: 8.68e-03, grad_scale: 8.0 +2023-03-09 04:50:24,931 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-03-09 04:51:15,928 INFO [train.py:898] (0/4) Epoch 13, batch 2600, loss[loss=0.1684, simple_loss=0.2587, pruned_loss=0.03911, over 18526.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2686, pruned_loss=0.04787, over 3593747.67 frames. ], batch size: 47, lr: 8.68e-03, grad_scale: 8.0 +2023-03-09 04:51:23,621 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46215.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:51:34,015 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.3400, 2.0060, 1.9720, 2.0828, 2.3835, 2.4702, 2.3150, 2.0449], + device='cuda:0'), covar=tensor([0.0194, 0.0189, 0.0394, 0.0386, 0.0155, 0.0148, 0.0305, 0.0269], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0115, 0.0153, 0.0145, 0.0107, 0.0096, 0.0138, 0.0139], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 04:52:10,077 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.796e+02 3.019e+02 3.586e+02 4.587e+02 8.864e+02, threshold=7.172e+02, percent-clipped=4.0 +2023-03-09 04:52:14,633 INFO [train.py:898] (0/4) Epoch 13, batch 2650, loss[loss=0.1817, simple_loss=0.2725, pruned_loss=0.0455, over 18394.00 frames. ], tot_loss[loss=0.1814, simple_loss=0.2678, pruned_loss=0.0475, over 3591787.87 frames. ], batch size: 52, lr: 8.67e-03, grad_scale: 8.0 +2023-03-09 04:52:27,409 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6858, 3.0078, 4.3488, 3.7192, 2.7724, 4.5879, 4.0096, 3.0077], + device='cuda:0'), covar=tensor([0.0435, 0.1258, 0.0195, 0.0372, 0.1388, 0.0197, 0.0363, 0.0908], + device='cuda:0'), in_proj_covar=tensor([0.0197, 0.0227, 0.0166, 0.0149, 0.0214, 0.0194, 0.0216, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 04:52:32,013 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5489, 3.5062, 2.0436, 4.3926, 2.9336, 4.3074, 2.2208, 4.0382], + device='cuda:0'), covar=tensor([0.0581, 0.0770, 0.1466, 0.0454, 0.0861, 0.0284, 0.1240, 0.0313], + device='cuda:0'), in_proj_covar=tensor([0.0202, 0.0218, 0.0183, 0.0253, 0.0182, 0.0253, 0.0195, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 04:52:35,521 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46276.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:52:47,212 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2024, 2.5135, 2.3049, 2.6573, 3.3686, 3.3463, 2.8541, 2.6615], + device='cuda:0'), covar=tensor([0.0223, 0.0311, 0.0582, 0.0367, 0.0185, 0.0171, 0.0383, 0.0390], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0116, 0.0154, 0.0146, 0.0108, 0.0098, 0.0139, 0.0140], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 04:53:14,122 INFO [train.py:898] (0/4) Epoch 13, batch 2700, loss[loss=0.2145, simple_loss=0.2971, pruned_loss=0.06595, over 18218.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2676, pruned_loss=0.04709, over 3582499.31 frames. ], batch size: 60, lr: 8.67e-03, grad_scale: 8.0 +2023-03-09 04:53:36,241 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8408, 2.9597, 4.4175, 3.9143, 2.6595, 4.7076, 3.9365, 3.1040], + device='cuda:0'), covar=tensor([0.0445, 0.1468, 0.0201, 0.0335, 0.1686, 0.0190, 0.0483, 0.0976], + device='cuda:0'), in_proj_covar=tensor([0.0198, 0.0227, 0.0166, 0.0149, 0.0215, 0.0195, 0.0217, 0.0192], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 04:53:47,824 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46337.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:54:08,568 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.122e+02 2.828e+02 3.269e+02 4.141e+02 7.295e+02, threshold=6.537e+02, percent-clipped=1.0 +2023-03-09 04:54:10,129 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4021, 5.3327, 4.9502, 5.2623, 5.2922, 4.6014, 5.2042, 4.9369], + device='cuda:0'), covar=tensor([0.0411, 0.0416, 0.1487, 0.0790, 0.0510, 0.0478, 0.0459, 0.1063], + device='cuda:0'), in_proj_covar=tensor([0.0419, 0.0488, 0.0641, 0.0385, 0.0373, 0.0443, 0.0469, 0.0606], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 04:54:12,335 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46358.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:54:13,366 INFO [train.py:898] (0/4) Epoch 13, batch 2750, loss[loss=0.1667, simple_loss=0.2614, pruned_loss=0.03595, over 18496.00 frames. ], tot_loss[loss=0.1816, simple_loss=0.2683, pruned_loss=0.04743, over 3587425.72 frames. ], batch size: 51, lr: 8.66e-03, grad_scale: 8.0 +2023-03-09 04:55:09,367 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=46406.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:55:12,701 INFO [train.py:898] (0/4) Epoch 13, batch 2800, loss[loss=0.1741, simple_loss=0.2661, pruned_loss=0.04107, over 18392.00 frames. ], tot_loss[loss=0.1811, simple_loss=0.2679, pruned_loss=0.04716, over 3593977.98 frames. ], batch size: 52, lr: 8.66e-03, grad_scale: 8.0 +2023-03-09 04:55:38,015 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46430.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:55:41,494 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7726, 4.7249, 4.8516, 4.6321, 4.6226, 4.6853, 4.9767, 4.9277], + device='cuda:0'), covar=tensor([0.0071, 0.0084, 0.0066, 0.0093, 0.0068, 0.0120, 0.0071, 0.0123], + device='cuda:0'), in_proj_covar=tensor([0.0084, 0.0060, 0.0062, 0.0080, 0.0067, 0.0092, 0.0077, 0.0077], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 04:56:07,317 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.145e+02 3.038e+02 3.515e+02 4.163e+02 1.533e+03, threshold=7.029e+02, percent-clipped=2.0 +2023-03-09 04:56:11,748 INFO [train.py:898] (0/4) Epoch 13, batch 2850, loss[loss=0.1723, simple_loss=0.2623, pruned_loss=0.04118, over 18396.00 frames. ], tot_loss[loss=0.181, simple_loss=0.2674, pruned_loss=0.04727, over 3585662.17 frames. ], batch size: 48, lr: 8.65e-03, grad_scale: 8.0 +2023-03-09 04:56:34,721 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=46478.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:56:40,688 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4074, 5.3491, 4.9789, 5.3307, 5.3024, 4.7272, 5.2351, 4.9407], + device='cuda:0'), covar=tensor([0.0356, 0.0394, 0.1331, 0.0631, 0.0511, 0.0396, 0.0390, 0.0966], + device='cuda:0'), in_proj_covar=tensor([0.0424, 0.0493, 0.0649, 0.0389, 0.0379, 0.0447, 0.0473, 0.0616], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 04:57:10,806 INFO [train.py:898] (0/4) Epoch 13, batch 2900, loss[loss=0.1556, simple_loss=0.2348, pruned_loss=0.03816, over 18379.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.267, pruned_loss=0.04736, over 3585552.68 frames. ], batch size: 42, lr: 8.65e-03, grad_scale: 8.0 +2023-03-09 04:57:18,093 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46515.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:58:05,311 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.838e+02 2.971e+02 3.539e+02 4.206e+02 8.689e+02, threshold=7.079e+02, percent-clipped=3.0 +2023-03-09 04:58:09,939 INFO [train.py:898] (0/4) Epoch 13, batch 2950, loss[loss=0.1762, simple_loss=0.2616, pruned_loss=0.04539, over 18546.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2662, pruned_loss=0.04674, over 3587418.66 frames. ], batch size: 49, lr: 8.65e-03, grad_scale: 8.0 +2023-03-09 04:58:14,681 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=46563.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:58:30,160 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-09 04:58:52,082 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46595.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:58:58,691 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3155, 4.7706, 4.3352, 4.5993, 4.4058, 4.4333, 4.8140, 4.7728], + device='cuda:0'), covar=tensor([0.1152, 0.0711, 0.1716, 0.0716, 0.1441, 0.0700, 0.0667, 0.0674], + device='cuda:0'), in_proj_covar=tensor([0.0551, 0.0462, 0.0350, 0.0499, 0.0676, 0.0497, 0.0653, 0.0489], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 04:59:01,057 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46602.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 04:59:09,182 INFO [train.py:898] (0/4) Epoch 13, batch 3000, loss[loss=0.1654, simple_loss=0.2569, pruned_loss=0.03694, over 18293.00 frames. ], tot_loss[loss=0.1809, simple_loss=0.2674, pruned_loss=0.04722, over 3584381.10 frames. ], batch size: 49, lr: 8.64e-03, grad_scale: 8.0 +2023-03-09 04:59:09,184 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 04:59:21,020 INFO [train.py:932] (0/4) Epoch 13, validation: loss=0.1542, simple_loss=0.256, pruned_loss=0.02615, over 944034.00 frames. +2023-03-09 04:59:21,021 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-09 04:59:47,889 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46632.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:00:15,394 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.844e+02 2.916e+02 3.524e+02 4.268e+02 9.781e+02, threshold=7.048e+02, percent-clipped=5.0 +2023-03-09 05:00:17,020 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46656.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 05:00:19,955 INFO [train.py:898] (0/4) Epoch 13, batch 3050, loss[loss=0.209, simple_loss=0.2916, pruned_loss=0.06318, over 18280.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2668, pruned_loss=0.04724, over 3585299.93 frames. ], batch size: 57, lr: 8.64e-03, grad_scale: 8.0 +2023-03-09 05:00:24,847 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46663.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:00:25,223 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-09 05:01:02,455 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8594, 3.2374, 4.3473, 3.9834, 3.1131, 4.7202, 4.0220, 3.0397], + device='cuda:0'), covar=tensor([0.0392, 0.1105, 0.0213, 0.0344, 0.1304, 0.0137, 0.0460, 0.0947], + device='cuda:0'), in_proj_covar=tensor([0.0197, 0.0225, 0.0167, 0.0151, 0.0215, 0.0194, 0.0217, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 05:01:18,395 INFO [train.py:898] (0/4) Epoch 13, batch 3100, loss[loss=0.1621, simple_loss=0.2477, pruned_loss=0.03825, over 18287.00 frames. ], tot_loss[loss=0.1803, simple_loss=0.2665, pruned_loss=0.04704, over 3594580.02 frames. ], batch size: 49, lr: 8.63e-03, grad_scale: 8.0 +2023-03-09 05:01:21,720 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.64 vs. limit=2.0 +2023-03-09 05:02:00,964 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-09 05:02:12,195 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.124e+02 3.204e+02 3.767e+02 4.418e+02 1.241e+03, threshold=7.535e+02, percent-clipped=5.0 +2023-03-09 05:02:16,820 INFO [train.py:898] (0/4) Epoch 13, batch 3150, loss[loss=0.1671, simple_loss=0.2512, pruned_loss=0.04154, over 18510.00 frames. ], tot_loss[loss=0.1807, simple_loss=0.2665, pruned_loss=0.04743, over 3577617.24 frames. ], batch size: 47, lr: 8.63e-03, grad_scale: 8.0 +2023-03-09 05:02:54,722 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3340, 4.4184, 2.6750, 4.2630, 5.4452, 2.5720, 3.9245, 4.2937], + device='cuda:0'), covar=tensor([0.0095, 0.1059, 0.1389, 0.0548, 0.0054, 0.1177, 0.0594, 0.0597], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0241, 0.0194, 0.0191, 0.0096, 0.0175, 0.0205, 0.0208], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 05:03:16,830 INFO [train.py:898] (0/4) Epoch 13, batch 3200, loss[loss=0.1624, simple_loss=0.2494, pruned_loss=0.03771, over 18487.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2651, pruned_loss=0.04678, over 3586856.82 frames. ], batch size: 47, lr: 8.62e-03, grad_scale: 8.0 +2023-03-09 05:03:18,307 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=46810.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:03:51,434 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-09 05:04:10,408 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.338e+02 3.187e+02 3.650e+02 4.694e+02 1.027e+03, threshold=7.300e+02, percent-clipped=4.0 +2023-03-09 05:04:15,581 INFO [train.py:898] (0/4) Epoch 13, batch 3250, loss[loss=0.1942, simple_loss=0.2869, pruned_loss=0.05074, over 18348.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.2657, pruned_loss=0.04709, over 3571949.61 frames. ], batch size: 55, lr: 8.62e-03, grad_scale: 8.0 +2023-03-09 05:04:29,639 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=46871.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:05:14,058 INFO [train.py:898] (0/4) Epoch 13, batch 3300, loss[loss=0.1587, simple_loss=0.2452, pruned_loss=0.0361, over 18426.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2657, pruned_loss=0.04719, over 3565044.26 frames. ], batch size: 48, lr: 8.61e-03, grad_scale: 8.0 +2023-03-09 05:05:20,188 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6355, 2.1685, 2.4833, 2.7905, 3.0573, 5.1950, 4.8150, 3.7806], + device='cuda:0'), covar=tensor([0.1677, 0.2642, 0.3165, 0.1608, 0.2859, 0.0141, 0.0349, 0.0686], + device='cuda:0'), in_proj_covar=tensor([0.0260, 0.0316, 0.0338, 0.0254, 0.0369, 0.0202, 0.0272, 0.0220], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 05:05:41,234 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=46932.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:06:03,362 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46951.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 05:06:07,663 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.054e+02 2.913e+02 3.556e+02 4.553e+02 1.517e+03, threshold=7.113e+02, percent-clipped=7.0 +2023-03-09 05:06:11,357 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=46958.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:06:12,410 INFO [train.py:898] (0/4) Epoch 13, batch 3350, loss[loss=0.1786, simple_loss=0.2623, pruned_loss=0.04745, over 18415.00 frames. ], tot_loss[loss=0.1805, simple_loss=0.2667, pruned_loss=0.04714, over 3583681.73 frames. ], batch size: 48, lr: 8.61e-03, grad_scale: 8.0 +2023-03-09 05:06:38,279 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=46980.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:07:12,263 INFO [train.py:898] (0/4) Epoch 13, batch 3400, loss[loss=0.1664, simple_loss=0.2542, pruned_loss=0.03925, over 18355.00 frames. ], tot_loss[loss=0.1806, simple_loss=0.2667, pruned_loss=0.04731, over 3582169.27 frames. ], batch size: 55, lr: 8.60e-03, grad_scale: 8.0 +2023-03-09 05:07:44,547 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-09 05:07:46,897 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-09 05:08:07,266 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.855e+02 2.739e+02 3.192e+02 3.909e+02 7.528e+02, threshold=6.383e+02, percent-clipped=0.0 +2023-03-09 05:08:11,965 INFO [train.py:898] (0/4) Epoch 13, batch 3450, loss[loss=0.1723, simple_loss=0.2667, pruned_loss=0.03892, over 18385.00 frames. ], tot_loss[loss=0.1799, simple_loss=0.266, pruned_loss=0.04691, over 3585973.40 frames. ], batch size: 52, lr: 8.60e-03, grad_scale: 8.0 +2023-03-09 05:09:11,355 INFO [train.py:898] (0/4) Epoch 13, batch 3500, loss[loss=0.1826, simple_loss=0.2632, pruned_loss=0.05098, over 18569.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2647, pruned_loss=0.04636, over 3593594.36 frames. ], batch size: 45, lr: 8.60e-03, grad_scale: 8.0 +2023-03-09 05:10:03,843 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.269e+02 2.980e+02 3.583e+02 4.065e+02 6.381e+02, threshold=7.167e+02, percent-clipped=1.0 +2023-03-09 05:10:08,140 INFO [train.py:898] (0/4) Epoch 13, batch 3550, loss[loss=0.1575, simple_loss=0.2362, pruned_loss=0.03942, over 18499.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2654, pruned_loss=0.04644, over 3600707.00 frames. ], batch size: 44, lr: 8.59e-03, grad_scale: 8.0 +2023-03-09 05:10:15,814 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=47166.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:10:30,882 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5708, 3.5608, 3.3262, 2.9009, 3.2052, 2.6024, 2.7821, 3.4253], + device='cuda:0'), covar=tensor([0.0069, 0.0100, 0.0106, 0.0158, 0.0128, 0.0218, 0.0199, 0.0078], + device='cuda:0'), in_proj_covar=tensor([0.0109, 0.0132, 0.0114, 0.0163, 0.0115, 0.0160, 0.0162, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 05:11:02,843 INFO [train.py:898] (0/4) Epoch 13, batch 3600, loss[loss=0.1582, simple_loss=0.2345, pruned_loss=0.04093, over 18512.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2647, pruned_loss=0.04628, over 3603417.85 frames. ], batch size: 44, lr: 8.59e-03, grad_scale: 8.0 +2023-03-09 05:11:36,437 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47241.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:11:38,682 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-13.pt +2023-03-09 05:12:09,422 INFO [train.py:898] (0/4) Epoch 14, batch 0, loss[loss=0.2044, simple_loss=0.2922, pruned_loss=0.0583, over 18278.00 frames. ], tot_loss[loss=0.2044, simple_loss=0.2922, pruned_loss=0.0583, over 18278.00 frames. ], batch size: 57, lr: 8.27e-03, grad_scale: 8.0 +2023-03-09 05:12:09,424 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 05:12:21,336 INFO [train.py:932] (0/4) Epoch 14, validation: loss=0.155, simple_loss=0.2569, pruned_loss=0.0266, over 944034.00 frames. +2023-03-09 05:12:21,337 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19781MB +2023-03-09 05:12:28,514 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6863, 2.8255, 2.6419, 2.9656, 3.6814, 3.6215, 3.2030, 2.9803], + device='cuda:0'), covar=tensor([0.0150, 0.0309, 0.0542, 0.0302, 0.0144, 0.0131, 0.0315, 0.0311], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0114, 0.0151, 0.0142, 0.0107, 0.0096, 0.0137, 0.0136], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 05:12:31,276 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=47251.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:12:35,444 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.980e+02 3.395e+02 4.114e+02 5.070e+02 1.381e+03, threshold=8.228e+02, percent-clipped=11.0 +2023-03-09 05:12:39,235 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=47258.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:13:19,478 INFO [train.py:898] (0/4) Epoch 14, batch 50, loss[loss=0.1539, simple_loss=0.2415, pruned_loss=0.03317, over 18471.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2658, pruned_loss=0.04632, over 813645.71 frames. ], batch size: 44, lr: 8.27e-03, grad_scale: 8.0 +2023-03-09 05:13:26,861 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=47299.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:13:30,389 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47302.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:13:35,367 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=47306.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:14:18,546 INFO [train.py:898] (0/4) Epoch 14, batch 100, loss[loss=0.179, simple_loss=0.2708, pruned_loss=0.04359, over 18297.00 frames. ], tot_loss[loss=0.1822, simple_loss=0.2693, pruned_loss=0.04749, over 1424507.29 frames. ], batch size: 57, lr: 8.26e-03, grad_scale: 8.0 +2023-03-09 05:14:29,861 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7400, 3.7221, 5.0295, 4.5214, 3.2661, 2.9193, 4.4816, 5.3002], + device='cuda:0'), covar=tensor([0.0821, 0.1548, 0.0163, 0.0347, 0.0934, 0.1195, 0.0355, 0.0151], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0250, 0.0115, 0.0163, 0.0178, 0.0176, 0.0176, 0.0159], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 05:14:32,862 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.774e+02 2.933e+02 3.364e+02 4.152e+02 6.819e+02, threshold=6.727e+02, percent-clipped=0.0 +2023-03-09 05:15:02,495 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47379.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:15:16,218 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0880, 5.1167, 2.4783, 5.0602, 4.8312, 5.1450, 4.8708, 2.4223], + device='cuda:0'), covar=tensor([0.0168, 0.0091, 0.0954, 0.0085, 0.0091, 0.0099, 0.0137, 0.1317], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0071, 0.0089, 0.0086, 0.0078, 0.0067, 0.0078, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 05:15:18,119 INFO [train.py:898] (0/4) Epoch 14, batch 150, loss[loss=0.1843, simple_loss=0.2787, pruned_loss=0.0449, over 18400.00 frames. ], tot_loss[loss=0.1808, simple_loss=0.2679, pruned_loss=0.04686, over 1915078.12 frames. ], batch size: 52, lr: 8.26e-03, grad_scale: 8.0 +2023-03-09 05:16:15,661 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47440.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:16:18,708 INFO [train.py:898] (0/4) Epoch 14, batch 200, loss[loss=0.1899, simple_loss=0.278, pruned_loss=0.05089, over 18625.00 frames. ], tot_loss[loss=0.1785, simple_loss=0.2656, pruned_loss=0.04565, over 2297229.79 frames. ], batch size: 52, lr: 8.25e-03, grad_scale: 8.0 +2023-03-09 05:16:25,376 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-09 05:16:32,554 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.856e+02 2.798e+02 3.533e+02 4.067e+02 1.064e+03, threshold=7.066e+02, percent-clipped=5.0 +2023-03-09 05:16:36,990 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6469, 4.6328, 4.6993, 4.5131, 4.4915, 4.5977, 4.8308, 4.7972], + device='cuda:0'), covar=tensor([0.0072, 0.0074, 0.0069, 0.0097, 0.0072, 0.0111, 0.0078, 0.0113], + device='cuda:0'), in_proj_covar=tensor([0.0084, 0.0060, 0.0063, 0.0080, 0.0067, 0.0091, 0.0077, 0.0077], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 05:16:46,973 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=47466.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:17:18,352 INFO [train.py:898] (0/4) Epoch 14, batch 250, loss[loss=0.188, simple_loss=0.2754, pruned_loss=0.05025, over 18262.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2659, pruned_loss=0.04637, over 2575200.30 frames. ], batch size: 47, lr: 8.25e-03, grad_scale: 8.0 +2023-03-09 05:17:43,431 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=47514.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:18:16,913 INFO [train.py:898] (0/4) Epoch 14, batch 300, loss[loss=0.179, simple_loss=0.2658, pruned_loss=0.04614, over 18539.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2669, pruned_loss=0.04657, over 2801527.20 frames. ], batch size: 49, lr: 8.24e-03, grad_scale: 8.0 +2023-03-09 05:18:30,803 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.041e+02 3.040e+02 3.610e+02 4.364e+02 8.752e+02, threshold=7.221e+02, percent-clipped=2.0 +2023-03-09 05:19:16,484 INFO [train.py:898] (0/4) Epoch 14, batch 350, loss[loss=0.2289, simple_loss=0.3146, pruned_loss=0.07162, over 18482.00 frames. ], tot_loss[loss=0.18, simple_loss=0.2665, pruned_loss=0.04672, over 2978628.86 frames. ], batch size: 59, lr: 8.24e-03, grad_scale: 8.0 +2023-03-09 05:19:21,120 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=47597.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:20:15,234 INFO [train.py:898] (0/4) Epoch 14, batch 400, loss[loss=0.1462, simple_loss=0.224, pruned_loss=0.03417, over 18461.00 frames. ], tot_loss[loss=0.1795, simple_loss=0.2663, pruned_loss=0.04634, over 3112590.93 frames. ], batch size: 43, lr: 8.24e-03, grad_scale: 8.0 +2023-03-09 05:20:28,438 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.702e+02 2.895e+02 3.399e+02 4.132e+02 9.040e+02, threshold=6.798e+02, percent-clipped=2.0 +2023-03-09 05:21:14,181 INFO [train.py:898] (0/4) Epoch 14, batch 450, loss[loss=0.1534, simple_loss=0.243, pruned_loss=0.03193, over 18385.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2647, pruned_loss=0.04558, over 3224929.85 frames. ], batch size: 46, lr: 8.23e-03, grad_scale: 8.0 +2023-03-09 05:21:20,788 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-09 05:21:25,443 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-09 05:22:04,128 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=47735.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:22:07,628 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47738.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:22:13,624 INFO [train.py:898] (0/4) Epoch 14, batch 500, loss[loss=0.1816, simple_loss=0.2798, pruned_loss=0.04169, over 17209.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2642, pruned_loss=0.04536, over 3314516.49 frames. ], batch size: 78, lr: 8.23e-03, grad_scale: 8.0 +2023-03-09 05:22:27,493 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.904e+02 2.746e+02 3.457e+02 4.087e+02 8.380e+02, threshold=6.915e+02, percent-clipped=1.0 +2023-03-09 05:22:40,868 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8680, 3.8533, 5.2990, 4.6591, 3.2819, 3.0064, 4.4731, 5.3844], + device='cuda:0'), covar=tensor([0.0812, 0.1614, 0.0124, 0.0334, 0.0983, 0.1254, 0.0438, 0.0186], + device='cuda:0'), in_proj_covar=tensor([0.0142, 0.0258, 0.0118, 0.0170, 0.0184, 0.0184, 0.0183, 0.0167], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 05:22:50,322 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.47 vs. limit=5.0 +2023-03-09 05:23:06,873 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47788.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:23:12,624 INFO [train.py:898] (0/4) Epoch 14, batch 550, loss[loss=0.1802, simple_loss=0.2693, pruned_loss=0.04557, over 18410.00 frames. ], tot_loss[loss=0.1787, simple_loss=0.2654, pruned_loss=0.04603, over 3377675.10 frames. ], batch size: 52, lr: 8.22e-03, grad_scale: 8.0 +2023-03-09 05:23:20,542 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47799.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:23:31,826 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=47809.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 05:24:12,015 INFO [train.py:898] (0/4) Epoch 14, batch 600, loss[loss=0.1999, simple_loss=0.2876, pruned_loss=0.05612, over 17934.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2661, pruned_loss=0.04627, over 3425109.78 frames. ], batch size: 65, lr: 8.22e-03, grad_scale: 8.0 +2023-03-09 05:24:19,649 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47849.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:24:25,847 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.360e+02 2.943e+02 3.448e+02 4.485e+02 1.001e+03, threshold=6.896e+02, percent-clipped=3.0 +2023-03-09 05:24:43,104 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=47870.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 05:25:10,121 INFO [train.py:898] (0/4) Epoch 14, batch 650, loss[loss=0.2285, simple_loss=0.2992, pruned_loss=0.07884, over 12428.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2663, pruned_loss=0.0463, over 3452696.26 frames. ], batch size: 129, lr: 8.21e-03, grad_scale: 8.0 +2023-03-09 05:25:15,361 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=47897.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:26:07,450 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.81 vs. limit=2.0 +2023-03-09 05:26:09,521 INFO [train.py:898] (0/4) Epoch 14, batch 700, loss[loss=0.1979, simple_loss=0.2817, pruned_loss=0.05704, over 17179.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2652, pruned_loss=0.04601, over 3484193.39 frames. ], batch size: 78, lr: 8.21e-03, grad_scale: 8.0 +2023-03-09 05:26:12,030 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=47945.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:26:19,376 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.62 vs. limit=2.0 +2023-03-09 05:26:23,811 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.052e+02 2.964e+02 3.596e+02 4.648e+02 7.874e+02, threshold=7.192e+02, percent-clipped=3.0 +2023-03-09 05:26:59,048 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8063, 3.6155, 4.8565, 2.9149, 4.1453, 2.4497, 3.0984, 1.8821], + device='cuda:0'), covar=tensor([0.1032, 0.0809, 0.0111, 0.0752, 0.0632, 0.2366, 0.2408, 0.1771], + device='cuda:0'), in_proj_covar=tensor([0.0206, 0.0224, 0.0134, 0.0178, 0.0236, 0.0252, 0.0296, 0.0212], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 05:27:08,327 INFO [train.py:898] (0/4) Epoch 14, batch 750, loss[loss=0.1611, simple_loss=0.2411, pruned_loss=0.04053, over 18444.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2657, pruned_loss=0.04614, over 3511095.19 frames. ], batch size: 43, lr: 8.21e-03, grad_scale: 16.0 +2023-03-09 05:27:17,391 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-48000.pt +2023-03-09 05:28:03,347 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48035.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:28:12,891 INFO [train.py:898] (0/4) Epoch 14, batch 800, loss[loss=0.1752, simple_loss=0.2685, pruned_loss=0.04093, over 18362.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2656, pruned_loss=0.04597, over 3536137.89 frames. ], batch size: 50, lr: 8.20e-03, grad_scale: 8.0 +2023-03-09 05:28:28,354 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.329e+02 3.109e+02 3.556e+02 4.248e+02 1.366e+03, threshold=7.111e+02, percent-clipped=4.0 +2023-03-09 05:28:47,984 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48072.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:29:00,370 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=48083.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:29:11,643 INFO [train.py:898] (0/4) Epoch 14, batch 850, loss[loss=0.1794, simple_loss=0.2627, pruned_loss=0.0481, over 18503.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2658, pruned_loss=0.04606, over 3547735.95 frames. ], batch size: 47, lr: 8.20e-03, grad_scale: 8.0 +2023-03-09 05:29:13,079 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48094.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:29:59,974 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48133.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:30:10,878 INFO [train.py:898] (0/4) Epoch 14, batch 900, loss[loss=0.1719, simple_loss=0.264, pruned_loss=0.0399, over 18354.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2658, pruned_loss=0.04597, over 3571137.06 frames. ], batch size: 56, lr: 8.19e-03, grad_scale: 8.0 +2023-03-09 05:30:12,171 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48144.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:30:26,934 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.713e+02 2.967e+02 3.519e+02 4.098e+02 9.060e+02, threshold=7.037e+02, percent-clipped=1.0 +2023-03-09 05:30:38,213 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48165.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 05:31:10,237 INFO [train.py:898] (0/4) Epoch 14, batch 950, loss[loss=0.164, simple_loss=0.2398, pruned_loss=0.04407, over 18557.00 frames. ], tot_loss[loss=0.1788, simple_loss=0.2652, pruned_loss=0.04615, over 3553508.96 frames. ], batch size: 45, lr: 8.19e-03, grad_scale: 8.0 +2023-03-09 05:31:31,953 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-09 05:31:34,372 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48212.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 05:31:45,196 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9071, 4.8624, 4.9779, 4.7606, 4.7739, 4.8592, 5.1188, 5.0620], + device='cuda:0'), covar=tensor([0.0063, 0.0072, 0.0059, 0.0098, 0.0059, 0.0108, 0.0084, 0.0101], + device='cuda:0'), in_proj_covar=tensor([0.0083, 0.0059, 0.0063, 0.0079, 0.0066, 0.0090, 0.0075, 0.0076], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 05:32:09,656 INFO [train.py:898] (0/4) Epoch 14, batch 1000, loss[loss=0.1802, simple_loss=0.2735, pruned_loss=0.04348, over 18398.00 frames. ], tot_loss[loss=0.1794, simple_loss=0.2662, pruned_loss=0.04631, over 3551336.32 frames. ], batch size: 52, lr: 8.19e-03, grad_scale: 8.0 +2023-03-09 05:32:26,019 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.854e+02 2.987e+02 3.542e+02 4.398e+02 9.134e+02, threshold=7.083e+02, percent-clipped=3.0 +2023-03-09 05:32:46,754 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48273.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 05:32:50,124 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4421, 6.0672, 5.5565, 5.8351, 5.6893, 5.4928, 6.1436, 6.0913], + device='cuda:0'), covar=tensor([0.1339, 0.0694, 0.0396, 0.0664, 0.1348, 0.0634, 0.0519, 0.0576], + device='cuda:0'), in_proj_covar=tensor([0.0550, 0.0466, 0.0342, 0.0492, 0.0668, 0.0492, 0.0649, 0.0481], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 05:32:55,826 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5485, 3.5927, 5.2171, 4.5540, 3.3287, 3.2424, 4.6224, 5.3618], + device='cuda:0'), covar=tensor([0.0836, 0.1853, 0.0105, 0.0307, 0.0891, 0.0993, 0.0317, 0.0176], + device='cuda:0'), in_proj_covar=tensor([0.0139, 0.0258, 0.0118, 0.0170, 0.0181, 0.0181, 0.0183, 0.0166], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 05:33:09,317 INFO [train.py:898] (0/4) Epoch 14, batch 1050, loss[loss=0.1772, simple_loss=0.2569, pruned_loss=0.04879, over 18264.00 frames. ], tot_loss[loss=0.179, simple_loss=0.266, pruned_loss=0.04602, over 3568146.96 frames. ], batch size: 47, lr: 8.18e-03, grad_scale: 8.0 +2023-03-09 05:33:46,612 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48324.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:34:08,881 INFO [train.py:898] (0/4) Epoch 14, batch 1100, loss[loss=0.1375, simple_loss=0.2232, pruned_loss=0.02591, over 18167.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.2657, pruned_loss=0.04601, over 3580022.61 frames. ], batch size: 44, lr: 8.18e-03, grad_scale: 8.0 +2023-03-09 05:34:23,494 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.341e+02 3.220e+02 3.737e+02 4.236e+02 6.762e+02, threshold=7.474e+02, percent-clipped=0.0 +2023-03-09 05:34:59,339 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48385.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:35:08,046 INFO [train.py:898] (0/4) Epoch 14, batch 1150, loss[loss=0.1678, simple_loss=0.2514, pruned_loss=0.04212, over 18411.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2651, pruned_loss=0.0457, over 3592807.94 frames. ], batch size: 48, lr: 8.17e-03, grad_scale: 8.0 +2023-03-09 05:35:09,486 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48394.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:35:12,833 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48397.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:35:23,514 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7955, 3.6642, 4.9125, 2.8832, 4.3313, 2.5034, 2.9891, 1.8715], + device='cuda:0'), covar=tensor([0.0993, 0.0842, 0.0148, 0.0751, 0.0559, 0.2511, 0.2625, 0.1824], + device='cuda:0'), in_proj_covar=tensor([0.0207, 0.0226, 0.0135, 0.0180, 0.0239, 0.0253, 0.0300, 0.0216], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 05:35:40,810 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0068, 3.9633, 5.2302, 3.1434, 4.6133, 2.8152, 3.2235, 1.9844], + device='cuda:0'), covar=tensor([0.0928, 0.0742, 0.0083, 0.0701, 0.0492, 0.2206, 0.2526, 0.1764], + device='cuda:0'), in_proj_covar=tensor([0.0206, 0.0224, 0.0135, 0.0179, 0.0238, 0.0252, 0.0298, 0.0215], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 05:35:44,598 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48423.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:35:49,900 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48428.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:36:06,282 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=48442.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:36:07,210 INFO [train.py:898] (0/4) Epoch 14, batch 1200, loss[loss=0.1819, simple_loss=0.267, pruned_loss=0.0484, over 16959.00 frames. ], tot_loss[loss=0.1789, simple_loss=0.266, pruned_loss=0.04592, over 3595809.81 frames. ], batch size: 78, lr: 8.17e-03, grad_scale: 8.0 +2023-03-09 05:36:08,590 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48444.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:36:22,327 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.965e+02 3.035e+02 3.656e+02 4.283e+02 1.032e+03, threshold=7.311e+02, percent-clipped=2.0 +2023-03-09 05:36:23,905 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2164, 2.6310, 2.3399, 2.5303, 3.4419, 3.3630, 2.9123, 2.6945], + device='cuda:0'), covar=tensor([0.0195, 0.0271, 0.0550, 0.0402, 0.0168, 0.0147, 0.0383, 0.0363], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0114, 0.0152, 0.0143, 0.0108, 0.0095, 0.0136, 0.0137], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0002, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 05:36:24,948 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48458.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:36:32,236 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.75 vs. limit=5.0 +2023-03-09 05:36:32,992 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48465.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 05:36:49,437 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0478, 5.0691, 5.1469, 4.8869, 4.7605, 4.9972, 5.2999, 5.2382], + device='cuda:0'), covar=tensor([0.0062, 0.0060, 0.0056, 0.0091, 0.0064, 0.0116, 0.0064, 0.0087], + device='cuda:0'), in_proj_covar=tensor([0.0084, 0.0060, 0.0064, 0.0081, 0.0067, 0.0091, 0.0076, 0.0077], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 05:36:56,409 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48484.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:37:05,997 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=48492.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:37:06,944 INFO [train.py:898] (0/4) Epoch 14, batch 1250, loss[loss=0.1828, simple_loss=0.2682, pruned_loss=0.04869, over 17239.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2655, pruned_loss=0.04569, over 3597929.22 frames. ], batch size: 78, lr: 8.16e-03, grad_scale: 8.0 +2023-03-09 05:37:30,367 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=48513.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 05:37:42,520 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6390, 3.5206, 3.4315, 3.0087, 3.3274, 2.6136, 2.6155, 3.6104], + device='cuda:0'), covar=tensor([0.0053, 0.0070, 0.0071, 0.0137, 0.0085, 0.0193, 0.0187, 0.0050], + device='cuda:0'), in_proj_covar=tensor([0.0111, 0.0136, 0.0117, 0.0167, 0.0118, 0.0163, 0.0165, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0001, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 05:38:04,674 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2541, 5.1669, 5.4326, 5.4512, 5.1166, 6.0295, 5.5324, 5.2563], + device='cuda:0'), covar=tensor([0.1064, 0.0646, 0.0658, 0.0707, 0.1476, 0.0705, 0.0626, 0.1643], + device='cuda:0'), in_proj_covar=tensor([0.0325, 0.0256, 0.0268, 0.0272, 0.0307, 0.0379, 0.0249, 0.0373], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0004, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 05:38:06,732 INFO [train.py:898] (0/4) Epoch 14, batch 1300, loss[loss=0.1896, simple_loss=0.2817, pruned_loss=0.04878, over 18454.00 frames. ], tot_loss[loss=0.1791, simple_loss=0.266, pruned_loss=0.04605, over 3599110.61 frames. ], batch size: 59, lr: 8.16e-03, grad_scale: 8.0 +2023-03-09 05:38:21,329 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.028e+02 2.821e+02 3.494e+02 4.256e+02 8.799e+02, threshold=6.987e+02, percent-clipped=3.0 +2023-03-09 05:38:35,199 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48568.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 05:39:05,484 INFO [train.py:898] (0/4) Epoch 14, batch 1350, loss[loss=0.1994, simple_loss=0.2887, pruned_loss=0.05507, over 15987.00 frames. ], tot_loss[loss=0.1793, simple_loss=0.2662, pruned_loss=0.04623, over 3581432.62 frames. ], batch size: 94, lr: 8.16e-03, grad_scale: 8.0 +2023-03-09 05:39:29,091 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5632, 3.6604, 4.9998, 4.3393, 3.3789, 3.0163, 4.4308, 5.2371], + device='cuda:0'), covar=tensor([0.0810, 0.1683, 0.0155, 0.0369, 0.0823, 0.1074, 0.0355, 0.0160], + device='cuda:0'), in_proj_covar=tensor([0.0137, 0.0253, 0.0117, 0.0167, 0.0178, 0.0178, 0.0181, 0.0164], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 05:40:05,026 INFO [train.py:898] (0/4) Epoch 14, batch 1400, loss[loss=0.179, simple_loss=0.277, pruned_loss=0.04049, over 18350.00 frames. ], tot_loss[loss=0.179, simple_loss=0.266, pruned_loss=0.04604, over 3588328.29 frames. ], batch size: 55, lr: 8.15e-03, grad_scale: 8.0 +2023-03-09 05:40:19,853 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.683e+02 2.979e+02 3.618e+02 4.043e+02 7.873e+02, threshold=7.235e+02, percent-clipped=2.0 +2023-03-09 05:40:48,792 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48680.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:41:04,620 INFO [train.py:898] (0/4) Epoch 14, batch 1450, loss[loss=0.1706, simple_loss=0.2604, pruned_loss=0.04035, over 18295.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2655, pruned_loss=0.04587, over 3581060.83 frames. ], batch size: 54, lr: 8.15e-03, grad_scale: 8.0 +2023-03-09 05:41:28,004 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0543, 5.5202, 2.7190, 5.3458, 5.1859, 5.6022, 5.3512, 3.0507], + device='cuda:0'), covar=tensor([0.0181, 0.0058, 0.0810, 0.0067, 0.0077, 0.0057, 0.0076, 0.0834], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0071, 0.0089, 0.0085, 0.0078, 0.0066, 0.0078, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 05:41:45,729 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48728.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:41:56,078 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8263, 4.8511, 4.9116, 4.6324, 4.6459, 4.7232, 5.0492, 5.0271], + device='cuda:0'), covar=tensor([0.0074, 0.0066, 0.0072, 0.0102, 0.0061, 0.0116, 0.0069, 0.0094], + device='cuda:0'), in_proj_covar=tensor([0.0086, 0.0061, 0.0064, 0.0083, 0.0068, 0.0093, 0.0077, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 05:42:03,618 INFO [train.py:898] (0/4) Epoch 14, batch 1500, loss[loss=0.1826, simple_loss=0.2595, pruned_loss=0.05289, over 18365.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2653, pruned_loss=0.04568, over 3578536.55 frames. ], batch size: 46, lr: 8.14e-03, grad_scale: 8.0 +2023-03-09 05:42:15,599 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48753.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:42:18,856 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.036e+02 2.981e+02 3.460e+02 4.226e+02 1.044e+03, threshold=6.921e+02, percent-clipped=2.0 +2023-03-09 05:42:42,122 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=48776.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:42:45,485 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=48779.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:42:45,609 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9096, 5.4497, 5.4305, 5.4075, 4.9430, 5.3534, 4.7350, 5.3059], + device='cuda:0'), covar=tensor([0.0269, 0.0254, 0.0190, 0.0369, 0.0357, 0.0219, 0.1021, 0.0295], + device='cuda:0'), in_proj_covar=tensor([0.0192, 0.0237, 0.0226, 0.0273, 0.0240, 0.0236, 0.0293, 0.0230], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0005, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 05:42:49,662 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8037, 2.9866, 4.3372, 3.9528, 2.6431, 4.6605, 3.9646, 2.9616], + device='cuda:0'), covar=tensor([0.0400, 0.1355, 0.0218, 0.0282, 0.1520, 0.0191, 0.0475, 0.0934], + device='cuda:0'), in_proj_covar=tensor([0.0196, 0.0225, 0.0169, 0.0146, 0.0214, 0.0189, 0.0218, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 05:43:00,924 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.99 vs. limit=5.0 +2023-03-09 05:43:02,592 INFO [train.py:898] (0/4) Epoch 14, batch 1550, loss[loss=0.1727, simple_loss=0.2643, pruned_loss=0.04054, over 18379.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2643, pruned_loss=0.04523, over 3594589.29 frames. ], batch size: 50, lr: 8.14e-03, grad_scale: 8.0 +2023-03-09 05:43:35,857 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48821.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:43:55,866 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=48838.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:44:02,359 INFO [train.py:898] (0/4) Epoch 14, batch 1600, loss[loss=0.1594, simple_loss=0.2492, pruned_loss=0.03478, over 18406.00 frames. ], tot_loss[loss=0.1776, simple_loss=0.2645, pruned_loss=0.04534, over 3586625.93 frames. ], batch size: 48, lr: 8.14e-03, grad_scale: 8.0 +2023-03-09 05:44:02,636 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9417, 3.9326, 3.9785, 3.8465, 3.8792, 3.9472, 4.0738, 4.0630], + device='cuda:0'), covar=tensor([0.0082, 0.0066, 0.0068, 0.0092, 0.0068, 0.0097, 0.0073, 0.0092], + device='cuda:0'), in_proj_covar=tensor([0.0084, 0.0060, 0.0063, 0.0081, 0.0067, 0.0091, 0.0076, 0.0077], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 05:44:17,658 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.056e+02 2.816e+02 3.506e+02 4.400e+02 1.387e+03, threshold=7.012e+02, percent-clipped=5.0 +2023-03-09 05:44:31,408 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48868.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 05:44:47,247 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48882.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:44:59,852 INFO [train.py:898] (0/4) Epoch 14, batch 1650, loss[loss=0.1668, simple_loss=0.2432, pruned_loss=0.04513, over 17654.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2649, pruned_loss=0.04574, over 3577288.67 frames. ], batch size: 39, lr: 8.13e-03, grad_scale: 8.0 +2023-03-09 05:45:07,547 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=48899.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:45:26,217 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8233, 4.3591, 4.4579, 3.1504, 3.6299, 3.3985, 2.5551, 2.2848], + device='cuda:0'), covar=tensor([0.0190, 0.0180, 0.0084, 0.0325, 0.0338, 0.0207, 0.0755, 0.0969], + device='cuda:0'), in_proj_covar=tensor([0.0061, 0.0051, 0.0052, 0.0062, 0.0082, 0.0058, 0.0073, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0004, 0.0004, 0.0006, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 05:45:27,194 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=48916.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 05:45:58,700 INFO [train.py:898] (0/4) Epoch 14, batch 1700, loss[loss=0.2291, simple_loss=0.2966, pruned_loss=0.08079, over 12696.00 frames. ], tot_loss[loss=0.1786, simple_loss=0.2652, pruned_loss=0.04594, over 3574749.79 frames. ], batch size: 129, lr: 8.13e-03, grad_scale: 8.0 +2023-03-09 05:46:15,064 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.904e+02 3.125e+02 3.663e+02 4.529e+02 8.545e+02, threshold=7.326e+02, percent-clipped=5.0 +2023-03-09 05:46:42,361 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=48980.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:46:42,411 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9226, 3.0712, 4.4242, 4.0630, 2.9259, 4.7752, 4.0847, 3.1125], + device='cuda:0'), covar=tensor([0.0364, 0.1316, 0.0227, 0.0291, 0.1271, 0.0160, 0.0421, 0.0885], + device='cuda:0'), in_proj_covar=tensor([0.0201, 0.0231, 0.0174, 0.0150, 0.0219, 0.0194, 0.0224, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 05:46:57,431 INFO [train.py:898] (0/4) Epoch 14, batch 1750, loss[loss=0.2197, simple_loss=0.3117, pruned_loss=0.0638, over 18270.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2653, pruned_loss=0.04579, over 3587618.54 frames. ], batch size: 57, lr: 8.12e-03, grad_scale: 8.0 +2023-03-09 05:47:04,057 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6350, 3.4188, 2.1918, 4.4178, 3.1245, 4.3577, 2.3342, 3.8812], + device='cuda:0'), covar=tensor([0.0544, 0.0775, 0.1343, 0.0374, 0.0786, 0.0289, 0.1164, 0.0347], + device='cuda:0'), in_proj_covar=tensor([0.0200, 0.0215, 0.0180, 0.0256, 0.0181, 0.0250, 0.0195, 0.0189], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 05:47:11,789 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9836, 5.3311, 2.7849, 5.1901, 5.0197, 5.3666, 5.1358, 2.8069], + device='cuda:0'), covar=tensor([0.0189, 0.0070, 0.0818, 0.0092, 0.0079, 0.0070, 0.0093, 0.0961], + device='cuda:0'), in_proj_covar=tensor([0.0080, 0.0072, 0.0092, 0.0087, 0.0080, 0.0068, 0.0080, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 05:47:39,584 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=49028.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:47:56,945 INFO [train.py:898] (0/4) Epoch 14, batch 1800, loss[loss=0.1841, simple_loss=0.2747, pruned_loss=0.04679, over 18472.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2644, pruned_loss=0.04548, over 3585192.88 frames. ], batch size: 59, lr: 8.12e-03, grad_scale: 8.0 +2023-03-09 05:48:09,358 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49053.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:48:13,402 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.931e+02 2.916e+02 3.364e+02 3.970e+02 9.603e+02, threshold=6.727e+02, percent-clipped=4.0 +2023-03-09 05:48:40,845 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49079.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:48:56,153 INFO [train.py:898] (0/4) Epoch 14, batch 1850, loss[loss=0.2074, simple_loss=0.2946, pruned_loss=0.06008, over 17813.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.2648, pruned_loss=0.04581, over 3590587.48 frames. ], batch size: 70, lr: 8.12e-03, grad_scale: 8.0 +2023-03-09 05:49:05,988 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=49101.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:49:10,279 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0747, 4.2102, 2.5038, 4.2473, 5.2056, 2.5382, 3.8261, 3.8315], + device='cuda:0'), covar=tensor([0.0125, 0.1129, 0.1676, 0.0576, 0.0067, 0.1342, 0.0702, 0.0817], + device='cuda:0'), in_proj_covar=tensor([0.0137, 0.0250, 0.0196, 0.0194, 0.0100, 0.0179, 0.0209, 0.0216], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 05:49:13,488 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-09 05:49:37,849 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=49127.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:49:40,216 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1858, 5.6069, 2.8629, 5.3937, 5.2479, 5.6762, 5.4528, 3.0226], + device='cuda:0'), covar=tensor([0.0153, 0.0053, 0.0711, 0.0064, 0.0070, 0.0050, 0.0063, 0.0824], + device='cuda:0'), in_proj_covar=tensor([0.0079, 0.0071, 0.0090, 0.0085, 0.0079, 0.0067, 0.0078, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 05:49:55,816 INFO [train.py:898] (0/4) Epoch 14, batch 1900, loss[loss=0.1821, simple_loss=0.269, pruned_loss=0.04762, over 18396.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2649, pruned_loss=0.04564, over 3587442.51 frames. ], batch size: 50, lr: 8.11e-03, grad_scale: 8.0 +2023-03-09 05:50:11,575 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.722e+02 2.855e+02 3.283e+02 4.141e+02 6.742e+02, threshold=6.565e+02, percent-clipped=1.0 +2023-03-09 05:50:33,776 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.3648, 3.5835, 4.9186, 4.2612, 3.1154, 2.9389, 4.4065, 5.1580], + device='cuda:0'), covar=tensor([0.1002, 0.1830, 0.0188, 0.0416, 0.1116, 0.1311, 0.0415, 0.0221], + device='cuda:0'), in_proj_covar=tensor([0.0139, 0.0256, 0.0119, 0.0168, 0.0181, 0.0181, 0.0183, 0.0167], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 05:50:36,986 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49177.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:50:54,815 INFO [train.py:898] (0/4) Epoch 14, batch 1950, loss[loss=0.2095, simple_loss=0.3013, pruned_loss=0.05888, over 18493.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2636, pruned_loss=0.0453, over 3581536.19 frames. ], batch size: 51, lr: 8.11e-03, grad_scale: 8.0 +2023-03-09 05:50:56,123 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49194.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:51:26,522 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4830, 5.3995, 5.6372, 5.5608, 5.3548, 6.2489, 5.8768, 5.5626], + device='cuda:0'), covar=tensor([0.1111, 0.0568, 0.0694, 0.0752, 0.1438, 0.0684, 0.0617, 0.1600], + device='cuda:0'), in_proj_covar=tensor([0.0329, 0.0259, 0.0272, 0.0272, 0.0307, 0.0381, 0.0253, 0.0375], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0004, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 05:51:46,874 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49236.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:51:54,544 INFO [train.py:898] (0/4) Epoch 14, batch 2000, loss[loss=0.1924, simple_loss=0.2826, pruned_loss=0.05107, over 18001.00 frames. ], tot_loss[loss=0.1771, simple_loss=0.2637, pruned_loss=0.0452, over 3590973.04 frames. ], batch size: 65, lr: 8.10e-03, grad_scale: 8.0 +2023-03-09 05:52:09,969 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.669e+02 2.933e+02 3.417e+02 3.982e+02 2.319e+03, threshold=6.833e+02, percent-clipped=9.0 +2023-03-09 05:52:53,831 INFO [train.py:898] (0/4) Epoch 14, batch 2050, loss[loss=0.1785, simple_loss=0.2556, pruned_loss=0.05072, over 18392.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2639, pruned_loss=0.04506, over 3591966.96 frames. ], batch size: 48, lr: 8.10e-03, grad_scale: 8.0 +2023-03-09 05:52:58,929 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49297.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:53:54,013 INFO [train.py:898] (0/4) Epoch 14, batch 2100, loss[loss=0.2064, simple_loss=0.2958, pruned_loss=0.05843, over 17801.00 frames. ], tot_loss[loss=0.1779, simple_loss=0.2646, pruned_loss=0.04562, over 3582767.84 frames. ], batch size: 70, lr: 8.09e-03, grad_scale: 8.0 +2023-03-09 05:54:08,753 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.284e+02 2.957e+02 3.376e+02 4.046e+02 6.341e+02, threshold=6.752e+02, percent-clipped=0.0 +2023-03-09 05:54:40,042 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6713, 2.2069, 2.6588, 2.5469, 3.3376, 5.1647, 4.7997, 3.9827], + device='cuda:0'), covar=tensor([0.1428, 0.2229, 0.2716, 0.1713, 0.2036, 0.0131, 0.0341, 0.0577], + device='cuda:0'), in_proj_covar=tensor([0.0262, 0.0317, 0.0339, 0.0258, 0.0368, 0.0204, 0.0274, 0.0222], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 05:54:53,552 INFO [train.py:898] (0/4) Epoch 14, batch 2150, loss[loss=0.1802, simple_loss=0.2717, pruned_loss=0.04432, over 18494.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2648, pruned_loss=0.04563, over 3590204.30 frames. ], batch size: 51, lr: 8.09e-03, grad_scale: 8.0 +2023-03-09 05:55:27,881 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49422.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:55:28,997 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8630, 4.6625, 4.7734, 4.5054, 4.6207, 4.6531, 4.9394, 4.8910], + device='cuda:0'), covar=tensor([0.0079, 0.0103, 0.0077, 0.0122, 0.0075, 0.0132, 0.0096, 0.0114], + device='cuda:0'), in_proj_covar=tensor([0.0084, 0.0060, 0.0063, 0.0081, 0.0066, 0.0090, 0.0076, 0.0077], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 05:55:39,972 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49432.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 05:55:52,833 INFO [train.py:898] (0/4) Epoch 14, batch 2200, loss[loss=0.166, simple_loss=0.2489, pruned_loss=0.04148, over 18249.00 frames. ], tot_loss[loss=0.1782, simple_loss=0.265, pruned_loss=0.04572, over 3597448.93 frames. ], batch size: 47, lr: 8.09e-03, grad_scale: 8.0 +2023-03-09 05:56:04,482 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7084, 4.6714, 2.7639, 4.5197, 4.4152, 4.6935, 4.4991, 2.6318], + device='cuda:0'), covar=tensor([0.0210, 0.0072, 0.0674, 0.0119, 0.0083, 0.0071, 0.0099, 0.0944], + device='cuda:0'), in_proj_covar=tensor([0.0080, 0.0072, 0.0090, 0.0084, 0.0079, 0.0068, 0.0079, 0.0093], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 05:56:07,493 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.048e+02 2.848e+02 3.284e+02 4.322e+02 8.537e+02, threshold=6.567e+02, percent-clipped=2.0 +2023-03-09 05:56:33,202 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49477.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:56:40,680 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49483.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:56:52,244 INFO [train.py:898] (0/4) Epoch 14, batch 2250, loss[loss=0.1762, simple_loss=0.2633, pruned_loss=0.04454, over 18379.00 frames. ], tot_loss[loss=0.1792, simple_loss=0.2659, pruned_loss=0.04626, over 3576953.92 frames. ], batch size: 46, lr: 8.08e-03, grad_scale: 8.0 +2023-03-09 05:56:52,570 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49493.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 05:56:53,621 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49494.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:57:29,509 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=49525.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:57:44,494 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1339, 5.1388, 4.6054, 4.9807, 5.0005, 4.4948, 4.9990, 4.6336], + device='cuda:0'), covar=tensor([0.0558, 0.0614, 0.1918, 0.1075, 0.0751, 0.0578, 0.0566, 0.1251], + device='cuda:0'), in_proj_covar=tensor([0.0437, 0.0500, 0.0658, 0.0394, 0.0386, 0.0453, 0.0478, 0.0626], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 05:57:49,603 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=49542.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:57:50,495 INFO [train.py:898] (0/4) Epoch 14, batch 2300, loss[loss=0.2189, simple_loss=0.3057, pruned_loss=0.06607, over 18071.00 frames. ], tot_loss[loss=0.179, simple_loss=0.266, pruned_loss=0.04604, over 3592048.93 frames. ], batch size: 62, lr: 8.08e-03, grad_scale: 8.0 +2023-03-09 05:58:05,068 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.043e+02 3.294e+02 3.736e+02 4.526e+02 1.029e+03, threshold=7.472e+02, percent-clipped=8.0 +2023-03-09 05:58:47,135 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49592.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 05:58:48,639 INFO [train.py:898] (0/4) Epoch 14, batch 2350, loss[loss=0.2017, simple_loss=0.2946, pruned_loss=0.05444, over 18286.00 frames. ], tot_loss[loss=0.1784, simple_loss=0.2649, pruned_loss=0.04595, over 3584684.02 frames. ], batch size: 57, lr: 8.07e-03, grad_scale: 8.0 +2023-03-09 05:58:57,962 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-09 05:59:25,929 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6552, 3.5119, 3.4921, 2.9896, 3.2986, 2.5859, 2.6441, 3.5850], + device='cuda:0'), covar=tensor([0.0044, 0.0084, 0.0064, 0.0136, 0.0086, 0.0207, 0.0196, 0.0058], + device='cuda:0'), in_proj_covar=tensor([0.0114, 0.0137, 0.0117, 0.0171, 0.0121, 0.0166, 0.0168, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 05:59:48,077 INFO [train.py:898] (0/4) Epoch 14, batch 2400, loss[loss=0.1773, simple_loss=0.2686, pruned_loss=0.04302, over 18408.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2646, pruned_loss=0.04598, over 3558025.43 frames. ], batch size: 52, lr: 8.07e-03, grad_scale: 8.0 +2023-03-09 06:00:03,325 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.270e+02 3.059e+02 3.720e+02 4.555e+02 1.609e+03, threshold=7.441e+02, percent-clipped=3.0 +2023-03-09 06:00:19,422 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5477, 5.1398, 5.0466, 5.0428, 4.6468, 4.9694, 4.4406, 4.9530], + device='cuda:0'), covar=tensor([0.0257, 0.0288, 0.0246, 0.0475, 0.0384, 0.0239, 0.1118, 0.0341], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0235, 0.0228, 0.0277, 0.0239, 0.0238, 0.0291, 0.0227], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0005, 0.0007, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-09 06:00:38,643 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1079, 5.1533, 5.3059, 5.3465, 5.0256, 5.8798, 5.4172, 5.2054], + device='cuda:0'), covar=tensor([0.1065, 0.0655, 0.0707, 0.0670, 0.1402, 0.0731, 0.0643, 0.1659], + device='cuda:0'), in_proj_covar=tensor([0.0327, 0.0258, 0.0273, 0.0274, 0.0309, 0.0385, 0.0254, 0.0374], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0004, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 06:00:46,979 INFO [train.py:898] (0/4) Epoch 14, batch 2450, loss[loss=0.1763, simple_loss=0.267, pruned_loss=0.04277, over 18002.00 frames. ], tot_loss[loss=0.1783, simple_loss=0.2648, pruned_loss=0.04591, over 3560450.27 frames. ], batch size: 65, lr: 8.07e-03, grad_scale: 8.0 +2023-03-09 06:01:13,319 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.54 vs. limit=5.0 +2023-03-09 06:01:46,430 INFO [train.py:898] (0/4) Epoch 14, batch 2500, loss[loss=0.1878, simple_loss=0.2808, pruned_loss=0.0474, over 18106.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2649, pruned_loss=0.04554, over 3575342.09 frames. ], batch size: 62, lr: 8.06e-03, grad_scale: 8.0 +2023-03-09 06:02:01,768 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.909e+02 2.668e+02 3.143e+02 3.903e+02 7.073e+02, threshold=6.286e+02, percent-clipped=0.0 +2023-03-09 06:02:13,459 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=49766.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:02:27,056 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49778.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:02:34,122 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-09 06:02:39,140 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=49788.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 06:02:44,647 INFO [train.py:898] (0/4) Epoch 14, batch 2550, loss[loss=0.1795, simple_loss=0.2731, pruned_loss=0.04289, over 18353.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2649, pruned_loss=0.04555, over 3583672.85 frames. ], batch size: 55, lr: 8.06e-03, grad_scale: 8.0 +2023-03-09 06:03:24,804 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=49827.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:03:44,070 INFO [train.py:898] (0/4) Epoch 14, batch 2600, loss[loss=0.1713, simple_loss=0.26, pruned_loss=0.04126, over 15886.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2636, pruned_loss=0.04499, over 3593148.61 frames. ], batch size: 94, lr: 8.05e-03, grad_scale: 8.0 +2023-03-09 06:03:59,923 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.176e+02 2.916e+02 3.456e+02 4.257e+02 7.547e+02, threshold=6.912e+02, percent-clipped=3.0 +2023-03-09 06:04:42,284 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=49892.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:04:43,077 INFO [train.py:898] (0/4) Epoch 14, batch 2650, loss[loss=0.201, simple_loss=0.2878, pruned_loss=0.05709, over 18213.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2634, pruned_loss=0.04501, over 3592199.63 frames. ], batch size: 60, lr: 8.05e-03, grad_scale: 8.0 +2023-03-09 06:05:38,325 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=49940.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:05:42,196 INFO [train.py:898] (0/4) Epoch 14, batch 2700, loss[loss=0.1563, simple_loss=0.2341, pruned_loss=0.03925, over 18378.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2624, pruned_loss=0.04459, over 3603459.12 frames. ], batch size: 42, lr: 8.05e-03, grad_scale: 8.0 +2023-03-09 06:05:57,466 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.140e+02 2.972e+02 3.465e+02 4.147e+02 6.859e+02, threshold=6.931e+02, percent-clipped=0.0 +2023-03-09 06:06:31,553 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8913, 2.7577, 2.7526, 2.4603, 2.7418, 2.1371, 2.3323, 2.7686], + device='cuda:0'), covar=tensor([0.0079, 0.0121, 0.0099, 0.0134, 0.0111, 0.0230, 0.0216, 0.0091], + device='cuda:0'), in_proj_covar=tensor([0.0115, 0.0137, 0.0118, 0.0169, 0.0121, 0.0166, 0.0167, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 06:06:40,862 INFO [train.py:898] (0/4) Epoch 14, batch 2750, loss[loss=0.171, simple_loss=0.264, pruned_loss=0.039, over 18583.00 frames. ], tot_loss[loss=0.1756, simple_loss=0.2621, pruned_loss=0.04459, over 3606781.04 frames. ], batch size: 54, lr: 8.04e-03, grad_scale: 8.0 +2023-03-09 06:06:49,491 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-50000.pt +2023-03-09 06:07:43,616 INFO [train.py:898] (0/4) Epoch 14, batch 2800, loss[loss=0.1744, simple_loss=0.2646, pruned_loss=0.04211, over 18244.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2628, pruned_loss=0.04491, over 3600370.56 frames. ], batch size: 57, lr: 8.04e-03, grad_scale: 16.0 +2023-03-09 06:07:58,951 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.149e+02 3.000e+02 3.519e+02 4.137e+02 9.656e+02, threshold=7.037e+02, percent-clipped=4.0 +2023-03-09 06:08:13,723 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.55 vs. limit=5.0 +2023-03-09 06:08:25,620 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50078.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:08:37,100 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50088.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 06:08:42,957 INFO [train.py:898] (0/4) Epoch 14, batch 2850, loss[loss=0.1732, simple_loss=0.2608, pruned_loss=0.0428, over 18271.00 frames. ], tot_loss[loss=0.1772, simple_loss=0.2634, pruned_loss=0.04549, over 3575210.84 frames. ], batch size: 47, lr: 8.03e-03, grad_scale: 16.0 +2023-03-09 06:09:17,616 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50122.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:09:22,115 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=50126.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:09:24,545 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50128.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:09:33,407 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=50136.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 06:09:40,915 INFO [train.py:898] (0/4) Epoch 14, batch 2900, loss[loss=0.1525, simple_loss=0.2349, pruned_loss=0.03507, over 18386.00 frames. ], tot_loss[loss=0.1768, simple_loss=0.2636, pruned_loss=0.04505, over 3587178.77 frames. ], batch size: 46, lr: 8.03e-03, grad_scale: 16.0 +2023-03-09 06:09:52,609 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-09 06:09:56,510 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.084e+02 3.147e+02 3.827e+02 4.513e+02 7.864e+02, threshold=7.653e+02, percent-clipped=1.0 +2023-03-09 06:10:34,446 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50189.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:10:38,564 INFO [train.py:898] (0/4) Epoch 14, batch 2950, loss[loss=0.1818, simple_loss=0.2747, pruned_loss=0.04446, over 18319.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2646, pruned_loss=0.04542, over 3586070.55 frames. ], batch size: 54, lr: 8.03e-03, grad_scale: 16.0 +2023-03-09 06:10:43,220 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50197.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 06:10:55,756 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.6326, 6.1181, 5.5557, 5.9203, 5.7180, 5.5871, 6.1981, 6.0710], + device='cuda:0'), covar=tensor([0.1100, 0.0684, 0.0450, 0.0687, 0.1228, 0.0716, 0.0498, 0.0689], + device='cuda:0'), in_proj_covar=tensor([0.0554, 0.0472, 0.0348, 0.0504, 0.0684, 0.0502, 0.0664, 0.0499], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 06:11:37,782 INFO [train.py:898] (0/4) Epoch 14, batch 3000, loss[loss=0.1529, simple_loss=0.242, pruned_loss=0.03185, over 18410.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.2648, pruned_loss=0.0457, over 3572097.69 frames. ], batch size: 48, lr: 8.02e-03, grad_scale: 16.0 +2023-03-09 06:11:37,784 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 06:11:49,728 INFO [train.py:932] (0/4) Epoch 14, validation: loss=0.1532, simple_loss=0.2546, pruned_loss=0.02587, over 944034.00 frames. +2023-03-09 06:11:49,729 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 06:12:04,122 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.273e+02 3.285e+02 3.966e+02 4.720e+02 1.017e+03, threshold=7.933e+02, percent-clipped=5.0 +2023-03-09 06:12:04,905 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-09 06:12:07,454 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50258.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 06:12:16,459 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9731, 5.0348, 5.1101, 4.8563, 4.7971, 4.8697, 5.1909, 5.1884], + device='cuda:0'), covar=tensor([0.0064, 0.0069, 0.0059, 0.0095, 0.0073, 0.0121, 0.0068, 0.0105], + device='cuda:0'), in_proj_covar=tensor([0.0087, 0.0062, 0.0065, 0.0084, 0.0069, 0.0094, 0.0079, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-09 06:12:43,860 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-09 06:12:46,941 INFO [train.py:898] (0/4) Epoch 14, batch 3050, loss[loss=0.1729, simple_loss=0.2556, pruned_loss=0.04506, over 18280.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2644, pruned_loss=0.04553, over 3583881.23 frames. ], batch size: 47, lr: 8.02e-03, grad_scale: 16.0 +2023-03-09 06:13:29,841 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-09 06:13:33,136 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.59 vs. limit=2.0 +2023-03-09 06:13:45,660 INFO [train.py:898] (0/4) Epoch 14, batch 3100, loss[loss=0.175, simple_loss=0.2649, pruned_loss=0.04252, over 18002.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2647, pruned_loss=0.04563, over 3581316.15 frames. ], batch size: 65, lr: 8.01e-03, grad_scale: 16.0 +2023-03-09 06:14:00,590 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.982e+02 2.722e+02 3.409e+02 4.335e+02 9.164e+02, threshold=6.818e+02, percent-clipped=2.0 +2023-03-09 06:14:04,683 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-09 06:14:42,799 INFO [train.py:898] (0/4) Epoch 14, batch 3150, loss[loss=0.1814, simple_loss=0.2744, pruned_loss=0.04419, over 18577.00 frames. ], tot_loss[loss=0.179, simple_loss=0.2654, pruned_loss=0.04628, over 3573381.64 frames. ], batch size: 54, lr: 8.01e-03, grad_scale: 16.0 +2023-03-09 06:15:18,660 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50422.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:15:42,274 INFO [train.py:898] (0/4) Epoch 14, batch 3200, loss[loss=0.1874, simple_loss=0.2784, pruned_loss=0.04822, over 18120.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2645, pruned_loss=0.04573, over 3588471.07 frames. ], batch size: 62, lr: 8.01e-03, grad_scale: 16.0 +2023-03-09 06:15:58,904 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.049e+02 3.101e+02 3.675e+02 4.644e+02 1.158e+03, threshold=7.350e+02, percent-clipped=4.0 +2023-03-09 06:16:14,118 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=50470.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:16:30,661 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50484.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:16:40,180 INFO [train.py:898] (0/4) Epoch 14, batch 3250, loss[loss=0.1718, simple_loss=0.2522, pruned_loss=0.04574, over 18372.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.264, pruned_loss=0.04537, over 3599220.32 frames. ], batch size: 46, lr: 8.00e-03, grad_scale: 8.0 +2023-03-09 06:17:39,140 INFO [train.py:898] (0/4) Epoch 14, batch 3300, loss[loss=0.1687, simple_loss=0.2576, pruned_loss=0.03985, over 18297.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.263, pruned_loss=0.04506, over 3592705.73 frames. ], batch size: 49, lr: 8.00e-03, grad_scale: 8.0 +2023-03-09 06:17:42,605 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50546.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:17:51,055 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50553.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 06:17:55,808 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.637e+02 3.010e+02 3.418e+02 4.071e+02 6.649e+02, threshold=6.837e+02, percent-clipped=0.0 +2023-03-09 06:18:37,883 INFO [train.py:898] (0/4) Epoch 14, batch 3350, loss[loss=0.1973, simple_loss=0.2896, pruned_loss=0.05254, over 18619.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2627, pruned_loss=0.04492, over 3586733.30 frames. ], batch size: 52, lr: 8.00e-03, grad_scale: 8.0 +2023-03-09 06:18:55,353 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50607.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:19:36,705 INFO [train.py:898] (0/4) Epoch 14, batch 3400, loss[loss=0.1548, simple_loss=0.2422, pruned_loss=0.03369, over 18365.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2622, pruned_loss=0.04462, over 3588586.84 frames. ], batch size: 46, lr: 7.99e-03, grad_scale: 8.0 +2023-03-09 06:19:53,311 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.118e+02 2.885e+02 3.458e+02 4.304e+02 7.306e+02, threshold=6.916e+02, percent-clipped=1.0 +2023-03-09 06:20:35,019 INFO [train.py:898] (0/4) Epoch 14, batch 3450, loss[loss=0.1698, simple_loss=0.2631, pruned_loss=0.03825, over 18519.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2631, pruned_loss=0.04479, over 3584949.04 frames. ], batch size: 51, lr: 7.99e-03, grad_scale: 8.0 +2023-03-09 06:21:12,323 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6276, 5.1234, 5.0962, 5.1512, 4.5974, 4.9532, 4.4568, 4.9805], + device='cuda:0'), covar=tensor([0.0258, 0.0290, 0.0230, 0.0366, 0.0381, 0.0269, 0.1105, 0.0322], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0236, 0.0231, 0.0279, 0.0242, 0.0241, 0.0293, 0.0230], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0005], + device='cuda:0') +2023-03-09 06:21:19,450 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.14 vs. limit=5.0 +2023-03-09 06:21:21,538 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7281, 3.8822, 5.2594, 4.6812, 3.3725, 3.1654, 4.6143, 5.4839], + device='cuda:0'), covar=tensor([0.0838, 0.1511, 0.0124, 0.0284, 0.0880, 0.1034, 0.0344, 0.0212], + device='cuda:0'), in_proj_covar=tensor([0.0139, 0.0256, 0.0122, 0.0170, 0.0181, 0.0182, 0.0183, 0.0166], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 06:21:32,283 INFO [train.py:898] (0/4) Epoch 14, batch 3500, loss[loss=0.1782, simple_loss=0.2689, pruned_loss=0.04375, over 18273.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.2644, pruned_loss=0.04521, over 3581925.73 frames. ], batch size: 57, lr: 7.98e-03, grad_scale: 8.0 +2023-03-09 06:21:47,960 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.084e+02 3.008e+02 3.421e+02 4.567e+02 7.789e+02, threshold=6.843e+02, percent-clipped=2.0 +2023-03-09 06:22:15,552 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50782.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:22:17,504 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50784.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:22:26,426 INFO [train.py:898] (0/4) Epoch 14, batch 3550, loss[loss=0.1814, simple_loss=0.2754, pruned_loss=0.04366, over 18582.00 frames. ], tot_loss[loss=0.1777, simple_loss=0.2648, pruned_loss=0.04532, over 3578142.99 frames. ], batch size: 54, lr: 7.98e-03, grad_scale: 8.0 +2023-03-09 06:22:43,138 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50808.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 06:23:08,048 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=50832.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:23:09,397 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.3995, 3.2427, 4.4428, 3.8831, 2.9450, 2.7687, 3.9132, 4.5751], + device='cuda:0'), covar=tensor([0.0838, 0.1570, 0.0189, 0.0391, 0.1007, 0.1199, 0.0425, 0.0232], + device='cuda:0'), in_proj_covar=tensor([0.0140, 0.0259, 0.0123, 0.0172, 0.0184, 0.0184, 0.0185, 0.0168], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 06:23:19,507 INFO [train.py:898] (0/4) Epoch 14, batch 3600, loss[loss=0.1943, simple_loss=0.2844, pruned_loss=0.05215, over 18110.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2647, pruned_loss=0.04517, over 3587870.80 frames. ], batch size: 62, lr: 7.98e-03, grad_scale: 8.0 +2023-03-09 06:23:19,786 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50843.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:23:30,333 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=50853.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 06:23:34,501 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.994e+02 3.100e+02 3.552e+02 4.596e+02 8.414e+02, threshold=7.104e+02, percent-clipped=7.0 +2023-03-09 06:23:47,641 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=50869.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 06:23:51,498 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.4467, 2.9612, 3.7830, 3.5436, 2.7787, 2.7874, 3.4819, 3.9467], + device='cuda:0'), covar=tensor([0.0780, 0.1380, 0.0281, 0.0401, 0.0923, 0.1001, 0.0489, 0.0326], + device='cuda:0'), in_proj_covar=tensor([0.0139, 0.0257, 0.0122, 0.0170, 0.0182, 0.0183, 0.0183, 0.0167], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 06:23:55,300 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-14.pt +2023-03-09 06:24:25,704 INFO [train.py:898] (0/4) Epoch 15, batch 0, loss[loss=0.1849, simple_loss=0.2755, pruned_loss=0.04715, over 17649.00 frames. ], tot_loss[loss=0.1849, simple_loss=0.2755, pruned_loss=0.04715, over 17649.00 frames. ], batch size: 70, lr: 7.70e-03, grad_scale: 8.0 +2023-03-09 06:24:25,706 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 06:24:32,819 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1902, 4.2843, 3.1904, 4.8666, 3.8657, 4.7543, 3.4579, 4.7451], + device='cuda:0'), covar=tensor([0.0452, 0.0560, 0.1055, 0.0385, 0.0613, 0.0244, 0.0821, 0.0242], + device='cuda:0'), in_proj_covar=tensor([0.0202, 0.0216, 0.0181, 0.0258, 0.0184, 0.0252, 0.0194, 0.0192], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 06:24:37,512 INFO [train.py:932] (0/4) Epoch 15, validation: loss=0.1543, simple_loss=0.2557, pruned_loss=0.02649, over 944034.00 frames. +2023-03-09 06:24:37,513 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 06:25:06,589 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=50901.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 06:25:07,607 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=50902.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:25:35,410 INFO [train.py:898] (0/4) Epoch 15, batch 50, loss[loss=0.1554, simple_loss=0.2439, pruned_loss=0.03345, over 18371.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2642, pruned_loss=0.04517, over 815196.11 frames. ], batch size: 46, lr: 7.70e-03, grad_scale: 8.0 +2023-03-09 06:25:58,099 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=50946.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:26:11,146 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.078e+02 2.874e+02 3.277e+02 4.282e+02 1.387e+03, threshold=6.554e+02, percent-clipped=5.0 +2023-03-09 06:26:33,992 INFO [train.py:898] (0/4) Epoch 15, batch 100, loss[loss=0.1992, simple_loss=0.2885, pruned_loss=0.05497, over 17034.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2627, pruned_loss=0.04477, over 1423069.26 frames. ], batch size: 78, lr: 7.69e-03, grad_scale: 8.0 +2023-03-09 06:27:10,672 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51007.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:27:11,859 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5191, 2.7234, 2.4495, 2.8257, 3.6054, 3.5726, 3.0600, 2.8784], + device='cuda:0'), covar=tensor([0.0179, 0.0256, 0.0504, 0.0343, 0.0154, 0.0121, 0.0327, 0.0353], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0121, 0.0155, 0.0147, 0.0112, 0.0099, 0.0139, 0.0139], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 06:27:33,101 INFO [train.py:898] (0/4) Epoch 15, batch 150, loss[loss=0.1771, simple_loss=0.2707, pruned_loss=0.04177, over 18587.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2619, pruned_loss=0.0445, over 1906127.09 frames. ], batch size: 54, lr: 7.69e-03, grad_scale: 8.0 +2023-03-09 06:27:49,551 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.75 vs. limit=5.0 +2023-03-09 06:28:09,467 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.782e+02 2.876e+02 3.303e+02 4.141e+02 9.283e+02, threshold=6.606e+02, percent-clipped=4.0 +2023-03-09 06:28:15,454 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51062.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:28:31,739 INFO [train.py:898] (0/4) Epoch 15, batch 200, loss[loss=0.1694, simple_loss=0.2488, pruned_loss=0.04499, over 18434.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2617, pruned_loss=0.04432, over 2286072.29 frames. ], batch size: 43, lr: 7.69e-03, grad_scale: 8.0 +2023-03-09 06:28:48,169 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8658, 3.0525, 4.3983, 3.9187, 2.6594, 4.7084, 4.0783, 2.9708], + device='cuda:0'), covar=tensor([0.0462, 0.1467, 0.0202, 0.0361, 0.1730, 0.0198, 0.0460, 0.1118], + device='cuda:0'), in_proj_covar=tensor([0.0202, 0.0233, 0.0179, 0.0151, 0.0219, 0.0196, 0.0224, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 06:29:25,536 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51123.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:29:29,637 INFO [train.py:898] (0/4) Epoch 15, batch 250, loss[loss=0.1786, simple_loss=0.2715, pruned_loss=0.04282, over 18491.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2633, pruned_loss=0.04509, over 2567993.28 frames. ], batch size: 53, lr: 7.68e-03, grad_scale: 8.0 +2023-03-09 06:29:41,452 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51138.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:30:03,163 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.733e+02 3.055e+02 3.759e+02 4.768e+02 8.337e+02, threshold=7.517e+02, percent-clipped=9.0 +2023-03-09 06:30:12,396 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51164.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 06:30:21,590 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51172.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 06:30:27,048 INFO [train.py:898] (0/4) Epoch 15, batch 300, loss[loss=0.2344, simple_loss=0.3042, pruned_loss=0.08225, over 12530.00 frames. ], tot_loss[loss=0.1774, simple_loss=0.264, pruned_loss=0.04541, over 2792181.97 frames. ], batch size: 130, lr: 7.68e-03, grad_scale: 8.0 +2023-03-09 06:30:56,731 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51202.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:31:26,182 INFO [train.py:898] (0/4) Epoch 15, batch 350, loss[loss=0.225, simple_loss=0.3, pruned_loss=0.07498, over 12769.00 frames. ], tot_loss[loss=0.178, simple_loss=0.2647, pruned_loss=0.0456, over 2958277.06 frames. ], batch size: 130, lr: 7.67e-03, grad_scale: 8.0 +2023-03-09 06:31:33,438 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51233.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 06:31:53,021 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=51250.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:32:01,198 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.843e+02 2.755e+02 3.363e+02 4.160e+02 1.249e+03, threshold=6.726e+02, percent-clipped=1.0 +2023-03-09 06:32:25,070 INFO [train.py:898] (0/4) Epoch 15, batch 400, loss[loss=0.1607, simple_loss=0.2388, pruned_loss=0.0413, over 18482.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2628, pruned_loss=0.04443, over 3110447.88 frames. ], batch size: 43, lr: 7.67e-03, grad_scale: 8.0 +2023-03-09 06:32:40,070 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8073, 5.2978, 4.9210, 5.0994, 4.9500, 4.8686, 5.3823, 5.2931], + device='cuda:0'), covar=tensor([0.1189, 0.0712, 0.1179, 0.0684, 0.1232, 0.0753, 0.0588, 0.0698], + device='cuda:0'), in_proj_covar=tensor([0.0565, 0.0473, 0.0351, 0.0507, 0.0689, 0.0504, 0.0678, 0.0503], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 06:32:53,507 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51302.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:33:24,231 INFO [train.py:898] (0/4) Epoch 15, batch 450, loss[loss=0.1753, simple_loss=0.2702, pruned_loss=0.04021, over 18359.00 frames. ], tot_loss[loss=0.1759, simple_loss=0.2633, pruned_loss=0.04422, over 3218765.76 frames. ], batch size: 55, lr: 7.67e-03, grad_scale: 8.0 +2023-03-09 06:33:27,786 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8758, 4.6009, 4.6796, 3.4975, 3.8632, 3.6865, 2.6577, 2.5812], + device='cuda:0'), covar=tensor([0.0212, 0.0141, 0.0079, 0.0288, 0.0309, 0.0196, 0.0739, 0.0844], + device='cuda:0'), in_proj_covar=tensor([0.0065, 0.0053, 0.0056, 0.0065, 0.0086, 0.0061, 0.0075, 0.0083], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 06:33:59,192 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.915e+02 2.968e+02 3.356e+02 4.062e+02 8.839e+02, threshold=6.712e+02, percent-clipped=1.0 +2023-03-09 06:34:23,279 INFO [train.py:898] (0/4) Epoch 15, batch 500, loss[loss=0.1963, simple_loss=0.2793, pruned_loss=0.05667, over 18294.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2633, pruned_loss=0.0445, over 3298910.13 frames. ], batch size: 57, lr: 7.66e-03, grad_scale: 8.0 +2023-03-09 06:34:28,969 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51382.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:34:39,371 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0387, 4.2133, 2.5100, 4.2268, 5.1432, 2.7270, 3.9416, 3.9872], + device='cuda:0'), covar=tensor([0.0107, 0.0916, 0.1508, 0.0471, 0.0063, 0.1116, 0.0554, 0.0628], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0249, 0.0195, 0.0191, 0.0100, 0.0178, 0.0209, 0.0215], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 06:34:40,353 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2510, 5.1385, 5.4422, 5.4637, 5.1263, 5.9931, 5.6270, 5.2960], + device='cuda:0'), covar=tensor([0.1083, 0.0657, 0.0736, 0.0661, 0.1499, 0.0774, 0.0597, 0.1680], + device='cuda:0'), in_proj_covar=tensor([0.0339, 0.0262, 0.0280, 0.0279, 0.0316, 0.0392, 0.0256, 0.0376], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0004, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 06:34:40,412 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51392.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:35:10,683 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51418.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:35:20,790 INFO [train.py:898] (0/4) Epoch 15, batch 550, loss[loss=0.17, simple_loss=0.2545, pruned_loss=0.0427, over 18288.00 frames. ], tot_loss[loss=0.176, simple_loss=0.263, pruned_loss=0.04453, over 3364754.54 frames. ], batch size: 49, lr: 7.66e-03, grad_scale: 8.0 +2023-03-09 06:35:33,774 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51438.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:35:39,485 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51443.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:35:44,139 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.98 vs. limit=2.0 +2023-03-09 06:35:50,514 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51453.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:35:54,565 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.873e+02 3.112e+02 3.809e+02 4.675e+02 7.627e+02, threshold=7.618e+02, percent-clipped=1.0 +2023-03-09 06:35:57,258 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1673, 5.4905, 2.9909, 5.3479, 5.2737, 5.5643, 5.3213, 2.8108], + device='cuda:0'), covar=tensor([0.0175, 0.0067, 0.0694, 0.0062, 0.0062, 0.0062, 0.0090, 0.0924], + device='cuda:0'), in_proj_covar=tensor([0.0081, 0.0072, 0.0090, 0.0087, 0.0080, 0.0068, 0.0079, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 06:36:03,336 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51464.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 06:36:04,323 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2395, 5.1255, 5.4494, 5.4201, 5.0486, 5.9652, 5.5892, 5.2433], + device='cuda:0'), covar=tensor([0.1068, 0.0679, 0.0694, 0.0689, 0.1426, 0.0759, 0.0671, 0.1638], + device='cuda:0'), in_proj_covar=tensor([0.0339, 0.0262, 0.0279, 0.0280, 0.0316, 0.0392, 0.0255, 0.0378], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0004, 0.0002, 0.0003], + device='cuda:0') +2023-03-09 06:36:12,216 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.97 vs. limit=2.0 +2023-03-09 06:36:18,185 INFO [train.py:898] (0/4) Epoch 15, batch 600, loss[loss=0.1989, simple_loss=0.282, pruned_loss=0.05788, over 17015.00 frames. ], tot_loss[loss=0.1766, simple_loss=0.2637, pruned_loss=0.04478, over 3419057.37 frames. ], batch size: 78, lr: 7.66e-03, grad_scale: 8.0 +2023-03-09 06:36:29,361 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=51486.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:36:36,336 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0893, 5.3678, 2.8225, 5.2148, 5.1375, 5.4237, 5.2670, 2.7982], + device='cuda:0'), covar=tensor([0.0170, 0.0056, 0.0710, 0.0072, 0.0064, 0.0058, 0.0067, 0.0896], + device='cuda:0'), in_proj_covar=tensor([0.0081, 0.0072, 0.0090, 0.0087, 0.0080, 0.0068, 0.0079, 0.0093], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 06:36:39,696 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51495.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:36:41,152 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-09 06:36:58,929 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=51512.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 06:37:16,748 INFO [train.py:898] (0/4) Epoch 15, batch 650, loss[loss=0.1671, simple_loss=0.2511, pruned_loss=0.04161, over 18401.00 frames. ], tot_loss[loss=0.1773, simple_loss=0.2644, pruned_loss=0.04509, over 3450601.11 frames. ], batch size: 48, lr: 7.65e-03, grad_scale: 8.0 +2023-03-09 06:37:18,066 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51528.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 06:37:51,101 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51556.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:37:51,421 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-09 06:37:51,901 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.157e+02 2.845e+02 3.389e+02 4.059e+02 1.145e+03, threshold=6.778e+02, percent-clipped=5.0 +2023-03-09 06:38:14,977 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7041, 4.4121, 4.4804, 3.3287, 3.7382, 3.5378, 2.8231, 2.5717], + device='cuda:0'), covar=tensor([0.0260, 0.0181, 0.0087, 0.0303, 0.0336, 0.0207, 0.0643, 0.0850], + device='cuda:0'), in_proj_covar=tensor([0.0064, 0.0052, 0.0055, 0.0064, 0.0084, 0.0060, 0.0074, 0.0082], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 06:38:15,829 INFO [train.py:898] (0/4) Epoch 15, batch 700, loss[loss=0.1629, simple_loss=0.2544, pruned_loss=0.03572, over 18573.00 frames. ], tot_loss[loss=0.177, simple_loss=0.2641, pruned_loss=0.04493, over 3487246.01 frames. ], batch size: 54, lr: 7.65e-03, grad_scale: 8.0 +2023-03-09 06:38:46,160 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51602.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:39:05,256 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7863, 3.6914, 3.5397, 3.0832, 3.5397, 2.7837, 2.8098, 3.6484], + device='cuda:0'), covar=tensor([0.0043, 0.0073, 0.0071, 0.0122, 0.0070, 0.0171, 0.0166, 0.0060], + device='cuda:0'), in_proj_covar=tensor([0.0118, 0.0139, 0.0120, 0.0174, 0.0124, 0.0167, 0.0170, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 06:39:14,427 INFO [train.py:898] (0/4) Epoch 15, batch 750, loss[loss=0.1774, simple_loss=0.2552, pruned_loss=0.0498, over 18510.00 frames. ], tot_loss[loss=0.1775, simple_loss=0.2645, pruned_loss=0.04524, over 3511582.51 frames. ], batch size: 44, lr: 7.65e-03, grad_scale: 8.0 +2023-03-09 06:39:15,155 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.00 vs. limit=2.0 +2023-03-09 06:39:21,311 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.82 vs. limit=5.0 +2023-03-09 06:39:41,992 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=51650.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:39:49,770 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.857e+02 2.830e+02 3.241e+02 3.850e+02 9.423e+02, threshold=6.481e+02, percent-clipped=1.0 +2023-03-09 06:40:12,982 INFO [train.py:898] (0/4) Epoch 15, batch 800, loss[loss=0.1502, simple_loss=0.228, pruned_loss=0.03618, over 18432.00 frames. ], tot_loss[loss=0.1767, simple_loss=0.2637, pruned_loss=0.04482, over 3529858.85 frames. ], batch size: 43, lr: 7.64e-03, grad_scale: 8.0 +2023-03-09 06:41:01,351 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51718.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:41:11,136 INFO [train.py:898] (0/4) Epoch 15, batch 850, loss[loss=0.1558, simple_loss=0.2346, pruned_loss=0.03855, over 18432.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2619, pruned_loss=0.0442, over 3549101.88 frames. ], batch size: 43, lr: 7.64e-03, grad_scale: 8.0 +2023-03-09 06:41:24,443 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51738.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:41:35,605 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=51747.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:41:36,543 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51748.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:41:47,044 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.547e+02 2.965e+02 3.491e+02 4.119e+02 9.035e+02, threshold=6.982e+02, percent-clipped=4.0 +2023-03-09 06:41:50,631 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1105, 5.5000, 2.9170, 5.3628, 5.2496, 5.5758, 5.2912, 2.9409], + device='cuda:0'), covar=tensor([0.0168, 0.0065, 0.0721, 0.0061, 0.0063, 0.0056, 0.0091, 0.0826], + device='cuda:0'), in_proj_covar=tensor([0.0082, 0.0073, 0.0091, 0.0087, 0.0080, 0.0069, 0.0080, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 06:41:57,083 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=51766.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:42:09,439 INFO [train.py:898] (0/4) Epoch 15, batch 900, loss[loss=0.197, simple_loss=0.2859, pruned_loss=0.05407, over 18352.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2613, pruned_loss=0.04397, over 3561478.23 frames. ], batch size: 55, lr: 7.63e-03, grad_scale: 8.0 +2023-03-09 06:42:39,094 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-09 06:42:47,067 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=51808.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:43:08,028 INFO [train.py:898] (0/4) Epoch 15, batch 950, loss[loss=0.1789, simple_loss=0.2649, pruned_loss=0.04642, over 18094.00 frames. ], tot_loss[loss=0.1744, simple_loss=0.2609, pruned_loss=0.04396, over 3568699.38 frames. ], batch size: 40, lr: 7.63e-03, grad_scale: 8.0 +2023-03-09 06:43:09,549 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=51828.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 06:43:37,185 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=51851.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:43:43,806 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.174e+02 2.846e+02 3.316e+02 4.013e+02 9.556e+02, threshold=6.632e+02, percent-clipped=3.0 +2023-03-09 06:44:06,101 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=51876.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 06:44:06,907 INFO [train.py:898] (0/4) Epoch 15, batch 1000, loss[loss=0.1817, simple_loss=0.2705, pruned_loss=0.04645, over 18413.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2609, pruned_loss=0.04359, over 3583786.52 frames. ], batch size: 48, lr: 7.63e-03, grad_scale: 8.0 +2023-03-09 06:45:05,511 INFO [train.py:898] (0/4) Epoch 15, batch 1050, loss[loss=0.1652, simple_loss=0.2487, pruned_loss=0.04091, over 18147.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2616, pruned_loss=0.04404, over 3578928.89 frames. ], batch size: 44, lr: 7.62e-03, grad_scale: 8.0 +2023-03-09 06:45:30,553 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6746, 3.5963, 3.3911, 2.9357, 3.4031, 2.5943, 2.4942, 3.7087], + device='cuda:0'), covar=tensor([0.0053, 0.0085, 0.0085, 0.0168, 0.0090, 0.0217, 0.0222, 0.0052], + device='cuda:0'), in_proj_covar=tensor([0.0118, 0.0141, 0.0120, 0.0175, 0.0125, 0.0169, 0.0174, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 06:45:37,265 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6190, 2.2453, 2.7784, 2.7205, 3.1360, 5.0326, 4.8067, 3.8505], + device='cuda:0'), covar=tensor([0.1515, 0.2166, 0.2595, 0.1543, 0.2229, 0.0152, 0.0329, 0.0615], + device='cuda:0'), in_proj_covar=tensor([0.0265, 0.0321, 0.0347, 0.0259, 0.0372, 0.0209, 0.0276, 0.0227], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 06:45:39,862 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.929e+02 2.891e+02 3.197e+02 3.857e+02 9.406e+02, threshold=6.393e+02, percent-clipped=2.0 +2023-03-09 06:46:03,749 INFO [train.py:898] (0/4) Epoch 15, batch 1100, loss[loss=0.1572, simple_loss=0.2466, pruned_loss=0.03388, over 18267.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2618, pruned_loss=0.04408, over 3580088.05 frames. ], batch size: 47, lr: 7.62e-03, grad_scale: 8.0 +2023-03-09 06:46:31,009 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-52000.pt +2023-03-09 06:47:06,139 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7236, 2.9072, 2.7715, 2.9728, 3.7954, 3.6749, 3.1676, 3.0599], + device='cuda:0'), covar=tensor([0.0166, 0.0281, 0.0425, 0.0388, 0.0127, 0.0149, 0.0340, 0.0339], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0122, 0.0155, 0.0149, 0.0113, 0.0101, 0.0141, 0.0142], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 06:47:06,916 INFO [train.py:898] (0/4) Epoch 15, batch 1150, loss[loss=0.1529, simple_loss=0.2381, pruned_loss=0.03381, over 18555.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.2616, pruned_loss=0.04391, over 3587397.44 frames. ], batch size: 45, lr: 7.62e-03, grad_scale: 8.0 +2023-03-09 06:47:19,871 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52038.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:47:22,255 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52040.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:47:31,718 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52048.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:47:42,030 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.912e+02 2.945e+02 3.374e+02 4.146e+02 9.750e+02, threshold=6.749e+02, percent-clipped=3.0 +2023-03-09 06:48:00,261 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6599, 3.5736, 2.3766, 4.5253, 3.0549, 4.4695, 2.5042, 4.2483], + device='cuda:0'), covar=tensor([0.0579, 0.0727, 0.1275, 0.0387, 0.0794, 0.0241, 0.1148, 0.0305], + device='cuda:0'), in_proj_covar=tensor([0.0204, 0.0218, 0.0182, 0.0264, 0.0187, 0.0255, 0.0197, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 06:48:05,575 INFO [train.py:898] (0/4) Epoch 15, batch 1200, loss[loss=0.2002, simple_loss=0.2857, pruned_loss=0.0573, over 18440.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2617, pruned_loss=0.04394, over 3591159.16 frames. ], batch size: 59, lr: 7.61e-03, grad_scale: 8.0 +2023-03-09 06:48:15,901 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=52086.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:48:27,181 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=52096.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:48:33,634 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52101.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:48:35,593 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52103.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:49:03,938 INFO [train.py:898] (0/4) Epoch 15, batch 1250, loss[loss=0.189, simple_loss=0.2793, pruned_loss=0.04937, over 17702.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2623, pruned_loss=0.04417, over 3588764.60 frames. ], batch size: 70, lr: 7.61e-03, grad_scale: 8.0 +2023-03-09 06:49:16,857 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2620, 3.1963, 2.0538, 4.0064, 2.7365, 3.8459, 2.2581, 3.6612], + device='cuda:0'), covar=tensor([0.0609, 0.0850, 0.1403, 0.0487, 0.0857, 0.0327, 0.1189, 0.0363], + device='cuda:0'), in_proj_covar=tensor([0.0203, 0.0218, 0.0182, 0.0263, 0.0186, 0.0255, 0.0197, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 06:49:31,456 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52151.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:49:38,915 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.755e+02 2.743e+02 3.530e+02 4.492e+02 1.221e+03, threshold=7.059e+02, percent-clipped=4.0 +2023-03-09 06:50:02,589 INFO [train.py:898] (0/4) Epoch 15, batch 1300, loss[loss=0.1751, simple_loss=0.2682, pruned_loss=0.04101, over 18486.00 frames. ], tot_loss[loss=0.1751, simple_loss=0.2621, pruned_loss=0.04407, over 3595449.83 frames. ], batch size: 51, lr: 7.61e-03, grad_scale: 8.0 +2023-03-09 06:50:21,047 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52193.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:50:27,602 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=52199.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:50:45,899 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0549, 5.1277, 5.1369, 4.9691, 4.8889, 4.8723, 5.2981, 5.2769], + device='cuda:0'), covar=tensor([0.0066, 0.0063, 0.0057, 0.0088, 0.0063, 0.0124, 0.0062, 0.0100], + device='cuda:0'), in_proj_covar=tensor([0.0087, 0.0062, 0.0066, 0.0084, 0.0069, 0.0094, 0.0080, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-09 06:50:54,637 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4878, 3.1974, 3.1756, 2.7425, 3.1168, 2.4979, 2.4398, 3.2793], + device='cuda:0'), covar=tensor([0.0073, 0.0167, 0.0141, 0.0183, 0.0134, 0.0272, 0.0284, 0.0089], + device='cuda:0'), in_proj_covar=tensor([0.0120, 0.0142, 0.0121, 0.0176, 0.0126, 0.0169, 0.0176, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 06:51:00,950 INFO [train.py:898] (0/4) Epoch 15, batch 1350, loss[loss=0.1537, simple_loss=0.2414, pruned_loss=0.03298, over 18394.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2623, pruned_loss=0.04377, over 3606757.71 frames. ], batch size: 52, lr: 7.60e-03, grad_scale: 8.0 +2023-03-09 06:51:32,048 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5743, 6.0497, 5.5833, 5.8818, 5.6306, 5.5880, 6.1311, 6.0496], + device='cuda:0'), covar=tensor([0.1143, 0.0686, 0.0460, 0.0614, 0.1376, 0.0674, 0.0541, 0.0668], + device='cuda:0'), in_proj_covar=tensor([0.0566, 0.0474, 0.0355, 0.0504, 0.0688, 0.0500, 0.0674, 0.0505], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 06:51:32,163 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9768, 4.5250, 4.7142, 3.6063, 3.8491, 3.5380, 2.7534, 2.6783], + device='cuda:0'), covar=tensor([0.0179, 0.0178, 0.0073, 0.0260, 0.0306, 0.0229, 0.0684, 0.0786], + device='cuda:0'), in_proj_covar=tensor([0.0063, 0.0052, 0.0053, 0.0063, 0.0082, 0.0059, 0.0073, 0.0079], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0004, 0.0006, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 06:51:32,210 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52254.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:51:33,440 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0216, 3.8371, 5.1676, 2.8380, 4.5165, 2.7919, 3.1680, 1.8294], + device='cuda:0'), covar=tensor([0.0968, 0.0883, 0.0110, 0.0873, 0.0514, 0.2340, 0.2550, 0.2033], + device='cuda:0'), in_proj_covar=tensor([0.0207, 0.0226, 0.0143, 0.0182, 0.0239, 0.0256, 0.0300, 0.0220], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 06:51:35,202 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.856e+02 2.890e+02 3.384e+02 4.121e+02 8.952e+02, threshold=6.768e+02, percent-clipped=2.0 +2023-03-09 06:51:52,121 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0340, 5.2685, 2.5342, 5.1221, 4.9786, 5.2954, 5.0809, 2.6466], + device='cuda:0'), covar=tensor([0.0194, 0.0052, 0.0836, 0.0075, 0.0069, 0.0068, 0.0091, 0.0950], + device='cuda:0'), in_proj_covar=tensor([0.0081, 0.0072, 0.0091, 0.0086, 0.0081, 0.0069, 0.0080, 0.0093], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 06:51:59,006 INFO [train.py:898] (0/4) Epoch 15, batch 1400, loss[loss=0.1684, simple_loss=0.2438, pruned_loss=0.04647, over 17650.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2624, pruned_loss=0.04378, over 3605515.05 frames. ], batch size: 39, lr: 7.60e-03, grad_scale: 8.0 +2023-03-09 06:52:40,416 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52312.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:52:57,734 INFO [train.py:898] (0/4) Epoch 15, batch 1450, loss[loss=0.1892, simple_loss=0.2785, pruned_loss=0.05002, over 17955.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2637, pruned_loss=0.04432, over 3590925.89 frames. ], batch size: 65, lr: 7.59e-03, grad_scale: 8.0 +2023-03-09 06:53:31,866 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.869e+02 2.900e+02 3.474e+02 4.248e+02 1.297e+03, threshold=6.947e+02, percent-clipped=2.0 +2023-03-09 06:53:46,596 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5399, 2.2052, 2.4885, 2.8307, 3.0083, 5.1905, 4.9853, 3.8177], + device='cuda:0'), covar=tensor([0.1803, 0.3024, 0.3547, 0.1828, 0.3443, 0.0275, 0.0319, 0.0725], + device='cuda:0'), in_proj_covar=tensor([0.0266, 0.0322, 0.0348, 0.0261, 0.0373, 0.0210, 0.0278, 0.0227], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 06:53:51,158 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52373.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:53:55,302 INFO [train.py:898] (0/4) Epoch 15, batch 1500, loss[loss=0.1586, simple_loss=0.2446, pruned_loss=0.03628, over 18145.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2624, pruned_loss=0.04378, over 3597788.30 frames. ], batch size: 44, lr: 7.59e-03, grad_scale: 8.0 +2023-03-09 06:54:18,115 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52396.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:54:26,394 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52403.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:54:54,576 INFO [train.py:898] (0/4) Epoch 15, batch 1550, loss[loss=0.1938, simple_loss=0.2897, pruned_loss=0.04896, over 18104.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2622, pruned_loss=0.04378, over 3595732.92 frames. ], batch size: 62, lr: 7.59e-03, grad_scale: 8.0 +2023-03-09 06:54:57,805 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9403, 3.0912, 4.6063, 3.8879, 3.0348, 4.8403, 4.0187, 3.1418], + device='cuda:0'), covar=tensor([0.0420, 0.1400, 0.0210, 0.0400, 0.1393, 0.0153, 0.0536, 0.0910], + device='cuda:0'), in_proj_covar=tensor([0.0204, 0.0234, 0.0182, 0.0152, 0.0220, 0.0197, 0.0227, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 06:55:14,312 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1269, 2.5474, 2.2563, 2.6087, 3.3407, 3.1984, 2.7777, 2.7126], + device='cuda:0'), covar=tensor([0.0214, 0.0333, 0.0652, 0.0454, 0.0185, 0.0173, 0.0406, 0.0361], + device='cuda:0'), in_proj_covar=tensor([0.0128, 0.0123, 0.0158, 0.0149, 0.0113, 0.0102, 0.0142, 0.0143], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 06:55:23,297 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=52451.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:55:29,980 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.173e+02 2.872e+02 3.351e+02 3.820e+02 1.204e+03, threshold=6.703e+02, percent-clipped=1.0 +2023-03-09 06:55:54,079 INFO [train.py:898] (0/4) Epoch 15, batch 1600, loss[loss=0.1819, simple_loss=0.2758, pruned_loss=0.04402, over 18346.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.262, pruned_loss=0.04369, over 3597392.33 frames. ], batch size: 56, lr: 7.58e-03, grad_scale: 16.0 +2023-03-09 06:56:38,787 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6116, 2.0946, 2.6668, 2.6880, 3.3343, 5.0392, 4.7838, 3.6135], + device='cuda:0'), covar=tensor([0.1554, 0.2440, 0.2801, 0.1600, 0.1999, 0.0141, 0.0357, 0.0740], + device='cuda:0'), in_proj_covar=tensor([0.0267, 0.0323, 0.0349, 0.0261, 0.0374, 0.0212, 0.0277, 0.0228], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 06:56:52,738 INFO [train.py:898] (0/4) Epoch 15, batch 1650, loss[loss=0.1536, simple_loss=0.235, pruned_loss=0.03607, over 18356.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2617, pruned_loss=0.04332, over 3600434.07 frames. ], batch size: 46, lr: 7.58e-03, grad_scale: 16.0 +2023-03-09 06:57:09,491 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7730, 4.3523, 4.4378, 3.1847, 3.5874, 3.4431, 2.3805, 2.3225], + device='cuda:0'), covar=tensor([0.0204, 0.0125, 0.0074, 0.0343, 0.0320, 0.0205, 0.0733, 0.0835], + device='cuda:0'), in_proj_covar=tensor([0.0064, 0.0052, 0.0054, 0.0064, 0.0083, 0.0060, 0.0073, 0.0080], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0004, 0.0006, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 06:57:12,094 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.17 vs. limit=5.0 +2023-03-09 06:57:18,380 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52549.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:57:25,803 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52555.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 06:57:27,650 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.792e+02 2.929e+02 3.401e+02 4.256e+02 6.511e+02, threshold=6.802e+02, percent-clipped=0.0 +2023-03-09 06:57:50,950 INFO [train.py:898] (0/4) Epoch 15, batch 1700, loss[loss=0.1766, simple_loss=0.2667, pruned_loss=0.04329, over 18402.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2626, pruned_loss=0.04364, over 3598807.19 frames. ], batch size: 50, lr: 7.58e-03, grad_scale: 16.0 +2023-03-09 06:58:18,983 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-09 06:58:37,309 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52616.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 06:58:49,952 INFO [train.py:898] (0/4) Epoch 15, batch 1750, loss[loss=0.1815, simple_loss=0.2792, pruned_loss=0.04192, over 18284.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2629, pruned_loss=0.04352, over 3610879.55 frames. ], batch size: 49, lr: 7.57e-03, grad_scale: 8.0 +2023-03-09 06:59:04,167 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8810, 4.0995, 2.3851, 3.9823, 5.0025, 2.6864, 3.6263, 3.8744], + device='cuda:0'), covar=tensor([0.0112, 0.0997, 0.1647, 0.0543, 0.0062, 0.1143, 0.0672, 0.0619], + device='cuda:0'), in_proj_covar=tensor([0.0140, 0.0252, 0.0198, 0.0191, 0.0103, 0.0180, 0.0211, 0.0216], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 06:59:26,519 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.883e+02 2.725e+02 3.268e+02 4.051e+02 8.662e+02, threshold=6.535e+02, percent-clipped=2.0 +2023-03-09 06:59:30,534 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-09 06:59:37,918 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52668.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 06:59:48,455 INFO [train.py:898] (0/4) Epoch 15, batch 1800, loss[loss=0.1676, simple_loss=0.2613, pruned_loss=0.03693, over 18548.00 frames. ], tot_loss[loss=0.1743, simple_loss=0.262, pruned_loss=0.04326, over 3605537.34 frames. ], batch size: 49, lr: 7.57e-03, grad_scale: 8.0 +2023-03-09 07:00:11,301 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52696.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:00:46,761 INFO [train.py:898] (0/4) Epoch 15, batch 1850, loss[loss=0.1634, simple_loss=0.2486, pruned_loss=0.03905, over 18512.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.263, pruned_loss=0.04378, over 3598050.64 frames. ], batch size: 47, lr: 7.57e-03, grad_scale: 8.0 +2023-03-09 07:01:07,603 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=52744.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:01:17,306 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=52752.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:01:23,879 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.980e+02 2.718e+02 3.382e+02 4.103e+02 9.791e+02, threshold=6.764e+02, percent-clipped=3.0 +2023-03-09 07:01:46,237 INFO [train.py:898] (0/4) Epoch 15, batch 1900, loss[loss=0.1916, simple_loss=0.2849, pruned_loss=0.04922, over 18344.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2621, pruned_loss=0.04349, over 3600129.59 frames. ], batch size: 56, lr: 7.56e-03, grad_scale: 8.0 +2023-03-09 07:02:11,683 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-09 07:02:29,970 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=52813.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:02:45,701 INFO [train.py:898] (0/4) Epoch 15, batch 1950, loss[loss=0.2012, simple_loss=0.2882, pruned_loss=0.05711, over 18488.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2617, pruned_loss=0.04339, over 3587834.27 frames. ], batch size: 53, lr: 7.56e-03, grad_scale: 8.0 +2023-03-09 07:02:58,248 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-03-09 07:03:01,756 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-09 07:03:12,871 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52849.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:03:22,534 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.851e+02 2.886e+02 3.501e+02 4.238e+02 6.243e+02, threshold=7.003e+02, percent-clipped=0.0 +2023-03-09 07:03:44,569 INFO [train.py:898] (0/4) Epoch 15, batch 2000, loss[loss=0.1714, simple_loss=0.2643, pruned_loss=0.03925, over 17877.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.262, pruned_loss=0.04378, over 3583368.01 frames. ], batch size: 70, lr: 7.56e-03, grad_scale: 8.0 +2023-03-09 07:04:08,581 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=52897.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:04:25,090 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=52911.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 07:04:43,167 INFO [train.py:898] (0/4) Epoch 15, batch 2050, loss[loss=0.1619, simple_loss=0.254, pruned_loss=0.03483, over 18305.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2624, pruned_loss=0.04407, over 3579994.42 frames. ], batch size: 54, lr: 7.55e-03, grad_scale: 8.0 +2023-03-09 07:05:19,673 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.048e+02 3.109e+02 3.661e+02 4.518e+02 9.006e+02, threshold=7.322e+02, percent-clipped=1.0 +2023-03-09 07:05:30,885 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=52968.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:05:41,828 INFO [train.py:898] (0/4) Epoch 15, batch 2100, loss[loss=0.1728, simple_loss=0.2701, pruned_loss=0.03775, over 18349.00 frames. ], tot_loss[loss=0.1761, simple_loss=0.2632, pruned_loss=0.04447, over 3583681.11 frames. ], batch size: 55, lr: 7.55e-03, grad_scale: 8.0 +2023-03-09 07:05:59,801 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2278, 4.1762, 3.9565, 4.1562, 4.1384, 3.6978, 4.1669, 3.9251], + device='cuda:0'), covar=tensor([0.0491, 0.0680, 0.1329, 0.0735, 0.0618, 0.0467, 0.0475, 0.1089], + device='cuda:0'), in_proj_covar=tensor([0.0454, 0.0509, 0.0663, 0.0402, 0.0393, 0.0462, 0.0495, 0.0635], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 07:06:28,578 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=53016.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:06:41,337 INFO [train.py:898] (0/4) Epoch 15, batch 2150, loss[loss=0.1591, simple_loss=0.2458, pruned_loss=0.03619, over 18239.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.263, pruned_loss=0.04424, over 3572178.40 frames. ], batch size: 47, lr: 7.54e-03, grad_scale: 8.0 +2023-03-09 07:06:47,525 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3333, 5.2979, 4.8893, 5.2755, 5.2249, 4.6948, 5.2015, 4.9127], + device='cuda:0'), covar=tensor([0.0425, 0.0375, 0.1305, 0.0650, 0.0505, 0.0363, 0.0375, 0.0863], + device='cuda:0'), in_proj_covar=tensor([0.0454, 0.0508, 0.0663, 0.0402, 0.0393, 0.0462, 0.0494, 0.0633], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 07:06:52,212 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9295, 4.8234, 5.0377, 4.7223, 4.7800, 4.7605, 5.1245, 5.1034], + device='cuda:0'), covar=tensor([0.0074, 0.0091, 0.0072, 0.0111, 0.0081, 0.0131, 0.0109, 0.0137], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0064, 0.0066, 0.0085, 0.0069, 0.0095, 0.0082, 0.0082], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 07:07:15,708 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53056.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:07:17,593 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.753e+02 2.734e+02 3.607e+02 4.426e+02 8.494e+02, threshold=7.214e+02, percent-clipped=2.0 +2023-03-09 07:07:33,377 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3659, 4.7586, 4.3369, 4.5972, 4.4562, 4.3911, 4.8317, 4.7477], + device='cuda:0'), covar=tensor([0.1120, 0.0762, 0.2115, 0.0686, 0.1341, 0.0756, 0.0648, 0.0731], + device='cuda:0'), in_proj_covar=tensor([0.0569, 0.0480, 0.0355, 0.0505, 0.0693, 0.0504, 0.0679, 0.0501], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 07:07:40,261 INFO [train.py:898] (0/4) Epoch 15, batch 2200, loss[loss=0.1895, simple_loss=0.2783, pruned_loss=0.05039, over 18360.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2631, pruned_loss=0.04449, over 3572974.28 frames. ], batch size: 50, lr: 7.54e-03, grad_scale: 8.0 +2023-03-09 07:07:47,503 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7484, 3.7267, 5.0330, 2.7395, 4.4673, 2.5833, 3.0269, 1.8236], + device='cuda:0'), covar=tensor([0.1085, 0.0893, 0.0111, 0.0905, 0.0495, 0.2573, 0.2668, 0.1993], + device='cuda:0'), in_proj_covar=tensor([0.0209, 0.0227, 0.0144, 0.0184, 0.0241, 0.0259, 0.0303, 0.0221], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 07:08:10,250 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-09 07:08:16,804 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53108.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:08:27,933 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53117.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:08:39,055 INFO [train.py:898] (0/4) Epoch 15, batch 2250, loss[loss=0.1476, simple_loss=0.2284, pruned_loss=0.03344, over 18406.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2635, pruned_loss=0.04442, over 3579311.77 frames. ], batch size: 42, lr: 7.54e-03, grad_scale: 8.0 +2023-03-09 07:09:15,603 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.829e+02 2.883e+02 3.334e+02 3.992e+02 8.132e+02, threshold=6.668e+02, percent-clipped=1.0 +2023-03-09 07:09:37,839 INFO [train.py:898] (0/4) Epoch 15, batch 2300, loss[loss=0.1498, simple_loss=0.2397, pruned_loss=0.02998, over 18355.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2638, pruned_loss=0.04447, over 3570930.93 frames. ], batch size: 46, lr: 7.53e-03, grad_scale: 8.0 +2023-03-09 07:10:18,729 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53211.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 07:10:30,673 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4136, 5.2400, 5.5683, 5.5960, 5.4027, 6.1698, 5.8251, 5.5770], + device='cuda:0'), covar=tensor([0.1101, 0.0565, 0.0690, 0.0601, 0.1208, 0.0662, 0.0569, 0.1587], + device='cuda:0'), in_proj_covar=tensor([0.0339, 0.0265, 0.0282, 0.0282, 0.0318, 0.0395, 0.0259, 0.0388], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 07:10:37,433 INFO [train.py:898] (0/4) Epoch 15, batch 2350, loss[loss=0.1685, simple_loss=0.2624, pruned_loss=0.03733, over 18357.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2629, pruned_loss=0.04434, over 3567040.09 frames. ], batch size: 55, lr: 7.53e-03, grad_scale: 8.0 +2023-03-09 07:10:54,360 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53241.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:11:10,951 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8937, 3.7733, 3.6380, 3.3708, 3.6353, 3.1065, 2.8627, 3.9116], + device='cuda:0'), covar=tensor([0.0047, 0.0078, 0.0069, 0.0106, 0.0071, 0.0147, 0.0163, 0.0046], + device='cuda:0'), in_proj_covar=tensor([0.0118, 0.0139, 0.0119, 0.0172, 0.0123, 0.0167, 0.0170, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 07:11:14,248 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.169e+02 3.108e+02 3.704e+02 4.390e+02 1.359e+03, threshold=7.409e+02, percent-clipped=1.0 +2023-03-09 07:11:15,591 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=53259.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 07:11:22,383 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0958, 4.5621, 4.2528, 4.3130, 4.1903, 4.7395, 4.5021, 4.2257], + device='cuda:0'), covar=tensor([0.1474, 0.1115, 0.0921, 0.0898, 0.1660, 0.1227, 0.0713, 0.2044], + device='cuda:0'), in_proj_covar=tensor([0.0339, 0.0266, 0.0282, 0.0283, 0.0319, 0.0396, 0.0259, 0.0389], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 07:11:36,235 INFO [train.py:898] (0/4) Epoch 15, batch 2400, loss[loss=0.1904, simple_loss=0.2758, pruned_loss=0.05247, over 18267.00 frames. ], tot_loss[loss=0.1757, simple_loss=0.2627, pruned_loss=0.04433, over 3566056.85 frames. ], batch size: 57, lr: 7.53e-03, grad_scale: 8.0 +2023-03-09 07:12:04,942 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53302.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:12:07,133 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53304.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:12:20,505 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53315.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:12:34,309 INFO [train.py:898] (0/4) Epoch 15, batch 2450, loss[loss=0.1747, simple_loss=0.2603, pruned_loss=0.04458, over 18375.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.262, pruned_loss=0.04391, over 3590220.44 frames. ], batch size: 50, lr: 7.52e-03, grad_scale: 8.0 +2023-03-09 07:12:41,554 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9462, 5.3556, 5.3584, 5.3651, 4.8689, 5.1962, 4.6485, 5.2205], + device='cuda:0'), covar=tensor([0.0196, 0.0278, 0.0190, 0.0381, 0.0361, 0.0229, 0.1127, 0.0309], + device='cuda:0'), in_proj_covar=tensor([0.0194, 0.0242, 0.0235, 0.0288, 0.0248, 0.0244, 0.0295, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 07:12:54,170 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-09 07:12:57,218 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7944, 3.7398, 3.5834, 3.3608, 3.5143, 3.0234, 3.0636, 3.7326], + device='cuda:0'), covar=tensor([0.0064, 0.0081, 0.0075, 0.0113, 0.0105, 0.0166, 0.0165, 0.0077], + device='cuda:0'), in_proj_covar=tensor([0.0118, 0.0138, 0.0119, 0.0172, 0.0123, 0.0166, 0.0170, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 07:13:01,947 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-09 07:13:10,220 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.975e+02 3.035e+02 3.526e+02 4.195e+02 8.583e+02, threshold=7.052e+02, percent-clipped=2.0 +2023-03-09 07:13:19,476 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53365.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:13:21,683 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=53367.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:13:32,516 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53376.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:13:33,325 INFO [train.py:898] (0/4) Epoch 15, batch 2500, loss[loss=0.1793, simple_loss=0.2713, pruned_loss=0.04362, over 18561.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2618, pruned_loss=0.04368, over 3593486.50 frames. ], batch size: 54, lr: 7.52e-03, grad_scale: 8.0 +2023-03-09 07:13:35,213 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.52 vs. limit=5.0 +2023-03-09 07:14:09,341 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53408.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:14:14,163 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53412.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:14:32,211 INFO [train.py:898] (0/4) Epoch 15, batch 2550, loss[loss=0.153, simple_loss=0.243, pruned_loss=0.03146, over 18421.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2623, pruned_loss=0.04408, over 3588660.38 frames. ], batch size: 48, lr: 7.52e-03, grad_scale: 8.0 +2023-03-09 07:14:33,585 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=53428.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 07:14:53,815 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-09 07:15:02,268 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9711, 3.7404, 4.9426, 2.9045, 4.3217, 2.5369, 3.0121, 1.7884], + device='cuda:0'), covar=tensor([0.1014, 0.0826, 0.0115, 0.0798, 0.0561, 0.2470, 0.2545, 0.2014], + device='cuda:0'), in_proj_covar=tensor([0.0206, 0.0225, 0.0144, 0.0182, 0.0240, 0.0255, 0.0299, 0.0218], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 07:15:05,448 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=53456.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:15:07,569 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.050e+02 2.986e+02 3.579e+02 4.675e+02 1.603e+03, threshold=7.159e+02, percent-clipped=6.0 +2023-03-09 07:15:08,996 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4976, 2.8135, 2.5593, 2.8871, 3.7014, 3.5667, 3.1195, 2.9100], + device='cuda:0'), covar=tensor([0.0214, 0.0241, 0.0512, 0.0333, 0.0145, 0.0134, 0.0337, 0.0360], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0123, 0.0158, 0.0150, 0.0114, 0.0103, 0.0142, 0.0145], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 07:15:30,224 INFO [train.py:898] (0/4) Epoch 15, batch 2600, loss[loss=0.1908, simple_loss=0.2774, pruned_loss=0.0521, over 16178.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.2629, pruned_loss=0.04392, over 3592571.34 frames. ], batch size: 94, lr: 7.51e-03, grad_scale: 8.0 +2023-03-09 07:16:28,752 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8049, 4.7033, 2.6532, 4.5359, 4.5119, 4.7374, 4.5166, 2.5848], + device='cuda:0'), covar=tensor([0.0206, 0.0068, 0.0758, 0.0113, 0.0075, 0.0075, 0.0109, 0.1024], + device='cuda:0'), in_proj_covar=tensor([0.0083, 0.0074, 0.0091, 0.0087, 0.0080, 0.0070, 0.0081, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 07:16:29,544 INFO [train.py:898] (0/4) Epoch 15, batch 2650, loss[loss=0.1799, simple_loss=0.2705, pruned_loss=0.04464, over 15972.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2629, pruned_loss=0.04386, over 3596382.57 frames. ], batch size: 94, lr: 7.51e-03, grad_scale: 8.0 +2023-03-09 07:17:05,490 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.951e+02 2.884e+02 3.324e+02 4.046e+02 9.778e+02, threshold=6.647e+02, percent-clipped=2.0 +2023-03-09 07:17:27,744 INFO [train.py:898] (0/4) Epoch 15, batch 2700, loss[loss=0.1768, simple_loss=0.2699, pruned_loss=0.04183, over 17009.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2631, pruned_loss=0.04378, over 3594819.40 frames. ], batch size: 78, lr: 7.51e-03, grad_scale: 8.0 +2023-03-09 07:17:51,335 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53597.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:18:17,512 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8170, 5.2769, 5.2555, 5.2280, 4.8219, 5.1706, 4.5691, 5.1403], + device='cuda:0'), covar=tensor([0.0230, 0.0275, 0.0193, 0.0378, 0.0357, 0.0231, 0.1091, 0.0307], + device='cuda:0'), in_proj_covar=tensor([0.0196, 0.0244, 0.0234, 0.0290, 0.0249, 0.0246, 0.0297, 0.0240], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 07:18:26,757 INFO [train.py:898] (0/4) Epoch 15, batch 2750, loss[loss=0.1942, simple_loss=0.2838, pruned_loss=0.05227, over 18087.00 frames. ], tot_loss[loss=0.175, simple_loss=0.2625, pruned_loss=0.04374, over 3603090.38 frames. ], batch size: 62, lr: 7.50e-03, grad_scale: 8.0 +2023-03-09 07:18:29,806 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6628, 5.1372, 5.1377, 5.1688, 4.6544, 5.0333, 4.3249, 5.0132], + device='cuda:0'), covar=tensor([0.0261, 0.0366, 0.0247, 0.0407, 0.0416, 0.0307, 0.1341, 0.0345], + device='cuda:0'), in_proj_covar=tensor([0.0197, 0.0246, 0.0236, 0.0293, 0.0251, 0.0247, 0.0299, 0.0241], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 07:19:03,251 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.144e+02 2.856e+02 3.450e+02 4.147e+02 9.079e+02, threshold=6.900e+02, percent-clipped=4.0 +2023-03-09 07:19:05,787 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53660.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:19:16,981 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0097, 3.1918, 4.6259, 4.0889, 3.0842, 4.7874, 4.1145, 3.1923], + device='cuda:0'), covar=tensor([0.0416, 0.1229, 0.0212, 0.0346, 0.1318, 0.0191, 0.0471, 0.0851], + device='cuda:0'), in_proj_covar=tensor([0.0199, 0.0228, 0.0181, 0.0150, 0.0217, 0.0196, 0.0225, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 07:19:17,938 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53671.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:19:24,960 INFO [train.py:898] (0/4) Epoch 15, batch 2800, loss[loss=0.1858, simple_loss=0.269, pruned_loss=0.05125, over 18497.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.263, pruned_loss=0.04394, over 3607774.42 frames. ], batch size: 44, lr: 7.50e-03, grad_scale: 8.0 +2023-03-09 07:20:06,923 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53712.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:20:19,427 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=53723.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 07:20:24,338 INFO [train.py:898] (0/4) Epoch 15, batch 2850, loss[loss=0.1886, simple_loss=0.2776, pruned_loss=0.04981, over 16136.00 frames. ], tot_loss[loss=0.1752, simple_loss=0.2628, pruned_loss=0.04376, over 3603813.48 frames. ], batch size: 94, lr: 7.50e-03, grad_scale: 8.0 +2023-03-09 07:20:55,751 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2755, 4.0646, 5.2461, 3.0681, 4.6527, 2.8125, 3.2481, 2.1141], + device='cuda:0'), covar=tensor([0.0839, 0.0694, 0.0126, 0.0753, 0.0455, 0.2180, 0.2390, 0.1756], + device='cuda:0'), in_proj_covar=tensor([0.0205, 0.0224, 0.0144, 0.0181, 0.0238, 0.0254, 0.0298, 0.0218], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 07:21:00,856 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.810e+02 2.667e+02 3.515e+02 4.128e+02 7.427e+02, threshold=7.030e+02, percent-clipped=1.0 +2023-03-09 07:21:03,166 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=53760.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:21:22,592 INFO [train.py:898] (0/4) Epoch 15, batch 2900, loss[loss=0.1775, simple_loss=0.2718, pruned_loss=0.0416, over 17944.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2617, pruned_loss=0.04316, over 3607207.73 frames. ], batch size: 65, lr: 7.49e-03, grad_scale: 8.0 +2023-03-09 07:21:34,348 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6998, 3.6880, 2.2496, 4.4648, 3.1727, 4.4768, 2.5785, 4.0898], + device='cuda:0'), covar=tensor([0.0555, 0.0751, 0.1382, 0.0527, 0.0761, 0.0349, 0.1077, 0.0361], + device='cuda:0'), in_proj_covar=tensor([0.0205, 0.0220, 0.0184, 0.0266, 0.0187, 0.0257, 0.0197, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 07:21:45,246 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-09 07:22:21,623 INFO [train.py:898] (0/4) Epoch 15, batch 2950, loss[loss=0.1738, simple_loss=0.2601, pruned_loss=0.04371, over 18410.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2617, pruned_loss=0.04328, over 3604181.39 frames. ], batch size: 48, lr: 7.49e-03, grad_scale: 8.0 +2023-03-09 07:22:48,738 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-09 07:22:53,012 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6907, 3.6100, 4.8143, 4.3388, 3.0604, 2.8640, 4.2132, 5.1611], + device='cuda:0'), covar=tensor([0.0831, 0.1471, 0.0210, 0.0362, 0.1041, 0.1238, 0.0405, 0.0174], + device='cuda:0'), in_proj_covar=tensor([0.0143, 0.0264, 0.0128, 0.0175, 0.0186, 0.0185, 0.0188, 0.0173], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 07:22:55,345 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5202, 2.2016, 2.4985, 2.6171, 3.3047, 4.9413, 4.6973, 3.7616], + device='cuda:0'), covar=tensor([0.1609, 0.2284, 0.2797, 0.1681, 0.2067, 0.0189, 0.0370, 0.0668], + device='cuda:0'), in_proj_covar=tensor([0.0272, 0.0327, 0.0352, 0.0263, 0.0378, 0.0215, 0.0278, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 07:22:58,162 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.012e+02 2.809e+02 3.403e+02 4.048e+02 1.283e+03, threshold=6.805e+02, percent-clipped=2.0 +2023-03-09 07:23:20,580 INFO [train.py:898] (0/4) Epoch 15, batch 3000, loss[loss=0.1836, simple_loss=0.269, pruned_loss=0.04914, over 17998.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2615, pruned_loss=0.04339, over 3597307.45 frames. ], batch size: 65, lr: 7.49e-03, grad_scale: 8.0 +2023-03-09 07:23:20,581 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 07:23:32,493 INFO [train.py:932] (0/4) Epoch 15, validation: loss=0.1532, simple_loss=0.254, pruned_loss=0.02619, over 944034.00 frames. +2023-03-09 07:23:32,493 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 07:23:36,229 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1022, 5.1262, 5.2062, 5.2533, 5.0273, 5.8507, 5.4997, 5.1722], + device='cuda:0'), covar=tensor([0.1068, 0.0721, 0.0769, 0.0673, 0.1463, 0.0787, 0.0659, 0.1844], + device='cuda:0'), in_proj_covar=tensor([0.0341, 0.0266, 0.0288, 0.0284, 0.0321, 0.0398, 0.0261, 0.0394], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 07:23:53,980 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3378, 3.2676, 3.1849, 2.8949, 3.1204, 2.5099, 2.5640, 3.3267], + device='cuda:0'), covar=tensor([0.0051, 0.0074, 0.0078, 0.0108, 0.0078, 0.0157, 0.0178, 0.0060], + device='cuda:0'), in_proj_covar=tensor([0.0119, 0.0140, 0.0121, 0.0172, 0.0125, 0.0168, 0.0172, 0.0102], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 07:23:56,040 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53897.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:24:26,236 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6400, 3.4878, 1.9517, 4.4327, 3.0779, 4.3982, 2.3733, 3.9595], + device='cuda:0'), covar=tensor([0.0589, 0.0855, 0.1594, 0.0423, 0.0873, 0.0268, 0.1213, 0.0399], + device='cuda:0'), in_proj_covar=tensor([0.0208, 0.0221, 0.0186, 0.0267, 0.0189, 0.0260, 0.0198, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 07:24:30,756 INFO [train.py:898] (0/4) Epoch 15, batch 3050, loss[loss=0.2073, simple_loss=0.2918, pruned_loss=0.06135, over 18495.00 frames. ], tot_loss[loss=0.1745, simple_loss=0.2618, pruned_loss=0.04363, over 3599942.22 frames. ], batch size: 51, lr: 7.48e-03, grad_scale: 4.0 +2023-03-09 07:24:33,336 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8436, 4.1542, 2.3567, 4.1446, 5.1491, 2.5720, 3.4313, 3.5672], + device='cuda:0'), covar=tensor([0.0185, 0.1124, 0.1756, 0.0568, 0.0071, 0.1311, 0.0916, 0.1033], + device='cuda:0'), in_proj_covar=tensor([0.0141, 0.0249, 0.0195, 0.0187, 0.0102, 0.0177, 0.0205, 0.0214], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 07:24:42,371 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.4779, 3.6576, 4.8883, 4.2352, 3.1082, 2.9461, 4.2717, 5.1776], + device='cuda:0'), covar=tensor([0.0856, 0.1410, 0.0222, 0.0404, 0.0949, 0.1175, 0.0400, 0.0234], + device='cuda:0'), in_proj_covar=tensor([0.0141, 0.0261, 0.0127, 0.0173, 0.0183, 0.0184, 0.0185, 0.0172], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 07:24:50,609 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2486, 4.1316, 5.2807, 3.1425, 4.6791, 2.8478, 3.1713, 2.1509], + device='cuda:0'), covar=tensor([0.0816, 0.0673, 0.0107, 0.0740, 0.0461, 0.2328, 0.2492, 0.1752], + device='cuda:0'), in_proj_covar=tensor([0.0205, 0.0224, 0.0144, 0.0181, 0.0239, 0.0253, 0.0296, 0.0217], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 07:24:52,730 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=53945.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:25:08,736 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.036e+02 2.885e+02 3.297e+02 3.882e+02 9.425e+02, threshold=6.594e+02, percent-clipped=2.0 +2023-03-09 07:25:10,176 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53960.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:25:22,549 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=53971.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:25:29,043 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.47 vs. limit=5.0 +2023-03-09 07:25:29,403 INFO [train.py:898] (0/4) Epoch 15, batch 3100, loss[loss=0.1999, simple_loss=0.2957, pruned_loss=0.05208, over 17218.00 frames. ], tot_loss[loss=0.1748, simple_loss=0.2621, pruned_loss=0.04375, over 3583430.03 frames. ], batch size: 78, lr: 7.48e-03, grad_scale: 4.0 +2023-03-09 07:25:56,691 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-54000.pt +2023-03-09 07:26:10,181 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=54008.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:26:23,189 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=54019.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:26:27,726 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=54023.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 07:26:31,791 INFO [train.py:898] (0/4) Epoch 15, batch 3150, loss[loss=0.1902, simple_loss=0.2809, pruned_loss=0.04973, over 18565.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.2625, pruned_loss=0.04409, over 3581591.23 frames. ], batch size: 54, lr: 7.48e-03, grad_scale: 4.0 +2023-03-09 07:26:33,306 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7415, 2.9637, 2.7669, 3.0089, 3.7174, 3.6685, 3.2128, 3.0087], + device='cuda:0'), covar=tensor([0.0168, 0.0281, 0.0483, 0.0315, 0.0187, 0.0154, 0.0326, 0.0373], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0126, 0.0160, 0.0151, 0.0115, 0.0106, 0.0144, 0.0147], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 07:27:09,867 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.040e+02 2.859e+02 3.550e+02 4.150e+02 8.480e+02, threshold=7.099e+02, percent-clipped=1.0 +2023-03-09 07:27:24,062 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=54071.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:27:30,816 INFO [train.py:898] (0/4) Epoch 15, batch 3200, loss[loss=0.1912, simple_loss=0.2798, pruned_loss=0.05133, over 18000.00 frames. ], tot_loss[loss=0.176, simple_loss=0.2633, pruned_loss=0.04431, over 3580609.56 frames. ], batch size: 65, lr: 7.47e-03, grad_scale: 8.0 +2023-03-09 07:28:25,013 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2957, 4.0750, 5.2913, 3.1631, 4.6437, 2.7270, 3.1274, 2.1310], + device='cuda:0'), covar=tensor([0.0863, 0.0778, 0.0092, 0.0744, 0.0473, 0.2359, 0.2702, 0.1861], + device='cuda:0'), in_proj_covar=tensor([0.0210, 0.0228, 0.0147, 0.0184, 0.0244, 0.0258, 0.0303, 0.0221], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 07:28:28,993 INFO [train.py:898] (0/4) Epoch 15, batch 3250, loss[loss=0.2251, simple_loss=0.3071, pruned_loss=0.07148, over 12713.00 frames. ], tot_loss[loss=0.1763, simple_loss=0.2636, pruned_loss=0.04452, over 3582465.07 frames. ], batch size: 131, lr: 7.47e-03, grad_scale: 8.0 +2023-03-09 07:29:07,308 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.872e+02 2.643e+02 3.129e+02 4.018e+02 8.489e+02, threshold=6.258e+02, percent-clipped=1.0 +2023-03-09 07:29:28,035 INFO [train.py:898] (0/4) Epoch 15, batch 3300, loss[loss=0.1509, simple_loss=0.2333, pruned_loss=0.03422, over 17729.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2637, pruned_loss=0.04433, over 3581771.53 frames. ], batch size: 39, lr: 7.46e-03, grad_scale: 8.0 +2023-03-09 07:30:27,872 INFO [train.py:898] (0/4) Epoch 15, batch 3350, loss[loss=0.1991, simple_loss=0.284, pruned_loss=0.0571, over 16106.00 frames. ], tot_loss[loss=0.1765, simple_loss=0.264, pruned_loss=0.04446, over 3569980.99 frames. ], batch size: 95, lr: 7.46e-03, grad_scale: 8.0 +2023-03-09 07:31:03,028 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.53 vs. limit=2.0 +2023-03-09 07:31:05,787 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.725e+02 2.934e+02 3.343e+02 4.171e+02 1.510e+03, threshold=6.685e+02, percent-clipped=4.0 +2023-03-09 07:31:26,833 INFO [train.py:898] (0/4) Epoch 15, batch 3400, loss[loss=0.1504, simple_loss=0.233, pruned_loss=0.03394, over 18485.00 frames. ], tot_loss[loss=0.1762, simple_loss=0.2639, pruned_loss=0.04425, over 3573984.48 frames. ], batch size: 44, lr: 7.46e-03, grad_scale: 8.0 +2023-03-09 07:32:24,911 INFO [train.py:898] (0/4) Epoch 15, batch 3450, loss[loss=0.1465, simple_loss=0.2239, pruned_loss=0.0346, over 18399.00 frames. ], tot_loss[loss=0.1764, simple_loss=0.2639, pruned_loss=0.04444, over 3567332.60 frames. ], batch size: 43, lr: 7.45e-03, grad_scale: 8.0 +2023-03-09 07:33:02,928 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.097e+02 3.124e+02 3.686e+02 4.996e+02 1.186e+03, threshold=7.372e+02, percent-clipped=7.0 +2023-03-09 07:33:23,261 INFO [train.py:898] (0/4) Epoch 15, batch 3500, loss[loss=0.1836, simple_loss=0.274, pruned_loss=0.04659, over 17086.00 frames. ], tot_loss[loss=0.1754, simple_loss=0.263, pruned_loss=0.04388, over 3570535.52 frames. ], batch size: 78, lr: 7.45e-03, grad_scale: 8.0 +2023-03-09 07:33:40,246 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.75 vs. limit=5.0 +2023-03-09 07:34:20,581 INFO [train.py:898] (0/4) Epoch 15, batch 3550, loss[loss=0.1853, simple_loss=0.267, pruned_loss=0.05184, over 18086.00 frames. ], tot_loss[loss=0.1753, simple_loss=0.263, pruned_loss=0.04382, over 3576036.73 frames. ], batch size: 62, lr: 7.45e-03, grad_scale: 8.0 +2023-03-09 07:34:54,890 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.932e+02 3.002e+02 3.469e+02 4.157e+02 6.561e+02, threshold=6.938e+02, percent-clipped=0.0 +2023-03-09 07:34:58,078 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2568, 5.1464, 5.4524, 5.3366, 5.2672, 5.9757, 5.6162, 5.2629], + device='cuda:0'), covar=tensor([0.0945, 0.0646, 0.0717, 0.0693, 0.1328, 0.0746, 0.0695, 0.1855], + device='cuda:0'), in_proj_covar=tensor([0.0337, 0.0264, 0.0285, 0.0285, 0.0318, 0.0397, 0.0259, 0.0389], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 07:35:13,936 INFO [train.py:898] (0/4) Epoch 15, batch 3600, loss[loss=0.1748, simple_loss=0.2646, pruned_loss=0.04248, over 18361.00 frames. ], tot_loss[loss=0.1758, simple_loss=0.2634, pruned_loss=0.0441, over 3578122.55 frames. ], batch size: 56, lr: 7.44e-03, grad_scale: 8.0 +2023-03-09 07:35:48,616 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-15.pt +2023-03-09 07:36:16,286 INFO [train.py:898] (0/4) Epoch 16, batch 0, loss[loss=0.1879, simple_loss=0.2775, pruned_loss=0.04914, over 18220.00 frames. ], tot_loss[loss=0.1879, simple_loss=0.2775, pruned_loss=0.04914, over 18220.00 frames. ], batch size: 60, lr: 7.20e-03, grad_scale: 8.0 +2023-03-09 07:36:16,288 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 07:36:28,071 INFO [train.py:932] (0/4) Epoch 16, validation: loss=0.1541, simple_loss=0.2552, pruned_loss=0.02651, over 944034.00 frames. +2023-03-09 07:36:28,072 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 07:37:06,270 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8727, 3.7903, 3.7207, 3.2683, 3.5643, 2.8466, 2.8689, 3.7562], + device='cuda:0'), covar=tensor([0.0043, 0.0072, 0.0065, 0.0117, 0.0075, 0.0188, 0.0180, 0.0062], + device='cuda:0'), in_proj_covar=tensor([0.0118, 0.0138, 0.0118, 0.0170, 0.0122, 0.0165, 0.0168, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 07:37:24,164 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.170e+02 3.007e+02 3.556e+02 4.370e+02 7.449e+02, threshold=7.113e+02, percent-clipped=5.0 +2023-03-09 07:37:26,546 INFO [train.py:898] (0/4) Epoch 16, batch 50, loss[loss=0.1731, simple_loss=0.271, pruned_loss=0.03762, over 18554.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2616, pruned_loss=0.04314, over 805951.51 frames. ], batch size: 54, lr: 7.20e-03, grad_scale: 8.0 +2023-03-09 07:38:25,630 INFO [train.py:898] (0/4) Epoch 16, batch 100, loss[loss=0.1864, simple_loss=0.2825, pruned_loss=0.04519, over 18247.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2596, pruned_loss=0.0423, over 1425412.15 frames. ], batch size: 60, lr: 7.20e-03, grad_scale: 8.0 +2023-03-09 07:39:21,644 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.042e+02 2.818e+02 3.254e+02 3.972e+02 9.345e+02, threshold=6.508e+02, percent-clipped=3.0 +2023-03-09 07:39:23,839 INFO [train.py:898] (0/4) Epoch 16, batch 150, loss[loss=0.1473, simple_loss=0.2269, pruned_loss=0.0338, over 18486.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2598, pruned_loss=0.04204, over 1907144.90 frames. ], batch size: 44, lr: 7.19e-03, grad_scale: 8.0 +2023-03-09 07:39:27,582 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=54664.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 07:39:47,464 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5429, 3.3098, 2.0301, 4.1944, 2.9514, 4.1044, 2.2983, 3.8473], + device='cuda:0'), covar=tensor([0.0589, 0.0900, 0.1619, 0.0514, 0.0912, 0.0285, 0.1287, 0.0418], + device='cuda:0'), in_proj_covar=tensor([0.0207, 0.0220, 0.0185, 0.0268, 0.0187, 0.0254, 0.0196, 0.0193], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 07:40:22,365 INFO [train.py:898] (0/4) Epoch 16, batch 200, loss[loss=0.1811, simple_loss=0.2701, pruned_loss=0.04608, over 18443.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.26, pruned_loss=0.04262, over 2279364.39 frames. ], batch size: 59, lr: 7.19e-03, grad_scale: 8.0 +2023-03-09 07:40:38,429 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=54725.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 07:40:39,499 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3570, 2.6467, 2.2970, 2.7466, 3.4843, 3.3713, 3.0034, 2.8425], + device='cuda:0'), covar=tensor([0.0177, 0.0273, 0.0576, 0.0364, 0.0131, 0.0135, 0.0310, 0.0344], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0127, 0.0161, 0.0150, 0.0114, 0.0105, 0.0145, 0.0146], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 07:41:17,921 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.955e+02 3.015e+02 3.636e+02 4.590e+02 9.600e+02, threshold=7.273e+02, percent-clipped=5.0 +2023-03-09 07:41:18,611 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-09 07:41:20,199 INFO [train.py:898] (0/4) Epoch 16, batch 250, loss[loss=0.2044, simple_loss=0.2804, pruned_loss=0.06421, over 12620.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2604, pruned_loss=0.0428, over 2572490.12 frames. ], batch size: 129, lr: 7.19e-03, grad_scale: 8.0 +2023-03-09 07:41:49,768 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6087, 3.5106, 3.4554, 3.0944, 3.3911, 2.6944, 2.7640, 3.5555], + device='cuda:0'), covar=tensor([0.0052, 0.0092, 0.0066, 0.0120, 0.0083, 0.0164, 0.0174, 0.0060], + device='cuda:0'), in_proj_covar=tensor([0.0117, 0.0138, 0.0117, 0.0168, 0.0123, 0.0164, 0.0166, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 07:42:17,952 INFO [train.py:898] (0/4) Epoch 16, batch 300, loss[loss=0.1663, simple_loss=0.2568, pruned_loss=0.03795, over 17146.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2593, pruned_loss=0.04264, over 2800972.46 frames. ], batch size: 78, lr: 7.18e-03, grad_scale: 8.0 +2023-03-09 07:42:40,204 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.18 vs. limit=5.0 +2023-03-09 07:43:14,226 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.101e+02 2.934e+02 3.524e+02 4.515e+02 1.434e+03, threshold=7.048e+02, percent-clipped=4.0 +2023-03-09 07:43:16,400 INFO [train.py:898] (0/4) Epoch 16, batch 350, loss[loss=0.152, simple_loss=0.2391, pruned_loss=0.03246, over 18221.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2597, pruned_loss=0.04293, over 2970809.62 frames. ], batch size: 45, lr: 7.18e-03, grad_scale: 8.0 +2023-03-09 07:43:17,953 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=54862.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:43:33,320 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1589, 5.3101, 2.8346, 5.1410, 4.9578, 5.2725, 4.9889, 2.3386], + device='cuda:0'), covar=tensor([0.0185, 0.0093, 0.0873, 0.0117, 0.0098, 0.0111, 0.0165, 0.1660], + device='cuda:0'), in_proj_covar=tensor([0.0083, 0.0074, 0.0091, 0.0087, 0.0080, 0.0071, 0.0081, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 07:44:15,025 INFO [train.py:898] (0/4) Epoch 16, batch 400, loss[loss=0.1495, simple_loss=0.2275, pruned_loss=0.03569, over 18493.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2602, pruned_loss=0.0429, over 3112757.39 frames. ], batch size: 44, lr: 7.18e-03, grad_scale: 8.0 +2023-03-09 07:44:26,673 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7502, 3.6727, 3.5972, 3.2157, 3.5583, 2.8407, 2.7891, 3.7672], + device='cuda:0'), covar=tensor([0.0052, 0.0101, 0.0071, 0.0133, 0.0081, 0.0172, 0.0188, 0.0058], + device='cuda:0'), in_proj_covar=tensor([0.0119, 0.0139, 0.0119, 0.0171, 0.0124, 0.0166, 0.0168, 0.0101], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 07:44:28,934 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=54923.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 07:44:46,422 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-09 07:44:55,719 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.41 vs. limit=5.0 +2023-03-09 07:45:00,649 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6801, 4.2065, 4.2925, 3.2210, 3.5761, 3.3493, 2.4404, 2.3873], + device='cuda:0'), covar=tensor([0.0220, 0.0149, 0.0086, 0.0303, 0.0345, 0.0219, 0.0747, 0.0845], + device='cuda:0'), in_proj_covar=tensor([0.0066, 0.0055, 0.0056, 0.0065, 0.0087, 0.0063, 0.0077, 0.0083], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 07:45:00,686 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4346, 3.2848, 1.9997, 4.1140, 2.8431, 4.0073, 2.2393, 3.5169], + device='cuda:0'), covar=tensor([0.0560, 0.0734, 0.1425, 0.0508, 0.0820, 0.0344, 0.1181, 0.0466], + device='cuda:0'), in_proj_covar=tensor([0.0207, 0.0221, 0.0186, 0.0268, 0.0188, 0.0256, 0.0196, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 07:45:11,970 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.963e+02 2.717e+02 3.272e+02 3.926e+02 7.148e+02, threshold=6.544e+02, percent-clipped=1.0 +2023-03-09 07:45:13,179 INFO [train.py:898] (0/4) Epoch 16, batch 450, loss[loss=0.1792, simple_loss=0.2721, pruned_loss=0.04319, over 18560.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2611, pruned_loss=0.04317, over 3224131.72 frames. ], batch size: 54, lr: 7.17e-03, grad_scale: 8.0 +2023-03-09 07:45:36,051 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8125, 3.5729, 4.9569, 4.2046, 3.1038, 2.8235, 4.2144, 5.1115], + device='cuda:0'), covar=tensor([0.0831, 0.1523, 0.0135, 0.0426, 0.1052, 0.1276, 0.0466, 0.0202], + device='cuda:0'), in_proj_covar=tensor([0.0144, 0.0266, 0.0130, 0.0174, 0.0184, 0.0185, 0.0186, 0.0175], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 07:46:12,498 INFO [train.py:898] (0/4) Epoch 16, batch 500, loss[loss=0.1747, simple_loss=0.2676, pruned_loss=0.0409, over 18506.00 frames. ], tot_loss[loss=0.1749, simple_loss=0.2625, pruned_loss=0.0437, over 3286937.54 frames. ], batch size: 53, lr: 7.17e-03, grad_scale: 8.0 +2023-03-09 07:46:23,061 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55020.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 07:46:37,208 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8344, 3.7171, 5.0103, 3.0240, 4.4447, 2.7715, 3.0921, 1.8558], + device='cuda:0'), covar=tensor([0.1090, 0.0807, 0.0140, 0.0802, 0.0535, 0.2246, 0.2651, 0.2051], + device='cuda:0'), in_proj_covar=tensor([0.0210, 0.0228, 0.0147, 0.0183, 0.0245, 0.0257, 0.0304, 0.0221], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 07:47:02,738 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5334, 2.7548, 2.5062, 2.8228, 3.5602, 3.6047, 3.1407, 2.9616], + device='cuda:0'), covar=tensor([0.0183, 0.0339, 0.0632, 0.0397, 0.0185, 0.0138, 0.0377, 0.0414], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0126, 0.0160, 0.0149, 0.0115, 0.0105, 0.0145, 0.0145], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 07:47:09,466 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.142e+02 2.826e+02 3.526e+02 4.225e+02 1.034e+03, threshold=7.052e+02, percent-clipped=4.0 +2023-03-09 07:47:10,629 INFO [train.py:898] (0/4) Epoch 16, batch 550, loss[loss=0.1506, simple_loss=0.2386, pruned_loss=0.03131, over 18347.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2602, pruned_loss=0.04278, over 3365714.55 frames. ], batch size: 46, lr: 7.17e-03, grad_scale: 8.0 +2023-03-09 07:47:21,431 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6756, 2.2192, 2.5724, 2.5884, 3.2346, 4.7962, 4.4302, 3.3752], + device='cuda:0'), covar=tensor([0.1548, 0.2326, 0.2939, 0.1739, 0.2013, 0.0154, 0.0437, 0.0764], + device='cuda:0'), in_proj_covar=tensor([0.0273, 0.0331, 0.0356, 0.0264, 0.0379, 0.0216, 0.0281, 0.0233], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 07:48:09,165 INFO [train.py:898] (0/4) Epoch 16, batch 600, loss[loss=0.1528, simple_loss=0.2302, pruned_loss=0.03772, over 18159.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2596, pruned_loss=0.0424, over 3414622.87 frames. ], batch size: 44, lr: 7.16e-03, grad_scale: 8.0 +2023-03-09 07:48:51,582 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=55148.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:49:05,764 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.722e+02 2.752e+02 3.250e+02 4.071e+02 8.362e+02, threshold=6.500e+02, percent-clipped=2.0 +2023-03-09 07:49:06,097 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9857, 5.0266, 5.1515, 4.8228, 4.7501, 4.9032, 5.2224, 5.2414], + device='cuda:0'), covar=tensor([0.0064, 0.0063, 0.0043, 0.0090, 0.0064, 0.0136, 0.0064, 0.0093], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0065, 0.0069, 0.0088, 0.0071, 0.0098, 0.0083, 0.0083], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 07:49:06,961 INFO [train.py:898] (0/4) Epoch 16, batch 650, loss[loss=0.1575, simple_loss=0.2437, pruned_loss=0.03567, over 18509.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2591, pruned_loss=0.04226, over 3454128.97 frames. ], batch size: 47, lr: 7.16e-03, grad_scale: 8.0 +2023-03-09 07:50:03,411 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55209.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:50:05,814 INFO [train.py:898] (0/4) Epoch 16, batch 700, loss[loss=0.1539, simple_loss=0.2319, pruned_loss=0.03792, over 18569.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2591, pruned_loss=0.04217, over 3483487.72 frames. ], batch size: 45, lr: 7.16e-03, grad_scale: 4.0 +2023-03-09 07:50:06,089 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=55211.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:50:14,494 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55218.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 07:51:02,204 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9752, 5.0930, 5.1080, 4.8634, 4.8915, 4.8896, 5.2203, 5.2258], + device='cuda:0'), covar=tensor([0.0066, 0.0059, 0.0055, 0.0092, 0.0053, 0.0137, 0.0076, 0.0114], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0064, 0.0068, 0.0087, 0.0070, 0.0097, 0.0082, 0.0083], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 07:51:02,974 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.749e+02 2.606e+02 3.222e+02 3.898e+02 7.462e+02, threshold=6.443e+02, percent-clipped=3.0 +2023-03-09 07:51:02,999 INFO [train.py:898] (0/4) Epoch 16, batch 750, loss[loss=0.1549, simple_loss=0.2389, pruned_loss=0.03543, over 18493.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2596, pruned_loss=0.04239, over 3510017.50 frames. ], batch size: 47, lr: 7.15e-03, grad_scale: 4.0 +2023-03-09 07:51:16,755 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55272.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:51:21,973 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9703, 4.1062, 2.4670, 4.1219, 5.2141, 2.8495, 3.5594, 3.7438], + device='cuda:0'), covar=tensor([0.0110, 0.1233, 0.1423, 0.0504, 0.0051, 0.0937, 0.0624, 0.0842], + device='cuda:0'), in_proj_covar=tensor([0.0145, 0.0252, 0.0196, 0.0190, 0.0104, 0.0177, 0.0208, 0.0217], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 07:52:01,599 INFO [train.py:898] (0/4) Epoch 16, batch 800, loss[loss=0.1807, simple_loss=0.2692, pruned_loss=0.04615, over 18581.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2596, pruned_loss=0.04244, over 3523948.45 frames. ], batch size: 54, lr: 7.15e-03, grad_scale: 8.0 +2023-03-09 07:52:13,197 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55320.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 07:52:46,423 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.3286, 2.0987, 1.9850, 2.0526, 2.4349, 2.4556, 2.3364, 2.1693], + device='cuda:0'), covar=tensor([0.0217, 0.0235, 0.0427, 0.0331, 0.0185, 0.0151, 0.0373, 0.0296], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0126, 0.0159, 0.0149, 0.0115, 0.0104, 0.0145, 0.0144], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 07:53:00,380 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.160e+02 2.840e+02 3.265e+02 3.800e+02 8.424e+02, threshold=6.530e+02, percent-clipped=5.0 +2023-03-09 07:53:00,409 INFO [train.py:898] (0/4) Epoch 16, batch 850, loss[loss=0.1531, simple_loss=0.2388, pruned_loss=0.03369, over 18341.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2595, pruned_loss=0.04233, over 3537488.52 frames. ], batch size: 46, lr: 7.15e-03, grad_scale: 8.0 +2023-03-09 07:53:08,444 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=55368.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 07:53:59,236 INFO [train.py:898] (0/4) Epoch 16, batch 900, loss[loss=0.2086, simple_loss=0.2945, pruned_loss=0.06139, over 18087.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2597, pruned_loss=0.04243, over 3553448.22 frames. ], batch size: 62, lr: 7.15e-03, grad_scale: 8.0 +2023-03-09 07:54:05,308 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8697, 5.2856, 2.7391, 5.1036, 5.0288, 5.2981, 5.0640, 2.5098], + device='cuda:0'), covar=tensor([0.0220, 0.0065, 0.0750, 0.0084, 0.0069, 0.0074, 0.0093, 0.1044], + device='cuda:0'), in_proj_covar=tensor([0.0083, 0.0075, 0.0092, 0.0087, 0.0080, 0.0071, 0.0081, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 07:54:36,511 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-09 07:54:57,055 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.979e+02 2.881e+02 3.338e+02 4.155e+02 1.042e+03, threshold=6.676e+02, percent-clipped=4.0 +2023-03-09 07:54:57,080 INFO [train.py:898] (0/4) Epoch 16, batch 950, loss[loss=0.1797, simple_loss=0.2712, pruned_loss=0.04409, over 17828.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2604, pruned_loss=0.0424, over 3568903.40 frames. ], batch size: 70, lr: 7.14e-03, grad_scale: 8.0 +2023-03-09 07:55:43,770 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.86 vs. limit=2.0 +2023-03-09 07:55:46,613 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55504.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:55:54,397 INFO [train.py:898] (0/4) Epoch 16, batch 1000, loss[loss=0.1585, simple_loss=0.2322, pruned_loss=0.04238, over 17577.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2603, pruned_loss=0.04256, over 3576008.53 frames. ], batch size: 39, lr: 7.14e-03, grad_scale: 8.0 +2023-03-09 07:56:03,260 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55518.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 07:56:53,729 INFO [train.py:898] (0/4) Epoch 16, batch 1050, loss[loss=0.1787, simple_loss=0.27, pruned_loss=0.04365, over 18368.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2607, pruned_loss=0.04279, over 3572908.02 frames. ], batch size: 55, lr: 7.14e-03, grad_scale: 4.0 +2023-03-09 07:56:54,803 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.207e+02 2.979e+02 3.472e+02 4.235e+02 7.011e+02, threshold=6.944e+02, percent-clipped=2.0 +2023-03-09 07:57:00,075 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=55566.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:57:01,342 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=55567.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 07:57:52,415 INFO [train.py:898] (0/4) Epoch 16, batch 1100, loss[loss=0.1457, simple_loss=0.2309, pruned_loss=0.03024, over 18237.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2613, pruned_loss=0.04326, over 3588425.96 frames. ], batch size: 45, lr: 7.13e-03, grad_scale: 4.0 +2023-03-09 07:58:52,113 INFO [train.py:898] (0/4) Epoch 16, batch 1150, loss[loss=0.1514, simple_loss=0.2346, pruned_loss=0.03408, over 18394.00 frames. ], tot_loss[loss=0.174, simple_loss=0.2611, pruned_loss=0.04348, over 3567149.57 frames. ], batch size: 42, lr: 7.13e-03, grad_scale: 4.0 +2023-03-09 07:58:53,244 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.682e+02 3.124e+02 3.807e+02 4.865e+02 2.142e+03, threshold=7.614e+02, percent-clipped=11.0 +2023-03-09 07:59:50,920 INFO [train.py:898] (0/4) Epoch 16, batch 1200, loss[loss=0.1635, simple_loss=0.2451, pruned_loss=0.04098, over 18252.00 frames. ], tot_loss[loss=0.174, simple_loss=0.261, pruned_loss=0.04345, over 3570904.85 frames. ], batch size: 45, lr: 7.13e-03, grad_scale: 8.0 +2023-03-09 08:00:34,409 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=55748.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:00:36,372 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.20 vs. limit=5.0 +2023-03-09 08:00:48,844 INFO [train.py:898] (0/4) Epoch 16, batch 1250, loss[loss=0.1604, simple_loss=0.2429, pruned_loss=0.039, over 18280.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2609, pruned_loss=0.04334, over 3573355.83 frames. ], batch size: 45, lr: 7.12e-03, grad_scale: 8.0 +2023-03-09 08:00:49,974 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.937e+02 2.771e+02 3.226e+02 3.849e+02 7.150e+02, threshold=6.452e+02, percent-clipped=0.0 +2023-03-09 08:01:30,947 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7781, 5.1360, 5.1256, 5.3934, 4.8221, 5.1693, 3.9795, 5.1521], + device='cuda:0'), covar=tensor([0.0303, 0.0616, 0.0395, 0.0357, 0.0438, 0.0340, 0.2067, 0.0379], + device='cuda:0'), in_proj_covar=tensor([0.0197, 0.0244, 0.0235, 0.0289, 0.0249, 0.0245, 0.0295, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 08:01:39,798 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55804.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:01:45,635 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55809.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:01:47,517 INFO [train.py:898] (0/4) Epoch 16, batch 1300, loss[loss=0.1528, simple_loss=0.2287, pruned_loss=0.03842, over 18383.00 frames. ], tot_loss[loss=0.1747, simple_loss=0.262, pruned_loss=0.04367, over 3578839.65 frames. ], batch size: 42, lr: 7.12e-03, grad_scale: 8.0 +2023-03-09 08:02:17,579 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6977, 2.2312, 2.5726, 2.6383, 3.2430, 4.9615, 4.6351, 3.4716], + device='cuda:0'), covar=tensor([0.1549, 0.2245, 0.2765, 0.1748, 0.2201, 0.0162, 0.0385, 0.0787], + device='cuda:0'), in_proj_covar=tensor([0.0275, 0.0331, 0.0357, 0.0265, 0.0381, 0.0219, 0.0283, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 08:02:30,123 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=55848.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:02:34,706 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=55852.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:02:45,455 INFO [train.py:898] (0/4) Epoch 16, batch 1350, loss[loss=0.1535, simple_loss=0.2418, pruned_loss=0.03262, over 18513.00 frames. ], tot_loss[loss=0.1746, simple_loss=0.2619, pruned_loss=0.0436, over 3570190.93 frames. ], batch size: 47, lr: 7.12e-03, grad_scale: 8.0 +2023-03-09 08:02:46,538 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.956e+02 2.894e+02 3.394e+02 4.145e+02 8.688e+02, threshold=6.789e+02, percent-clipped=2.0 +2023-03-09 08:02:52,173 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=55867.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:03:29,694 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8140, 2.9753, 4.3754, 3.7897, 2.8486, 4.7059, 4.0809, 2.9352], + device='cuda:0'), covar=tensor([0.0427, 0.1338, 0.0224, 0.0370, 0.1451, 0.0179, 0.0426, 0.0916], + device='cuda:0'), in_proj_covar=tensor([0.0200, 0.0228, 0.0183, 0.0150, 0.0217, 0.0199, 0.0229, 0.0192], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 08:03:42,138 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=55909.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:03:43,926 INFO [train.py:898] (0/4) Epoch 16, batch 1400, loss[loss=0.1875, simple_loss=0.272, pruned_loss=0.05154, over 16214.00 frames. ], tot_loss[loss=0.1742, simple_loss=0.2615, pruned_loss=0.04343, over 3562238.86 frames. ], batch size: 94, lr: 7.11e-03, grad_scale: 8.0 +2023-03-09 08:03:48,560 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=55915.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:04:42,842 INFO [train.py:898] (0/4) Epoch 16, batch 1450, loss[loss=0.1883, simple_loss=0.272, pruned_loss=0.05224, over 17730.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2608, pruned_loss=0.04282, over 3567582.21 frames. ], batch size: 70, lr: 7.11e-03, grad_scale: 8.0 +2023-03-09 08:04:43,988 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.940e+02 2.882e+02 3.410e+02 3.952e+02 8.442e+02, threshold=6.821e+02, percent-clipped=3.0 +2023-03-09 08:04:53,623 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0703, 4.2585, 2.5381, 4.2037, 5.3017, 2.6412, 3.9130, 4.1593], + device='cuda:0'), covar=tensor([0.0177, 0.1209, 0.1694, 0.0655, 0.0065, 0.1432, 0.0719, 0.0713], + device='cuda:0'), in_proj_covar=tensor([0.0143, 0.0253, 0.0194, 0.0189, 0.0103, 0.0176, 0.0208, 0.0217], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 08:04:56,949 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8710, 2.9606, 2.1344, 3.3030, 2.5034, 3.0562, 2.2210, 2.9070], + device='cuda:0'), covar=tensor([0.0519, 0.0663, 0.1018, 0.0589, 0.0726, 0.0335, 0.0964, 0.0469], + device='cuda:0'), in_proj_covar=tensor([0.0208, 0.0220, 0.0186, 0.0268, 0.0188, 0.0259, 0.0198, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 08:05:28,056 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-56000.pt +2023-03-09 08:05:40,396 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6452, 2.2672, 2.4664, 2.6414, 3.0070, 4.3844, 4.0790, 3.1774], + device='cuda:0'), covar=tensor([0.1560, 0.2260, 0.2729, 0.1625, 0.2172, 0.0220, 0.0498, 0.0796], + device='cuda:0'), in_proj_covar=tensor([0.0272, 0.0328, 0.0353, 0.0262, 0.0378, 0.0216, 0.0281, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 08:05:46,795 INFO [train.py:898] (0/4) Epoch 16, batch 1500, loss[loss=0.1799, simple_loss=0.2718, pruned_loss=0.044, over 18367.00 frames. ], tot_loss[loss=0.1727, simple_loss=0.2602, pruned_loss=0.04263, over 3569743.58 frames. ], batch size: 56, lr: 7.11e-03, grad_scale: 8.0 +2023-03-09 08:06:43,869 INFO [train.py:898] (0/4) Epoch 16, batch 1550, loss[loss=0.1689, simple_loss=0.2601, pruned_loss=0.03883, over 17097.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2608, pruned_loss=0.04241, over 3584691.47 frames. ], batch size: 78, lr: 7.10e-03, grad_scale: 8.0 +2023-03-09 08:06:44,922 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.796e+02 2.823e+02 3.373e+02 3.914e+02 6.631e+02, threshold=6.746e+02, percent-clipped=0.0 +2023-03-09 08:07:33,199 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56104.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:07:41,129 INFO [train.py:898] (0/4) Epoch 16, batch 1600, loss[loss=0.1549, simple_loss=0.2382, pruned_loss=0.03578, over 17562.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2617, pruned_loss=0.04294, over 3581262.47 frames. ], batch size: 39, lr: 7.10e-03, grad_scale: 8.0 +2023-03-09 08:08:21,183 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56145.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:08:39,549 INFO [train.py:898] (0/4) Epoch 16, batch 1650, loss[loss=0.1618, simple_loss=0.2528, pruned_loss=0.03544, over 18293.00 frames. ], tot_loss[loss=0.1739, simple_loss=0.2619, pruned_loss=0.04295, over 3587348.89 frames. ], batch size: 49, lr: 7.10e-03, grad_scale: 8.0 +2023-03-09 08:08:40,627 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.176e+02 2.932e+02 3.714e+02 4.558e+02 1.092e+03, threshold=7.428e+02, percent-clipped=5.0 +2023-03-09 08:08:55,943 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1423, 5.6466, 5.1916, 5.3935, 5.2433, 5.0840, 5.7041, 5.6700], + device='cuda:0'), covar=tensor([0.1098, 0.0728, 0.0609, 0.0724, 0.1322, 0.0757, 0.0557, 0.0660], + device='cuda:0'), in_proj_covar=tensor([0.0579, 0.0489, 0.0364, 0.0517, 0.0709, 0.0519, 0.0696, 0.0518], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 08:09:30,606 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56204.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:09:33,050 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56206.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:09:38,403 INFO [train.py:898] (0/4) Epoch 16, batch 1700, loss[loss=0.154, simple_loss=0.2302, pruned_loss=0.03895, over 18175.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2606, pruned_loss=0.04255, over 3591103.21 frames. ], batch size: 44, lr: 7.09e-03, grad_scale: 8.0 +2023-03-09 08:10:11,399 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3081, 4.0589, 5.2226, 2.9186, 4.7013, 2.8819, 3.3169, 2.1183], + device='cuda:0'), covar=tensor([0.0835, 0.0729, 0.0116, 0.0820, 0.0480, 0.2185, 0.2407, 0.1845], + device='cuda:0'), in_proj_covar=tensor([0.0207, 0.0227, 0.0150, 0.0181, 0.0243, 0.0256, 0.0303, 0.0223], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 08:10:23,391 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-09 08:10:36,637 INFO [train.py:898] (0/4) Epoch 16, batch 1750, loss[loss=0.1746, simple_loss=0.2642, pruned_loss=0.04248, over 18306.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2594, pruned_loss=0.04172, over 3603433.23 frames. ], batch size: 57, lr: 7.09e-03, grad_scale: 8.0 +2023-03-09 08:10:37,713 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.941e+02 2.914e+02 3.484e+02 4.158e+02 1.053e+03, threshold=6.969e+02, percent-clipped=1.0 +2023-03-09 08:10:58,211 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8609, 3.2597, 3.9719, 2.7667, 3.6249, 2.6678, 2.7877, 2.3744], + device='cuda:0'), covar=tensor([0.0817, 0.0759, 0.0198, 0.0628, 0.0623, 0.1930, 0.2039, 0.1418], + device='cuda:0'), in_proj_covar=tensor([0.0208, 0.0227, 0.0150, 0.0181, 0.0244, 0.0258, 0.0305, 0.0224], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 08:11:36,017 INFO [train.py:898] (0/4) Epoch 16, batch 1800, loss[loss=0.1767, simple_loss=0.2705, pruned_loss=0.04143, over 18380.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2594, pruned_loss=0.04185, over 3593410.93 frames. ], batch size: 50, lr: 7.09e-03, grad_scale: 8.0 +2023-03-09 08:12:34,567 INFO [train.py:898] (0/4) Epoch 16, batch 1850, loss[loss=0.1426, simple_loss=0.224, pruned_loss=0.0306, over 17668.00 frames. ], tot_loss[loss=0.173, simple_loss=0.2609, pruned_loss=0.04259, over 3585740.11 frames. ], batch size: 39, lr: 7.09e-03, grad_scale: 8.0 +2023-03-09 08:12:34,939 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7034, 3.5917, 4.9300, 4.1842, 3.0832, 2.9363, 4.2093, 5.0840], + device='cuda:0'), covar=tensor([0.0834, 0.1512, 0.0144, 0.0425, 0.1056, 0.1224, 0.0423, 0.0203], + device='cuda:0'), in_proj_covar=tensor([0.0144, 0.0265, 0.0131, 0.0175, 0.0185, 0.0186, 0.0187, 0.0177], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 08:12:35,491 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.821e+02 3.124e+02 3.834e+02 4.699e+02 1.584e+03, threshold=7.668e+02, percent-clipped=5.0 +2023-03-09 08:13:24,857 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.4911, 4.3402, 4.5320, 4.2249, 4.2122, 4.4195, 4.6728, 4.5094], + device='cuda:0'), covar=tensor([0.0117, 0.0122, 0.0101, 0.0158, 0.0116, 0.0171, 0.0101, 0.0173], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0064, 0.0067, 0.0087, 0.0070, 0.0097, 0.0081, 0.0082], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 08:13:24,872 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=56404.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:13:32,347 INFO [train.py:898] (0/4) Epoch 16, batch 1900, loss[loss=0.1737, simple_loss=0.2652, pruned_loss=0.04115, over 18393.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2617, pruned_loss=0.04297, over 3587259.65 frames. ], batch size: 52, lr: 7.08e-03, grad_scale: 8.0 +2023-03-09 08:14:19,990 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=56452.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:14:30,008 INFO [train.py:898] (0/4) Epoch 16, batch 1950, loss[loss=0.1576, simple_loss=0.2502, pruned_loss=0.03244, over 18413.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2613, pruned_loss=0.04317, over 3584568.84 frames. ], batch size: 48, lr: 7.08e-03, grad_scale: 8.0 +2023-03-09 08:14:31,033 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.810e+02 3.032e+02 3.372e+02 4.243e+02 1.557e+03, threshold=6.744e+02, percent-clipped=4.0 +2023-03-09 08:15:10,390 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8088, 3.1378, 4.4325, 3.8781, 2.8172, 4.7028, 4.1665, 3.1009], + device='cuda:0'), covar=tensor([0.0448, 0.1326, 0.0244, 0.0392, 0.1508, 0.0191, 0.0454, 0.0964], + device='cuda:0'), in_proj_covar=tensor([0.0203, 0.0232, 0.0187, 0.0154, 0.0220, 0.0203, 0.0234, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 08:15:16,853 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56501.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:15:20,430 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=56504.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:15:27,937 INFO [train.py:898] (0/4) Epoch 16, batch 2000, loss[loss=0.1747, simple_loss=0.2572, pruned_loss=0.04603, over 18550.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.261, pruned_loss=0.04305, over 3597459.08 frames. ], batch size: 49, lr: 7.08e-03, grad_scale: 8.0 +2023-03-09 08:15:42,626 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56523.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 08:16:02,080 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8537, 5.0649, 5.0686, 5.0177, 4.8959, 5.6209, 5.1678, 4.9818], + device='cuda:0'), covar=tensor([0.1140, 0.0654, 0.0760, 0.0634, 0.1423, 0.0737, 0.0673, 0.1569], + device='cuda:0'), in_proj_covar=tensor([0.0340, 0.0269, 0.0287, 0.0284, 0.0319, 0.0398, 0.0262, 0.0390], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 08:16:04,555 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8055, 4.3704, 4.4993, 3.3935, 3.6419, 3.5159, 2.5810, 2.4923], + device='cuda:0'), covar=tensor([0.0202, 0.0152, 0.0077, 0.0307, 0.0318, 0.0192, 0.0717, 0.0826], + device='cuda:0'), in_proj_covar=tensor([0.0065, 0.0053, 0.0056, 0.0064, 0.0085, 0.0062, 0.0075, 0.0081], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 08:16:16,920 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=56552.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:16:26,974 INFO [train.py:898] (0/4) Epoch 16, batch 2050, loss[loss=0.1827, simple_loss=0.2692, pruned_loss=0.04806, over 18559.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2614, pruned_loss=0.04315, over 3587862.76 frames. ], batch size: 54, lr: 7.07e-03, grad_scale: 8.0 +2023-03-09 08:16:28,115 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.109e+02 2.908e+02 3.297e+02 4.101e+02 7.290e+02, threshold=6.593e+02, percent-clipped=1.0 +2023-03-09 08:16:38,923 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-09 08:16:54,663 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56584.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 08:17:26,198 INFO [train.py:898] (0/4) Epoch 16, batch 2100, loss[loss=0.1794, simple_loss=0.2776, pruned_loss=0.04063, over 18574.00 frames. ], tot_loss[loss=0.1733, simple_loss=0.2609, pruned_loss=0.04287, over 3583652.52 frames. ], batch size: 54, lr: 7.07e-03, grad_scale: 4.0 +2023-03-09 08:18:25,142 INFO [train.py:898] (0/4) Epoch 16, batch 2150, loss[loss=0.1927, simple_loss=0.2881, pruned_loss=0.0486, over 18013.00 frames. ], tot_loss[loss=0.1722, simple_loss=0.2596, pruned_loss=0.04238, over 3596396.84 frames. ], batch size: 65, lr: 7.07e-03, grad_scale: 4.0 +2023-03-09 08:18:27,229 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.074e+02 2.907e+02 3.397e+02 4.221e+02 7.037e+02, threshold=6.794e+02, percent-clipped=2.0 +2023-03-09 08:18:41,057 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.38 vs. limit=5.0 +2023-03-09 08:19:23,204 INFO [train.py:898] (0/4) Epoch 16, batch 2200, loss[loss=0.1665, simple_loss=0.2515, pruned_loss=0.04079, over 18151.00 frames. ], tot_loss[loss=0.1723, simple_loss=0.2601, pruned_loss=0.04222, over 3605121.61 frames. ], batch size: 44, lr: 7.06e-03, grad_scale: 4.0 +2023-03-09 08:20:21,590 INFO [train.py:898] (0/4) Epoch 16, batch 2250, loss[loss=0.1938, simple_loss=0.2874, pruned_loss=0.05011, over 17121.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2608, pruned_loss=0.04277, over 3585634.91 frames. ], batch size: 78, lr: 7.06e-03, grad_scale: 4.0 +2023-03-09 08:20:23,717 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.871e+02 2.822e+02 3.228e+02 3.719e+02 7.082e+02, threshold=6.456e+02, percent-clipped=1.0 +2023-03-09 08:20:55,986 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8314, 5.3551, 5.2428, 5.3022, 4.8146, 5.2172, 4.5934, 5.1416], + device='cuda:0'), covar=tensor([0.0212, 0.0233, 0.0216, 0.0341, 0.0344, 0.0196, 0.1094, 0.0301], + device='cuda:0'), in_proj_covar=tensor([0.0196, 0.0244, 0.0242, 0.0296, 0.0252, 0.0249, 0.0298, 0.0243], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 08:21:09,315 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=56801.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:21:11,561 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56803.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:21:20,392 INFO [train.py:898] (0/4) Epoch 16, batch 2300, loss[loss=0.1749, simple_loss=0.2578, pruned_loss=0.04596, over 18542.00 frames. ], tot_loss[loss=0.1729, simple_loss=0.2606, pruned_loss=0.04262, over 3582047.33 frames. ], batch size: 49, lr: 7.06e-03, grad_scale: 4.0 +2023-03-09 08:22:04,699 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=56849.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:22:18,547 INFO [train.py:898] (0/4) Epoch 16, batch 2350, loss[loss=0.1851, simple_loss=0.2779, pruned_loss=0.04617, over 17191.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2614, pruned_loss=0.04301, over 3570918.80 frames. ], batch size: 78, lr: 7.05e-03, grad_scale: 4.0 +2023-03-09 08:22:20,836 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.203e+02 2.978e+02 3.446e+02 4.170e+02 7.854e+02, threshold=6.893e+02, percent-clipped=4.0 +2023-03-09 08:22:22,387 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56864.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:22:39,083 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=56879.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 08:22:42,997 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-09 08:23:13,901 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56908.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:23:16,988 INFO [train.py:898] (0/4) Epoch 16, batch 2400, loss[loss=0.2311, simple_loss=0.3023, pruned_loss=0.07992, over 12766.00 frames. ], tot_loss[loss=0.1736, simple_loss=0.2613, pruned_loss=0.04296, over 3563747.45 frames. ], batch size: 130, lr: 7.05e-03, grad_scale: 8.0 +2023-03-09 08:23:57,217 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8146, 5.2855, 5.2590, 5.3799, 4.7518, 5.1810, 4.3946, 5.0932], + device='cuda:0'), covar=tensor([0.0295, 0.0386, 0.0303, 0.0423, 0.0465, 0.0281, 0.1495, 0.0432], + device='cuda:0'), in_proj_covar=tensor([0.0198, 0.0247, 0.0242, 0.0297, 0.0254, 0.0250, 0.0299, 0.0244], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 08:24:15,260 INFO [train.py:898] (0/4) Epoch 16, batch 2450, loss[loss=0.1697, simple_loss=0.2592, pruned_loss=0.04009, over 17246.00 frames. ], tot_loss[loss=0.1741, simple_loss=0.2619, pruned_loss=0.04309, over 3572299.68 frames. ], batch size: 78, lr: 7.05e-03, grad_scale: 8.0 +2023-03-09 08:24:17,554 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.256e+02 2.910e+02 3.453e+02 4.291e+02 1.108e+03, threshold=6.907e+02, percent-clipped=4.0 +2023-03-09 08:24:24,710 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=56969.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:24:52,072 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56993.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:24:54,273 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=56995.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:25:13,034 INFO [train.py:898] (0/4) Epoch 16, batch 2500, loss[loss=0.1639, simple_loss=0.2507, pruned_loss=0.03855, over 18279.00 frames. ], tot_loss[loss=0.1737, simple_loss=0.2614, pruned_loss=0.04304, over 3582904.50 frames. ], batch size: 47, lr: 7.04e-03, grad_scale: 8.0 +2023-03-09 08:25:50,717 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6528, 3.0237, 4.2980, 3.6915, 2.4338, 4.5711, 3.9463, 2.7969], + device='cuda:0'), covar=tensor([0.0457, 0.1312, 0.0261, 0.0361, 0.1735, 0.0210, 0.0487, 0.0960], + device='cuda:0'), in_proj_covar=tensor([0.0203, 0.0232, 0.0191, 0.0154, 0.0220, 0.0203, 0.0236, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 08:26:03,205 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57054.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:26:05,395 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57056.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:26:11,275 INFO [train.py:898] (0/4) Epoch 16, batch 2550, loss[loss=0.187, simple_loss=0.2763, pruned_loss=0.04886, over 17942.00 frames. ], tot_loss[loss=0.1738, simple_loss=0.2615, pruned_loss=0.04306, over 3581949.76 frames. ], batch size: 65, lr: 7.04e-03, grad_scale: 8.0 +2023-03-09 08:26:13,775 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.929e+02 2.789e+02 3.538e+02 4.534e+02 7.082e+02, threshold=7.077e+02, percent-clipped=1.0 +2023-03-09 08:27:01,346 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.74 vs. limit=5.0 +2023-03-09 08:27:04,616 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7567, 3.8846, 3.7101, 3.4086, 3.6246, 3.0992, 3.0815, 3.9871], + device='cuda:0'), covar=tensor([0.0057, 0.0074, 0.0069, 0.0120, 0.0080, 0.0150, 0.0154, 0.0054], + device='cuda:0'), in_proj_covar=tensor([0.0121, 0.0140, 0.0120, 0.0173, 0.0127, 0.0165, 0.0169, 0.0104], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 08:27:09,957 INFO [train.py:898] (0/4) Epoch 16, batch 2600, loss[loss=0.1545, simple_loss=0.2444, pruned_loss=0.03227, over 18276.00 frames. ], tot_loss[loss=0.1731, simple_loss=0.2608, pruned_loss=0.04271, over 3587441.12 frames. ], batch size: 49, lr: 7.04e-03, grad_scale: 8.0 +2023-03-09 08:27:40,158 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7236, 2.5340, 2.4975, 2.5870, 2.9825, 3.7928, 3.7236, 3.3153], + device='cuda:0'), covar=tensor([0.1342, 0.1958, 0.2433, 0.1589, 0.1950, 0.0393, 0.0524, 0.0562], + device='cuda:0'), in_proj_covar=tensor([0.0276, 0.0332, 0.0358, 0.0266, 0.0381, 0.0221, 0.0285, 0.0236], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 08:27:57,732 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-09 08:28:05,192 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57159.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:28:07,950 INFO [train.py:898] (0/4) Epoch 16, batch 2650, loss[loss=0.1527, simple_loss=0.2335, pruned_loss=0.03601, over 18270.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2601, pruned_loss=0.04239, over 3580800.15 frames. ], batch size: 45, lr: 7.04e-03, grad_scale: 4.0 +2023-03-09 08:28:11,698 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.853e+02 2.763e+02 3.335e+02 4.015e+02 1.057e+03, threshold=6.669e+02, percent-clipped=2.0 +2023-03-09 08:28:20,361 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4196, 2.7991, 2.4677, 2.8475, 3.5553, 3.4906, 2.9646, 2.9437], + device='cuda:0'), covar=tensor([0.0165, 0.0247, 0.0528, 0.0381, 0.0144, 0.0140, 0.0410, 0.0342], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0125, 0.0158, 0.0147, 0.0117, 0.0105, 0.0147, 0.0143], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 08:28:29,449 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57179.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 08:28:37,445 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57186.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:28:49,745 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57197.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:29:06,036 INFO [train.py:898] (0/4) Epoch 16, batch 2700, loss[loss=0.1434, simple_loss=0.2239, pruned_loss=0.03146, over 18375.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2595, pruned_loss=0.04194, over 3583380.32 frames. ], batch size: 42, lr: 7.03e-03, grad_scale: 4.0 +2023-03-09 08:29:25,708 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=57227.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 08:29:44,858 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57244.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:29:48,216 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57247.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:30:01,243 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57258.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 08:30:04,239 INFO [train.py:898] (0/4) Epoch 16, batch 2750, loss[loss=0.1749, simple_loss=0.2673, pruned_loss=0.04122, over 18395.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2589, pruned_loss=0.04174, over 3582942.62 frames. ], batch size: 50, lr: 7.03e-03, grad_scale: 4.0 +2023-03-09 08:30:08,180 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.113e+02 2.970e+02 3.336e+02 3.947e+02 1.031e+03, threshold=6.671e+02, percent-clipped=3.0 +2023-03-09 08:30:08,381 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57264.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:30:55,625 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57305.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:31:02,570 INFO [train.py:898] (0/4) Epoch 16, batch 2800, loss[loss=0.1458, simple_loss=0.2289, pruned_loss=0.03133, over 18381.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2602, pruned_loss=0.04231, over 3582009.46 frames. ], batch size: 50, lr: 7.03e-03, grad_scale: 8.0 +2023-03-09 08:31:47,099 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57349.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:31:49,409 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57351.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:32:00,270 INFO [train.py:898] (0/4) Epoch 16, batch 2850, loss[loss=0.1575, simple_loss=0.2407, pruned_loss=0.03719, over 18490.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2607, pruned_loss=0.04241, over 3588954.39 frames. ], batch size: 47, lr: 7.02e-03, grad_scale: 8.0 +2023-03-09 08:32:04,038 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.919e+02 2.832e+02 3.405e+02 3.993e+02 9.421e+02, threshold=6.810e+02, percent-clipped=3.0 +2023-03-09 08:32:25,040 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0109, 5.0858, 5.1987, 5.2992, 5.0254, 5.7843, 5.3966, 5.1570], + device='cuda:0'), covar=tensor([0.1113, 0.0739, 0.0723, 0.0675, 0.1458, 0.0817, 0.0725, 0.1725], + device='cuda:0'), in_proj_covar=tensor([0.0340, 0.0271, 0.0288, 0.0289, 0.0318, 0.0400, 0.0263, 0.0394], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 08:32:35,145 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4246, 2.6943, 3.7285, 3.4010, 2.6113, 3.9425, 3.5652, 2.6173], + device='cuda:0'), covar=tensor([0.0445, 0.1336, 0.0330, 0.0350, 0.1439, 0.0213, 0.0536, 0.0955], + device='cuda:0'), in_proj_covar=tensor([0.0205, 0.0234, 0.0192, 0.0156, 0.0222, 0.0204, 0.0238, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 08:32:44,090 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57398.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:32:58,701 INFO [train.py:898] (0/4) Epoch 16, batch 2900, loss[loss=0.1514, simple_loss=0.2303, pruned_loss=0.03624, over 18245.00 frames. ], tot_loss[loss=0.1728, simple_loss=0.2606, pruned_loss=0.04248, over 3590492.27 frames. ], batch size: 45, lr: 7.02e-03, grad_scale: 8.0 +2023-03-09 08:33:01,400 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0223, 4.1785, 2.3767, 4.0291, 5.2746, 2.4860, 3.9068, 4.0984], + device='cuda:0'), covar=tensor([0.0156, 0.1171, 0.1646, 0.0670, 0.0060, 0.1387, 0.0669, 0.0678], + device='cuda:0'), in_proj_covar=tensor([0.0147, 0.0259, 0.0198, 0.0192, 0.0107, 0.0181, 0.0212, 0.0220], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 08:33:37,556 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7311, 3.6776, 5.0475, 4.2754, 3.2092, 2.9213, 4.3397, 5.2196], + device='cuda:0'), covar=tensor([0.0729, 0.1347, 0.0111, 0.0380, 0.0893, 0.1164, 0.0345, 0.0140], + device='cuda:0'), in_proj_covar=tensor([0.0140, 0.0259, 0.0129, 0.0172, 0.0181, 0.0182, 0.0183, 0.0174], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 08:33:42,048 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9393, 3.7661, 5.3161, 3.3770, 4.7464, 2.5536, 3.0768, 2.0518], + device='cuda:0'), covar=tensor([0.0978, 0.0851, 0.0100, 0.0587, 0.0376, 0.2398, 0.2485, 0.1879], + device='cuda:0'), in_proj_covar=tensor([0.0209, 0.0232, 0.0154, 0.0183, 0.0244, 0.0257, 0.0307, 0.0225], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 08:33:48,598 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0412, 5.1484, 5.0603, 4.8705, 4.8391, 4.8921, 5.1698, 5.2295], + device='cuda:0'), covar=tensor([0.0059, 0.0050, 0.0063, 0.0095, 0.0059, 0.0134, 0.0064, 0.0085], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0064, 0.0068, 0.0087, 0.0070, 0.0096, 0.0081, 0.0082], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-09 08:33:55,184 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57459.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:33:55,308 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57459.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:33:57,208 INFO [train.py:898] (0/4) Epoch 16, batch 2950, loss[loss=0.15, simple_loss=0.2274, pruned_loss=0.03635, over 18423.00 frames. ], tot_loss[loss=0.1721, simple_loss=0.2598, pruned_loss=0.04216, over 3589002.54 frames. ], batch size: 42, lr: 7.02e-03, grad_scale: 8.0 +2023-03-09 08:33:58,795 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57462.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:34:00,781 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.919e+02 2.681e+02 3.202e+02 3.928e+02 6.706e+02, threshold=6.405e+02, percent-clipped=1.0 +2023-03-09 08:34:24,041 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-09 08:34:38,270 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57495.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:34:51,785 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=57507.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:34:56,206 INFO [train.py:898] (0/4) Epoch 16, batch 3000, loss[loss=0.2081, simple_loss=0.3018, pruned_loss=0.05716, over 17673.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2602, pruned_loss=0.0423, over 3580488.08 frames. ], batch size: 70, lr: 7.01e-03, grad_scale: 8.0 +2023-03-09 08:34:56,209 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 08:35:07,227 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4836, 5.8514, 5.8568, 5.8285, 5.4572, 5.7775, 5.2239, 5.7682], + device='cuda:0'), covar=tensor([0.0159, 0.0182, 0.0112, 0.0333, 0.0267, 0.0166, 0.0803, 0.0263], + device='cuda:0'), in_proj_covar=tensor([0.0198, 0.0247, 0.0241, 0.0299, 0.0254, 0.0250, 0.0298, 0.0242], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 08:35:08,109 INFO [train.py:932] (0/4) Epoch 16, validation: loss=0.1522, simple_loss=0.2529, pruned_loss=0.02576, over 944034.00 frames. +2023-03-09 08:35:08,110 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 08:35:23,832 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57523.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:35:44,928 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57542.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:35:57,251 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57553.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 08:36:00,760 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57556.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:36:05,942 INFO [train.py:898] (0/4) Epoch 16, batch 3050, loss[loss=0.183, simple_loss=0.2674, pruned_loss=0.04927, over 18385.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.26, pruned_loss=0.04239, over 3568317.32 frames. ], batch size: 48, lr: 7.01e-03, grad_scale: 8.0 +2023-03-09 08:36:09,983 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.073e+02 2.813e+02 3.441e+02 4.204e+02 1.352e+03, threshold=6.882e+02, percent-clipped=6.0 +2023-03-09 08:36:10,189 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57564.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:36:19,830 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9073, 4.4128, 4.5765, 3.4462, 3.7247, 3.5796, 2.5817, 2.4683], + device='cuda:0'), covar=tensor([0.0197, 0.0165, 0.0085, 0.0298, 0.0303, 0.0196, 0.0770, 0.0877], + device='cuda:0'), in_proj_covar=tensor([0.0065, 0.0054, 0.0057, 0.0065, 0.0086, 0.0062, 0.0075, 0.0081], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 08:36:52,853 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57600.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:37:05,423 INFO [train.py:898] (0/4) Epoch 16, batch 3100, loss[loss=0.146, simple_loss=0.2298, pruned_loss=0.03105, over 18261.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2585, pruned_loss=0.04196, over 3570115.08 frames. ], batch size: 47, lr: 7.01e-03, grad_scale: 8.0 +2023-03-09 08:37:06,734 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=57612.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:37:40,965 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3830, 5.2499, 5.5716, 5.5721, 5.2162, 6.1093, 5.7010, 5.3849], + device='cuda:0'), covar=tensor([0.1045, 0.0648, 0.0775, 0.0654, 0.1605, 0.0791, 0.0668, 0.1910], + device='cuda:0'), in_proj_covar=tensor([0.0342, 0.0273, 0.0290, 0.0289, 0.0319, 0.0402, 0.0266, 0.0394], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 08:37:51,357 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57649.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:37:53,374 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57651.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:38:04,648 INFO [train.py:898] (0/4) Epoch 16, batch 3150, loss[loss=0.1504, simple_loss=0.232, pruned_loss=0.03443, over 18439.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2585, pruned_loss=0.04179, over 3573142.16 frames. ], batch size: 43, lr: 7.01e-03, grad_scale: 8.0 +2023-03-09 08:38:08,017 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.831e+02 2.982e+02 3.518e+02 4.061e+02 7.623e+02, threshold=7.037e+02, percent-clipped=2.0 +2023-03-09 08:38:36,988 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.46 vs. limit=5.0 +2023-03-09 08:38:47,967 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=57697.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:38:50,355 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=57699.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:39:03,753 INFO [train.py:898] (0/4) Epoch 16, batch 3200, loss[loss=0.1974, simple_loss=0.29, pruned_loss=0.05235, over 18135.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2594, pruned_loss=0.04189, over 3574302.34 frames. ], batch size: 62, lr: 7.00e-03, grad_scale: 8.0 +2023-03-09 08:39:05,101 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9566, 4.5370, 4.6492, 3.4316, 3.7821, 3.6072, 2.4932, 2.5190], + device='cuda:0'), covar=tensor([0.0192, 0.0132, 0.0075, 0.0286, 0.0313, 0.0189, 0.0728, 0.0849], + device='cuda:0'), in_proj_covar=tensor([0.0065, 0.0053, 0.0057, 0.0064, 0.0085, 0.0061, 0.0074, 0.0081], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 08:39:30,521 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=57733.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:39:54,224 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57754.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:40:02,254 INFO [train.py:898] (0/4) Epoch 16, batch 3250, loss[loss=0.1858, simple_loss=0.2717, pruned_loss=0.04999, over 18483.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2596, pruned_loss=0.04212, over 3569707.05 frames. ], batch size: 53, lr: 7.00e-03, grad_scale: 8.0 +2023-03-09 08:40:05,686 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.974e+02 2.787e+02 3.348e+02 4.125e+02 7.388e+02, threshold=6.695e+02, percent-clipped=1.0 +2023-03-09 08:40:41,768 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=57794.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:40:48,083 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.58 vs. limit=2.0 +2023-03-09 08:41:00,953 INFO [train.py:898] (0/4) Epoch 16, batch 3300, loss[loss=0.1632, simple_loss=0.2524, pruned_loss=0.03696, over 18537.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2597, pruned_loss=0.04211, over 3578073.02 frames. ], batch size: 49, lr: 7.00e-03, grad_scale: 8.0 +2023-03-09 08:41:02,681 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-09 08:41:03,705 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5570, 2.2233, 2.6140, 2.6343, 3.2070, 5.0450, 4.6509, 3.5235], + device='cuda:0'), covar=tensor([0.1666, 0.2284, 0.2596, 0.1703, 0.2149, 0.0144, 0.0371, 0.0760], + device='cuda:0'), in_proj_covar=tensor([0.0279, 0.0332, 0.0359, 0.0266, 0.0380, 0.0220, 0.0286, 0.0236], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 08:41:09,053 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57818.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:41:16,118 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5332, 2.2027, 2.6636, 2.7374, 3.2793, 5.0915, 4.6568, 3.5325], + device='cuda:0'), covar=tensor([0.1731, 0.2391, 0.2753, 0.1683, 0.2146, 0.0167, 0.0388, 0.0815], + device='cuda:0'), in_proj_covar=tensor([0.0279, 0.0332, 0.0359, 0.0266, 0.0381, 0.0221, 0.0286, 0.0236], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 08:41:37,735 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57842.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:41:47,803 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=57851.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:41:50,151 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57853.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 08:41:59,083 INFO [train.py:898] (0/4) Epoch 16, batch 3350, loss[loss=0.1894, simple_loss=0.2773, pruned_loss=0.05074, over 18471.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2602, pruned_loss=0.04249, over 3572211.93 frames. ], batch size: 59, lr: 6.99e-03, grad_scale: 8.0 +2023-03-09 08:42:02,564 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.007e+02 2.795e+02 3.328e+02 4.480e+02 9.325e+02, threshold=6.655e+02, percent-clipped=2.0 +2023-03-09 08:42:33,171 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=57890.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:42:44,835 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=57900.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:42:45,998 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=57901.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:42:57,107 INFO [train.py:898] (0/4) Epoch 16, batch 3400, loss[loss=0.1604, simple_loss=0.2443, pruned_loss=0.03827, over 18366.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2592, pruned_loss=0.04202, over 3580758.16 frames. ], batch size: 46, lr: 6.99e-03, grad_scale: 8.0 +2023-03-09 08:43:40,845 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=57948.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:43:55,559 INFO [train.py:898] (0/4) Epoch 16, batch 3450, loss[loss=0.1588, simple_loss=0.2541, pruned_loss=0.03177, over 18624.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2593, pruned_loss=0.04198, over 3581907.48 frames. ], batch size: 52, lr: 6.99e-03, grad_scale: 8.0 +2023-03-09 08:43:58,786 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.652e+02 2.617e+02 3.263e+02 4.009e+02 9.440e+02, threshold=6.526e+02, percent-clipped=3.0 +2023-03-09 08:44:01,663 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-09 08:44:09,454 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.46 vs. limit=2.0 +2023-03-09 08:44:18,144 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3576, 5.8636, 5.4428, 5.5976, 5.3908, 5.2882, 5.8370, 5.8494], + device='cuda:0'), covar=tensor([0.1070, 0.0664, 0.0488, 0.0708, 0.1356, 0.0692, 0.0597, 0.0644], + device='cuda:0'), in_proj_covar=tensor([0.0576, 0.0494, 0.0362, 0.0521, 0.0711, 0.0517, 0.0696, 0.0524], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 08:44:33,085 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-09 08:44:41,602 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-58000.pt +2023-03-09 08:44:59,166 INFO [train.py:898] (0/4) Epoch 16, batch 3500, loss[loss=0.2139, simple_loss=0.2959, pruned_loss=0.06594, over 18111.00 frames. ], tot_loss[loss=0.1718, simple_loss=0.2595, pruned_loss=0.04205, over 3578393.19 frames. ], batch size: 62, lr: 6.98e-03, grad_scale: 8.0 +2023-03-09 08:45:05,271 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-09 08:45:22,964 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4416, 5.9548, 5.5327, 5.7926, 5.5834, 5.4904, 6.0268, 5.9637], + device='cuda:0'), covar=tensor([0.1124, 0.0720, 0.0472, 0.0676, 0.1232, 0.0648, 0.0498, 0.0683], + device='cuda:0'), in_proj_covar=tensor([0.0579, 0.0496, 0.0363, 0.0524, 0.0712, 0.0519, 0.0699, 0.0528], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 08:45:28,432 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8930, 3.7753, 3.7264, 3.2536, 3.6359, 2.9737, 2.9433, 3.9321], + device='cuda:0'), covar=tensor([0.0053, 0.0093, 0.0071, 0.0132, 0.0081, 0.0165, 0.0173, 0.0048], + device='cuda:0'), in_proj_covar=tensor([0.0122, 0.0142, 0.0121, 0.0174, 0.0128, 0.0166, 0.0170, 0.0104], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 08:45:47,964 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58054.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:45:55,319 INFO [train.py:898] (0/4) Epoch 16, batch 3550, loss[loss=0.166, simple_loss=0.2539, pruned_loss=0.03906, over 18457.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2595, pruned_loss=0.04192, over 3589753.81 frames. ], batch size: 59, lr: 6.98e-03, grad_scale: 8.0 +2023-03-09 08:45:58,563 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.098e+02 2.999e+02 3.554e+02 4.304e+02 1.121e+03, threshold=7.108e+02, percent-clipped=3.0 +2023-03-09 08:46:26,670 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=58089.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:46:40,530 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=58102.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:46:50,129 INFO [train.py:898] (0/4) Epoch 16, batch 3600, loss[loss=0.1903, simple_loss=0.283, pruned_loss=0.04882, over 17218.00 frames. ], tot_loss[loss=0.1713, simple_loss=0.259, pruned_loss=0.04177, over 3585894.98 frames. ], batch size: 78, lr: 6.98e-03, grad_scale: 8.0 +2023-03-09 08:46:58,039 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58118.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:47:21,696 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58141.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:47:25,412 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-16.pt +2023-03-09 08:47:54,100 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58144.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:47:54,854 INFO [train.py:898] (0/4) Epoch 17, batch 0, loss[loss=0.1781, simple_loss=0.274, pruned_loss=0.04112, over 18302.00 frames. ], tot_loss[loss=0.1781, simple_loss=0.274, pruned_loss=0.04112, over 18302.00 frames. ], batch size: 54, lr: 6.77e-03, grad_scale: 8.0 +2023-03-09 08:47:54,856 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 08:48:01,230 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8427, 2.4388, 2.8004, 2.9617, 3.5904, 5.2640, 5.0508, 3.5056], + device='cuda:0'), covar=tensor([0.1587, 0.2375, 0.2527, 0.1634, 0.1921, 0.0134, 0.0272, 0.0859], + device='cuda:0'), in_proj_covar=tensor([0.0276, 0.0330, 0.0356, 0.0264, 0.0376, 0.0218, 0.0282, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 08:48:01,429 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5379, 4.8607, 4.9040, 4.7482, 4.5831, 4.5055, 4.9357, 4.9104], + device='cuda:0'), covar=tensor([0.1212, 0.0780, 0.0304, 0.0727, 0.1489, 0.0754, 0.0673, 0.0716], + device='cuda:0'), in_proj_covar=tensor([0.0578, 0.0495, 0.0361, 0.0524, 0.0711, 0.0516, 0.0696, 0.0525], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 08:48:03,421 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0245, 2.5761, 2.4626, 2.5946, 3.1976, 3.1904, 2.8951, 2.7041], + device='cuda:0'), covar=tensor([0.0218, 0.0275, 0.0574, 0.0436, 0.0233, 0.0182, 0.0421, 0.0459], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0127, 0.0158, 0.0149, 0.0115, 0.0103, 0.0146, 0.0145], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 08:48:06,682 INFO [train.py:932] (0/4) Epoch 17, validation: loss=0.1527, simple_loss=0.2537, pruned_loss=0.02582, over 944034.00 frames. +2023-03-09 08:48:06,683 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 08:48:13,717 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58151.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:48:30,194 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.965e+02 2.893e+02 3.453e+02 4.374e+02 8.967e+02, threshold=6.906e+02, percent-clipped=3.0 +2023-03-09 08:48:32,615 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=58166.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:49:05,025 INFO [train.py:898] (0/4) Epoch 17, batch 50, loss[loss=0.181, simple_loss=0.268, pruned_loss=0.04695, over 17959.00 frames. ], tot_loss[loss=0.1725, simple_loss=0.2602, pruned_loss=0.0424, over 810616.80 frames. ], batch size: 65, lr: 6.76e-03, grad_scale: 8.0 +2023-03-09 08:49:06,425 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9370, 5.2309, 2.2186, 5.1663, 4.9572, 5.2015, 4.9388, 2.2119], + device='cuda:0'), covar=tensor([0.0215, 0.0095, 0.1064, 0.0095, 0.0081, 0.0116, 0.0145, 0.1362], + device='cuda:0'), in_proj_covar=tensor([0.0082, 0.0075, 0.0092, 0.0089, 0.0081, 0.0070, 0.0081, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 08:49:09,568 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=58199.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:49:13,609 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58202.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:49:16,909 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58205.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:50:03,223 INFO [train.py:898] (0/4) Epoch 17, batch 100, loss[loss=0.1498, simple_loss=0.2345, pruned_loss=0.03254, over 18443.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2588, pruned_loss=0.04157, over 1432060.67 frames. ], batch size: 43, lr: 6.76e-03, grad_scale: 8.0 +2023-03-09 08:50:26,005 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.069e+02 2.900e+02 3.440e+02 4.043e+02 9.296e+02, threshold=6.881e+02, percent-clipped=1.0 +2023-03-09 08:50:36,379 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2573, 5.2562, 5.5070, 5.5942, 5.2356, 6.0962, 5.7744, 5.4130], + device='cuda:0'), covar=tensor([0.0993, 0.0585, 0.0664, 0.0603, 0.1308, 0.0675, 0.0582, 0.1584], + device='cuda:0'), in_proj_covar=tensor([0.0334, 0.0266, 0.0284, 0.0283, 0.0311, 0.0394, 0.0258, 0.0386], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 08:51:02,239 INFO [train.py:898] (0/4) Epoch 17, batch 150, loss[loss=0.1605, simple_loss=0.2467, pruned_loss=0.03719, over 18337.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2586, pruned_loss=0.04114, over 1916655.28 frames. ], batch size: 46, lr: 6.76e-03, grad_scale: 8.0 +2023-03-09 08:52:01,271 INFO [train.py:898] (0/4) Epoch 17, batch 200, loss[loss=0.1764, simple_loss=0.2707, pruned_loss=0.04106, over 18323.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.26, pruned_loss=0.04172, over 2285198.49 frames. ], batch size: 54, lr: 6.75e-03, grad_scale: 8.0 +2023-03-09 08:52:22,847 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.674e+02 2.879e+02 3.254e+02 3.939e+02 7.173e+02, threshold=6.508e+02, percent-clipped=1.0 +2023-03-09 08:52:53,457 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58389.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:52:59,835 INFO [train.py:898] (0/4) Epoch 17, batch 250, loss[loss=0.1644, simple_loss=0.2599, pruned_loss=0.03443, over 18576.00 frames. ], tot_loss[loss=0.1724, simple_loss=0.2607, pruned_loss=0.04207, over 2577643.99 frames. ], batch size: 54, lr: 6.75e-03, grad_scale: 8.0 +2023-03-09 08:53:01,548 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6852, 3.6351, 5.1332, 4.5164, 3.3935, 3.2741, 4.6041, 5.3700], + device='cuda:0'), covar=tensor([0.0817, 0.1733, 0.0139, 0.0368, 0.0850, 0.0991, 0.0316, 0.0193], + device='cuda:0'), in_proj_covar=tensor([0.0143, 0.0261, 0.0130, 0.0173, 0.0184, 0.0183, 0.0185, 0.0177], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 08:53:18,155 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-09 08:53:19,127 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.52 vs. limit=2.0 +2023-03-09 08:53:50,083 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=58437.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:53:59,033 INFO [train.py:898] (0/4) Epoch 17, batch 300, loss[loss=0.1644, simple_loss=0.2488, pruned_loss=0.04001, over 18488.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2583, pruned_loss=0.04164, over 2796899.04 frames. ], batch size: 44, lr: 6.75e-03, grad_scale: 8.0 +2023-03-09 08:54:03,762 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4049, 5.3406, 5.6243, 5.7166, 5.3307, 6.2521, 5.8398, 5.3879], + device='cuda:0'), covar=tensor([0.1139, 0.0637, 0.0778, 0.0708, 0.1467, 0.0776, 0.0747, 0.1703], + device='cuda:0'), in_proj_covar=tensor([0.0337, 0.0266, 0.0285, 0.0286, 0.0313, 0.0396, 0.0259, 0.0386], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 08:54:20,579 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.698e+02 2.907e+02 3.543e+02 4.450e+02 1.655e+03, threshold=7.087e+02, percent-clipped=7.0 +2023-03-09 08:54:44,034 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-09 08:54:57,844 INFO [train.py:898] (0/4) Epoch 17, batch 350, loss[loss=0.1523, simple_loss=0.2439, pruned_loss=0.03038, over 18542.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2583, pruned_loss=0.04125, over 2973646.69 frames. ], batch size: 49, lr: 6.75e-03, grad_scale: 8.0 +2023-03-09 08:55:00,144 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=58497.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:55:03,442 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=58500.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:55:56,655 INFO [train.py:898] (0/4) Epoch 17, batch 400, loss[loss=0.163, simple_loss=0.2506, pruned_loss=0.0377, over 18494.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2586, pruned_loss=0.04112, over 3110986.54 frames. ], batch size: 47, lr: 6.74e-03, grad_scale: 8.0 +2023-03-09 08:56:07,507 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-03-09 08:56:18,008 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.966e+02 2.743e+02 3.133e+02 4.342e+02 9.924e+02, threshold=6.265e+02, percent-clipped=3.0 +2023-03-09 08:56:54,598 INFO [train.py:898] (0/4) Epoch 17, batch 450, loss[loss=0.1693, simple_loss=0.2495, pruned_loss=0.04454, over 18565.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2591, pruned_loss=0.04154, over 3208429.26 frames. ], batch size: 45, lr: 6.74e-03, grad_scale: 8.0 +2023-03-09 08:56:55,934 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1566, 5.1304, 5.3843, 5.3933, 5.0195, 5.9420, 5.5199, 5.1577], + device='cuda:0'), covar=tensor([0.1006, 0.0605, 0.0689, 0.0650, 0.1440, 0.0710, 0.0574, 0.1664], + device='cuda:0'), in_proj_covar=tensor([0.0336, 0.0267, 0.0286, 0.0285, 0.0312, 0.0395, 0.0259, 0.0385], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 08:57:52,489 INFO [train.py:898] (0/4) Epoch 17, batch 500, loss[loss=0.1651, simple_loss=0.2602, pruned_loss=0.035, over 18620.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2595, pruned_loss=0.04132, over 3302737.97 frames. ], batch size: 52, lr: 6.74e-03, grad_scale: 8.0 +2023-03-09 08:58:13,821 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.815e+02 2.817e+02 3.097e+02 3.778e+02 7.071e+02, threshold=6.194e+02, percent-clipped=1.0 +2023-03-09 08:58:33,054 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58681.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:58:49,767 INFO [train.py:898] (0/4) Epoch 17, batch 550, loss[loss=0.178, simple_loss=0.2709, pruned_loss=0.04259, over 18215.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2591, pruned_loss=0.0411, over 3376754.15 frames. ], batch size: 60, lr: 6.73e-03, grad_scale: 8.0 +2023-03-09 08:59:16,715 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9943, 4.9500, 4.6167, 4.8713, 4.9423, 4.3508, 4.8219, 4.6201], + device='cuda:0'), covar=tensor([0.0412, 0.0449, 0.1226, 0.0831, 0.0502, 0.0422, 0.0410, 0.0964], + device='cuda:0'), in_proj_covar=tensor([0.0452, 0.0519, 0.0664, 0.0414, 0.0405, 0.0475, 0.0504, 0.0634], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 08:59:44,774 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58742.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 08:59:47,843 INFO [train.py:898] (0/4) Epoch 17, batch 600, loss[loss=0.1723, simple_loss=0.2617, pruned_loss=0.04145, over 18260.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2594, pruned_loss=0.04141, over 3417167.75 frames. ], batch size: 60, lr: 6.73e-03, grad_scale: 8.0 +2023-03-09 09:00:09,618 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.181e+02 2.820e+02 3.241e+02 3.843e+02 6.469e+02, threshold=6.481e+02, percent-clipped=2.0 +2023-03-09 09:00:22,457 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58775.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:00:45,231 INFO [train.py:898] (0/4) Epoch 17, batch 650, loss[loss=0.1536, simple_loss=0.2466, pruned_loss=0.03027, over 18371.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2595, pruned_loss=0.04105, over 3467553.87 frames. ], batch size: 50, lr: 6.73e-03, grad_scale: 8.0 +2023-03-09 09:00:48,191 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58797.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:00:52,636 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=58800.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:01:34,489 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58836.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:01:44,384 INFO [train.py:898] (0/4) Epoch 17, batch 700, loss[loss=0.147, simple_loss=0.2355, pruned_loss=0.02928, over 18259.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2589, pruned_loss=0.04091, over 3499919.94 frames. ], batch size: 45, lr: 6.73e-03, grad_scale: 8.0 +2023-03-09 09:01:44,568 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=58845.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:01:48,372 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=58848.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:02:07,828 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.002e+02 2.825e+02 3.273e+02 3.706e+02 6.863e+02, threshold=6.547e+02, percent-clipped=2.0 +2023-03-09 09:02:42,569 INFO [train.py:898] (0/4) Epoch 17, batch 750, loss[loss=0.1656, simple_loss=0.2514, pruned_loss=0.0399, over 18405.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.259, pruned_loss=0.04098, over 3526201.31 frames. ], batch size: 48, lr: 6.72e-03, grad_scale: 8.0 +2023-03-09 09:02:56,539 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=58906.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:03:40,647 INFO [train.py:898] (0/4) Epoch 17, batch 800, loss[loss=0.1487, simple_loss=0.2317, pruned_loss=0.03282, over 18253.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2588, pruned_loss=0.04113, over 3549842.61 frames. ], batch size: 45, lr: 6.72e-03, grad_scale: 8.0 +2023-03-09 09:04:04,577 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.057e+02 2.797e+02 3.307e+02 3.815e+02 9.263e+02, threshold=6.613e+02, percent-clipped=2.0 +2023-03-09 09:04:08,359 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=58967.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:04:20,545 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7751, 3.6523, 4.9485, 4.3068, 3.2876, 2.9501, 4.3677, 5.2518], + device='cuda:0'), covar=tensor([0.0791, 0.1422, 0.0208, 0.0406, 0.0942, 0.1174, 0.0414, 0.0186], + device='cuda:0'), in_proj_covar=tensor([0.0145, 0.0264, 0.0134, 0.0175, 0.0187, 0.0186, 0.0188, 0.0179], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 09:04:34,814 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-03-09 09:04:37,989 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1701, 4.3418, 2.4680, 4.1943, 5.3424, 2.7864, 3.9899, 4.1000], + device='cuda:0'), covar=tensor([0.0130, 0.0904, 0.1531, 0.0576, 0.0064, 0.1118, 0.0582, 0.0659], + device='cuda:0'), in_proj_covar=tensor([0.0148, 0.0258, 0.0198, 0.0190, 0.0108, 0.0178, 0.0211, 0.0216], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 09:04:38,750 INFO [train.py:898] (0/4) Epoch 17, batch 850, loss[loss=0.1636, simple_loss=0.2503, pruned_loss=0.03841, over 18356.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.259, pruned_loss=0.04089, over 3558112.46 frames. ], batch size: 46, lr: 6.72e-03, grad_scale: 8.0 +2023-03-09 09:05:28,759 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59037.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:05:37,758 INFO [train.py:898] (0/4) Epoch 17, batch 900, loss[loss=0.1448, simple_loss=0.2253, pruned_loss=0.03218, over 18145.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2582, pruned_loss=0.04083, over 3561270.17 frames. ], batch size: 44, lr: 6.71e-03, grad_scale: 8.0 +2023-03-09 09:05:49,541 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-09 09:05:59,616 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.891e+02 2.805e+02 3.225e+02 3.970e+02 5.951e+02, threshold=6.451e+02, percent-clipped=0.0 +2023-03-09 09:06:03,323 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-09 09:06:36,334 INFO [train.py:898] (0/4) Epoch 17, batch 950, loss[loss=0.1644, simple_loss=0.2559, pruned_loss=0.0364, over 17999.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2577, pruned_loss=0.04061, over 3568232.10 frames. ], batch size: 65, lr: 6.71e-03, grad_scale: 8.0 +2023-03-09 09:07:19,785 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59131.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:07:35,500 INFO [train.py:898] (0/4) Epoch 17, batch 1000, loss[loss=0.176, simple_loss=0.2701, pruned_loss=0.04093, over 16102.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2581, pruned_loss=0.04063, over 3574255.53 frames. ], batch size: 94, lr: 6.71e-03, grad_scale: 16.0 +2023-03-09 09:07:56,827 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.800e+02 2.712e+02 3.112e+02 3.820e+02 1.157e+03, threshold=6.224e+02, percent-clipped=3.0 +2023-03-09 09:07:57,121 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59164.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:08:33,588 INFO [train.py:898] (0/4) Epoch 17, batch 1050, loss[loss=0.1747, simple_loss=0.2661, pruned_loss=0.04163, over 16216.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2581, pruned_loss=0.04076, over 3581294.03 frames. ], batch size: 94, lr: 6.71e-03, grad_scale: 16.0 +2023-03-09 09:08:43,514 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2261, 4.3973, 2.5616, 4.2104, 5.3340, 2.8756, 3.9439, 4.2013], + device='cuda:0'), covar=tensor([0.0123, 0.0989, 0.1540, 0.0604, 0.0073, 0.1165, 0.0604, 0.0586], + device='cuda:0'), in_proj_covar=tensor([0.0147, 0.0256, 0.0196, 0.0189, 0.0108, 0.0176, 0.0208, 0.0215], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 09:08:51,254 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7612, 4.7324, 4.8683, 4.5812, 4.5957, 4.6103, 4.9888, 4.9725], + device='cuda:0'), covar=tensor([0.0073, 0.0071, 0.0061, 0.0110, 0.0067, 0.0159, 0.0082, 0.0106], + device='cuda:0'), in_proj_covar=tensor([0.0087, 0.0064, 0.0068, 0.0087, 0.0070, 0.0097, 0.0081, 0.0081], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-09 09:09:06,490 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.6465, 6.1646, 5.6191, 5.9822, 5.7948, 5.6125, 6.2548, 6.1413], + device='cuda:0'), covar=tensor([0.1155, 0.0732, 0.0365, 0.0649, 0.1284, 0.0688, 0.0496, 0.0682], + device='cuda:0'), in_proj_covar=tensor([0.0588, 0.0504, 0.0368, 0.0531, 0.0722, 0.0527, 0.0710, 0.0534], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 09:09:08,963 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59225.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:09:09,158 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.54 vs. limit=2.0 +2023-03-09 09:09:32,437 INFO [train.py:898] (0/4) Epoch 17, batch 1100, loss[loss=0.1585, simple_loss=0.2487, pruned_loss=0.03421, over 18248.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2586, pruned_loss=0.04071, over 3591385.39 frames. ], batch size: 45, lr: 6.70e-03, grad_scale: 16.0 +2023-03-09 09:09:51,944 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59262.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:09:54,097 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.737e+02 2.749e+02 3.337e+02 4.015e+02 7.609e+02, threshold=6.673e+02, percent-clipped=2.0 +2023-03-09 09:10:31,685 INFO [train.py:898] (0/4) Epoch 17, batch 1150, loss[loss=0.1603, simple_loss=0.2503, pruned_loss=0.03512, over 18504.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2584, pruned_loss=0.04059, over 3584180.97 frames. ], batch size: 47, lr: 6.70e-03, grad_scale: 16.0 +2023-03-09 09:11:19,732 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=59337.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:11:29,503 INFO [train.py:898] (0/4) Epoch 17, batch 1200, loss[loss=0.175, simple_loss=0.2643, pruned_loss=0.04284, over 18626.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.259, pruned_loss=0.04081, over 3587967.57 frames. ], batch size: 52, lr: 6.70e-03, grad_scale: 16.0 +2023-03-09 09:11:42,466 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6141, 3.5880, 4.9304, 4.2923, 3.2200, 3.0075, 4.2527, 5.1522], + device='cuda:0'), covar=tensor([0.0844, 0.1595, 0.0143, 0.0383, 0.0878, 0.1098, 0.0372, 0.0222], + device='cuda:0'), in_proj_covar=tensor([0.0144, 0.0262, 0.0132, 0.0173, 0.0184, 0.0184, 0.0186, 0.0177], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 09:11:51,084 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.028e+02 2.851e+02 3.293e+02 4.093e+02 6.542e+02, threshold=6.585e+02, percent-clipped=0.0 +2023-03-09 09:12:15,531 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=59385.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:12:25,249 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5249, 2.2526, 2.5754, 2.5511, 3.0338, 4.6999, 4.5661, 3.4394], + device='cuda:0'), covar=tensor([0.1611, 0.2291, 0.2657, 0.1732, 0.2251, 0.0199, 0.0355, 0.0753], + device='cuda:0'), in_proj_covar=tensor([0.0279, 0.0334, 0.0362, 0.0265, 0.0379, 0.0223, 0.0286, 0.0236], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 09:12:28,178 INFO [train.py:898] (0/4) Epoch 17, batch 1250, loss[loss=0.1897, simple_loss=0.2894, pruned_loss=0.04506, over 18612.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2594, pruned_loss=0.04095, over 3578148.45 frames. ], batch size: 52, lr: 6.69e-03, grad_scale: 8.0 +2023-03-09 09:13:05,707 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4178, 5.9421, 5.4572, 5.7676, 5.5449, 5.4796, 6.0497, 5.9907], + device='cuda:0'), covar=tensor([0.1253, 0.0841, 0.0520, 0.0744, 0.1490, 0.0671, 0.0637, 0.0680], + device='cuda:0'), in_proj_covar=tensor([0.0587, 0.0504, 0.0366, 0.0527, 0.0722, 0.0524, 0.0706, 0.0531], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 09:13:09,049 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=59431.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:13:21,810 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6907, 2.2979, 2.6188, 2.7642, 3.1821, 5.0026, 4.7602, 3.6199], + device='cuda:0'), covar=tensor([0.1605, 0.2323, 0.2821, 0.1647, 0.2372, 0.0164, 0.0379, 0.0750], + device='cuda:0'), in_proj_covar=tensor([0.0279, 0.0334, 0.0363, 0.0266, 0.0381, 0.0223, 0.0286, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 09:13:26,465 INFO [train.py:898] (0/4) Epoch 17, batch 1300, loss[loss=0.1621, simple_loss=0.2375, pruned_loss=0.04333, over 17690.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2594, pruned_loss=0.04097, over 3580067.85 frames. ], batch size: 39, lr: 6.69e-03, grad_scale: 8.0 +2023-03-09 09:13:48,930 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.812e+02 2.789e+02 3.310e+02 4.090e+02 6.733e+02, threshold=6.621e+02, percent-clipped=2.0 +2023-03-09 09:14:05,083 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=59479.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:14:06,270 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3918, 5.9226, 5.3993, 5.7418, 5.4820, 5.4801, 6.0038, 5.9513], + device='cuda:0'), covar=tensor([0.1152, 0.0737, 0.0446, 0.0669, 0.1365, 0.0611, 0.0570, 0.0641], + device='cuda:0'), in_proj_covar=tensor([0.0585, 0.0502, 0.0364, 0.0524, 0.0721, 0.0523, 0.0704, 0.0530], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 09:14:24,100 INFO [train.py:898] (0/4) Epoch 17, batch 1350, loss[loss=0.1571, simple_loss=0.237, pruned_loss=0.0386, over 18397.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2599, pruned_loss=0.04115, over 3580038.32 frames. ], batch size: 48, lr: 6.69e-03, grad_scale: 8.0 +2023-03-09 09:14:33,497 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59502.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:14:47,702 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-09 09:14:53,671 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59520.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:15:17,221 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.14 vs. limit=5.0 +2023-03-09 09:15:22,940 INFO [train.py:898] (0/4) Epoch 17, batch 1400, loss[loss=0.152, simple_loss=0.2322, pruned_loss=0.03589, over 18127.00 frames. ], tot_loss[loss=0.172, simple_loss=0.2606, pruned_loss=0.04174, over 3575936.39 frames. ], batch size: 44, lr: 6.69e-03, grad_scale: 8.0 +2023-03-09 09:15:43,344 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=59562.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:15:44,550 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59563.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:15:46,381 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.823e+02 2.959e+02 3.481e+02 4.444e+02 9.729e+02, threshold=6.962e+02, percent-clipped=6.0 +2023-03-09 09:15:58,815 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9849, 5.1106, 5.0678, 4.7428, 4.8010, 4.8287, 5.1738, 5.2318], + device='cuda:0'), covar=tensor([0.0071, 0.0054, 0.0054, 0.0113, 0.0058, 0.0145, 0.0080, 0.0088], + device='cuda:0'), in_proj_covar=tensor([0.0087, 0.0064, 0.0068, 0.0087, 0.0070, 0.0097, 0.0081, 0.0081], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0003], + device='cuda:0') +2023-03-09 09:16:21,020 INFO [train.py:898] (0/4) Epoch 17, batch 1450, loss[loss=0.1886, simple_loss=0.2724, pruned_loss=0.05245, over 18320.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2599, pruned_loss=0.04148, over 3581903.54 frames. ], batch size: 56, lr: 6.68e-03, grad_scale: 8.0 +2023-03-09 09:16:40,247 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=59610.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:16:47,369 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5391, 3.3365, 2.0920, 4.2942, 2.9967, 4.2262, 2.3833, 3.6613], + device='cuda:0'), covar=tensor([0.0611, 0.0829, 0.1523, 0.0488, 0.0875, 0.0292, 0.1249, 0.0488], + device='cuda:0'), in_proj_covar=tensor([0.0208, 0.0218, 0.0187, 0.0270, 0.0189, 0.0263, 0.0201, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 09:17:20,443 INFO [train.py:898] (0/4) Epoch 17, batch 1500, loss[loss=0.1607, simple_loss=0.2396, pruned_loss=0.04088, over 18279.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2584, pruned_loss=0.04107, over 3590371.02 frames. ], batch size: 47, lr: 6.68e-03, grad_scale: 8.0 +2023-03-09 09:17:44,116 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.960e+02 2.886e+02 3.557e+02 4.310e+02 1.324e+03, threshold=7.115e+02, percent-clipped=4.0 +2023-03-09 09:17:44,432 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4021, 5.9050, 5.4198, 5.6475, 5.4466, 5.4110, 5.9388, 5.9109], + device='cuda:0'), covar=tensor([0.1230, 0.0759, 0.0495, 0.0785, 0.1685, 0.0687, 0.0690, 0.0668], + device='cuda:0'), in_proj_covar=tensor([0.0585, 0.0498, 0.0365, 0.0524, 0.0722, 0.0521, 0.0707, 0.0531], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 09:17:52,728 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-09 09:18:18,453 INFO [train.py:898] (0/4) Epoch 17, batch 1550, loss[loss=0.1503, simple_loss=0.2275, pruned_loss=0.03659, over 17657.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2584, pruned_loss=0.04075, over 3589992.95 frames. ], batch size: 39, lr: 6.68e-03, grad_scale: 8.0 +2023-03-09 09:19:16,813 INFO [train.py:898] (0/4) Epoch 17, batch 1600, loss[loss=0.1682, simple_loss=0.2602, pruned_loss=0.03813, over 18453.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2585, pruned_loss=0.0406, over 3592789.05 frames. ], batch size: 59, lr: 6.67e-03, grad_scale: 8.0 +2023-03-09 09:19:41,443 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.908e+02 2.635e+02 3.092e+02 3.637e+02 7.676e+02, threshold=6.183e+02, percent-clipped=1.0 +2023-03-09 09:20:15,865 INFO [train.py:898] (0/4) Epoch 17, batch 1650, loss[loss=0.1609, simple_loss=0.2487, pruned_loss=0.03653, over 18534.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2588, pruned_loss=0.04091, over 3598829.84 frames. ], batch size: 49, lr: 6.67e-03, grad_scale: 8.0 +2023-03-09 09:20:46,637 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=59820.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:20:46,894 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.47 vs. limit=2.0 +2023-03-09 09:20:50,413 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-03-09 09:21:14,486 INFO [train.py:898] (0/4) Epoch 17, batch 1700, loss[loss=0.1617, simple_loss=0.2542, pruned_loss=0.03457, over 18400.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2592, pruned_loss=0.04109, over 3596782.14 frames. ], batch size: 52, lr: 6.67e-03, grad_scale: 8.0 +2023-03-09 09:21:18,266 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59848.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:21:18,474 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.74 vs. limit=5.0 +2023-03-09 09:21:30,977 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=59858.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:21:38,855 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.932e+02 2.816e+02 3.196e+02 3.802e+02 1.399e+03, threshold=6.391e+02, percent-clipped=5.0 +2023-03-09 09:21:43,456 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=59868.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:22:07,156 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=59889.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:22:13,214 INFO [train.py:898] (0/4) Epoch 17, batch 1750, loss[loss=0.2117, simple_loss=0.2942, pruned_loss=0.06459, over 18376.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2597, pruned_loss=0.04125, over 3596715.80 frames. ], batch size: 56, lr: 6.67e-03, grad_scale: 8.0 +2023-03-09 09:22:18,180 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7610, 3.9075, 5.1878, 4.6477, 3.2922, 3.1140, 4.5710, 5.4162], + device='cuda:0'), covar=tensor([0.0797, 0.1513, 0.0165, 0.0332, 0.0890, 0.1062, 0.0325, 0.0177], + device='cuda:0'), in_proj_covar=tensor([0.0143, 0.0263, 0.0134, 0.0174, 0.0184, 0.0184, 0.0186, 0.0179], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 09:22:19,175 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7835, 2.9328, 2.6734, 3.0176, 3.7466, 3.6111, 3.3085, 3.0340], + device='cuda:0'), covar=tensor([0.0172, 0.0220, 0.0507, 0.0300, 0.0132, 0.0136, 0.0288, 0.0293], + device='cuda:0'), in_proj_covar=tensor([0.0132, 0.0126, 0.0160, 0.0150, 0.0118, 0.0105, 0.0147, 0.0146], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 09:22:30,592 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59909.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:22:32,893 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9607, 3.8310, 3.7270, 3.3526, 3.6638, 3.0187, 2.9635, 3.8144], + device='cuda:0'), covar=tensor([0.0041, 0.0073, 0.0068, 0.0110, 0.0067, 0.0162, 0.0181, 0.0064], + device='cuda:0'), in_proj_covar=tensor([0.0123, 0.0145, 0.0124, 0.0177, 0.0130, 0.0168, 0.0172, 0.0109], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 09:22:54,006 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7482, 5.3492, 5.3174, 5.3467, 4.7881, 5.1913, 4.6009, 5.1802], + device='cuda:0'), covar=tensor([0.0293, 0.0302, 0.0215, 0.0465, 0.0455, 0.0243, 0.1259, 0.0373], + device='cuda:0'), in_proj_covar=tensor([0.0203, 0.0249, 0.0241, 0.0304, 0.0259, 0.0252, 0.0298, 0.0247], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 09:23:10,882 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9074, 4.6044, 4.7315, 3.5819, 3.8896, 3.5535, 2.8337, 2.7937], + device='cuda:0'), covar=tensor([0.0183, 0.0158, 0.0086, 0.0282, 0.0311, 0.0201, 0.0699, 0.0775], + device='cuda:0'), in_proj_covar=tensor([0.0066, 0.0055, 0.0058, 0.0065, 0.0087, 0.0063, 0.0076, 0.0082], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 09:23:11,677 INFO [train.py:898] (0/4) Epoch 17, batch 1800, loss[loss=0.1459, simple_loss=0.2323, pruned_loss=0.02975, over 18458.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2593, pruned_loss=0.04089, over 3605673.45 frames. ], batch size: 44, lr: 6.66e-03, grad_scale: 8.0 +2023-03-09 09:23:17,834 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=59950.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:23:35,353 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.961e+02 2.799e+02 3.065e+02 3.642e+02 5.911e+02, threshold=6.130e+02, percent-clipped=0.0 +2023-03-09 09:24:10,360 INFO [train.py:898] (0/4) Epoch 17, batch 1850, loss[loss=0.1459, simple_loss=0.227, pruned_loss=0.03238, over 18439.00 frames. ], tot_loss[loss=0.1707, simple_loss=0.2593, pruned_loss=0.04109, over 3594027.00 frames. ], batch size: 43, lr: 6.66e-03, grad_scale: 8.0 +2023-03-09 09:24:16,336 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-60000.pt +2023-03-09 09:25:08,339 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60040.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:25:13,959 INFO [train.py:898] (0/4) Epoch 17, batch 1900, loss[loss=0.178, simple_loss=0.2651, pruned_loss=0.04549, over 18626.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2596, pruned_loss=0.04124, over 3600349.30 frames. ], batch size: 52, lr: 6.66e-03, grad_scale: 8.0 +2023-03-09 09:25:14,431 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9530, 5.2264, 2.5755, 5.1213, 4.9057, 5.2563, 5.0464, 2.6429], + device='cuda:0'), covar=tensor([0.0187, 0.0067, 0.0784, 0.0064, 0.0077, 0.0069, 0.0091, 0.0992], + device='cuda:0'), in_proj_covar=tensor([0.0084, 0.0077, 0.0093, 0.0091, 0.0083, 0.0072, 0.0083, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 09:25:22,411 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60052.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:25:37,666 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.972e+02 2.751e+02 3.346e+02 4.320e+02 1.006e+03, threshold=6.692e+02, percent-clipped=5.0 +2023-03-09 09:25:38,020 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60065.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:25:53,045 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60078.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:26:12,278 INFO [train.py:898] (0/4) Epoch 17, batch 1950, loss[loss=0.1782, simple_loss=0.2729, pruned_loss=0.04175, over 18261.00 frames. ], tot_loss[loss=0.1716, simple_loss=0.2602, pruned_loss=0.04155, over 3596751.34 frames. ], batch size: 57, lr: 6.66e-03, grad_scale: 8.0 +2023-03-09 09:26:13,723 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.4778, 4.9153, 4.8837, 4.9567, 4.4690, 4.7912, 4.3259, 4.8126], + device='cuda:0'), covar=tensor([0.0218, 0.0283, 0.0199, 0.0359, 0.0339, 0.0215, 0.0922, 0.0312], + device='cuda:0'), in_proj_covar=tensor([0.0200, 0.0246, 0.0238, 0.0299, 0.0256, 0.0249, 0.0295, 0.0243], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 09:26:19,351 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60101.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 09:26:33,448 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60113.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 09:26:48,299 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60126.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:27:04,087 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60139.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:27:08,894 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-09 09:27:10,626 INFO [train.py:898] (0/4) Epoch 17, batch 2000, loss[loss=0.1674, simple_loss=0.2606, pruned_loss=0.0371, over 18572.00 frames. ], tot_loss[loss=0.1712, simple_loss=0.2595, pruned_loss=0.04142, over 3593638.91 frames. ], batch size: 54, lr: 6.65e-03, grad_scale: 8.0 +2023-03-09 09:27:25,203 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60158.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:27:33,337 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.223e+02 2.887e+02 3.368e+02 4.161e+02 9.381e+02, threshold=6.736e+02, percent-clipped=4.0 +2023-03-09 09:27:33,893 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.39 vs. limit=5.0 +2023-03-09 09:27:51,409 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7140, 2.8202, 2.6048, 3.0007, 3.6335, 3.5523, 3.2489, 2.9588], + device='cuda:0'), covar=tensor([0.0162, 0.0254, 0.0555, 0.0359, 0.0163, 0.0144, 0.0348, 0.0408], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0126, 0.0160, 0.0150, 0.0119, 0.0106, 0.0148, 0.0147], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 09:28:07,197 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6313, 2.5351, 2.5526, 2.4501, 2.5278, 2.2180, 2.2974, 2.6027], + device='cuda:0'), covar=tensor([0.0060, 0.0095, 0.0074, 0.0108, 0.0084, 0.0138, 0.0147, 0.0066], + device='cuda:0'), in_proj_covar=tensor([0.0124, 0.0147, 0.0125, 0.0179, 0.0131, 0.0170, 0.0174, 0.0110], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 09:28:08,977 INFO [train.py:898] (0/4) Epoch 17, batch 2050, loss[loss=0.1527, simple_loss=0.2341, pruned_loss=0.03564, over 18420.00 frames. ], tot_loss[loss=0.1714, simple_loss=0.2598, pruned_loss=0.04143, over 3586500.98 frames. ], batch size: 42, lr: 6.65e-03, grad_scale: 8.0 +2023-03-09 09:28:09,542 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-09 09:28:12,761 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60198.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:28:19,672 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60204.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:28:21,877 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=60206.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:28:28,887 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0877, 5.1173, 5.3920, 5.3694, 5.0996, 5.9599, 5.5341, 5.3152], + device='cuda:0'), covar=tensor([0.1389, 0.0667, 0.0785, 0.0821, 0.1556, 0.0814, 0.0694, 0.1721], + device='cuda:0'), in_proj_covar=tensor([0.0339, 0.0274, 0.0294, 0.0291, 0.0322, 0.0405, 0.0264, 0.0394], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 09:28:56,400 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0828, 3.7521, 5.1295, 3.0047, 4.5612, 2.7856, 3.2193, 1.8147], + device='cuda:0'), covar=tensor([0.0986, 0.0840, 0.0115, 0.0827, 0.0490, 0.2264, 0.2575, 0.1993], + device='cuda:0'), in_proj_covar=tensor([0.0215, 0.0233, 0.0160, 0.0187, 0.0246, 0.0260, 0.0312, 0.0225], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 09:28:58,144 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7910, 3.1698, 3.9279, 2.8819, 3.6499, 2.6393, 2.7970, 2.2615], + device='cuda:0'), covar=tensor([0.0930, 0.0834, 0.0244, 0.0666, 0.0640, 0.1931, 0.2217, 0.1517], + device='cuda:0'), in_proj_covar=tensor([0.0215, 0.0233, 0.0160, 0.0187, 0.0246, 0.0260, 0.0312, 0.0225], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 09:29:07,877 INFO [train.py:898] (0/4) Epoch 17, batch 2100, loss[loss=0.1811, simple_loss=0.2696, pruned_loss=0.04627, over 18302.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.26, pruned_loss=0.04154, over 3586569.61 frames. ], batch size: 54, lr: 6.65e-03, grad_scale: 8.0 +2023-03-09 09:29:08,080 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60245.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:29:24,076 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60259.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:29:30,556 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.013e+02 2.803e+02 3.288e+02 3.914e+02 1.145e+03, threshold=6.576e+02, percent-clipped=2.0 +2023-03-09 09:30:07,000 INFO [train.py:898] (0/4) Epoch 17, batch 2150, loss[loss=0.1745, simple_loss=0.2637, pruned_loss=0.04262, over 17097.00 frames. ], tot_loss[loss=0.1709, simple_loss=0.2594, pruned_loss=0.04122, over 3587676.61 frames. ], batch size: 78, lr: 6.64e-03, grad_scale: 8.0 +2023-03-09 09:30:19,963 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.94 vs. limit=5.0 +2023-03-09 09:31:03,786 INFO [train.py:898] (0/4) Epoch 17, batch 2200, loss[loss=0.1611, simple_loss=0.2499, pruned_loss=0.03616, over 18503.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2581, pruned_loss=0.04079, over 3592028.26 frames. ], batch size: 47, lr: 6.64e-03, grad_scale: 8.0 +2023-03-09 09:31:26,299 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.966e+02 2.999e+02 3.869e+02 4.912e+02 1.337e+03, threshold=7.738e+02, percent-clipped=7.0 +2023-03-09 09:32:01,255 INFO [train.py:898] (0/4) Epoch 17, batch 2250, loss[loss=0.1572, simple_loss=0.2544, pruned_loss=0.03003, over 18335.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.2581, pruned_loss=0.04073, over 3595166.59 frames. ], batch size: 55, lr: 6.64e-03, grad_scale: 4.0 +2023-03-09 09:32:02,662 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60396.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 09:32:08,101 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.6086, 6.0684, 5.5351, 5.8522, 5.6357, 5.5052, 6.1340, 6.0776], + device='cuda:0'), covar=tensor([0.1129, 0.0772, 0.0420, 0.0681, 0.1444, 0.0688, 0.0621, 0.0687], + device='cuda:0'), in_proj_covar=tensor([0.0585, 0.0502, 0.0366, 0.0521, 0.0717, 0.0525, 0.0705, 0.0530], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 09:32:13,010 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-09 09:32:16,900 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60408.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 09:32:31,460 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60421.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:32:36,533 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.16 vs. limit=5.0 +2023-03-09 09:32:46,534 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60434.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:32:58,979 INFO [train.py:898] (0/4) Epoch 17, batch 2300, loss[loss=0.1762, simple_loss=0.2616, pruned_loss=0.04542, over 18490.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2579, pruned_loss=0.04065, over 3601404.50 frames. ], batch size: 51, lr: 6.64e-03, grad_scale: 4.0 +2023-03-09 09:32:59,312 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2109, 5.4731, 3.1585, 5.3550, 5.1910, 5.5340, 5.3917, 3.1375], + device='cuda:0'), covar=tensor([0.0156, 0.0070, 0.0620, 0.0063, 0.0071, 0.0064, 0.0073, 0.0741], + device='cuda:0'), in_proj_covar=tensor([0.0084, 0.0078, 0.0094, 0.0092, 0.0084, 0.0072, 0.0083, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 09:33:16,353 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-09 09:33:23,540 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.591e+02 2.714e+02 3.247e+02 3.774e+02 7.565e+02, threshold=6.493e+02, percent-clipped=0.0 +2023-03-09 09:33:55,696 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9648, 5.0554, 5.1502, 4.8236, 4.8851, 4.8265, 5.1761, 5.2158], + device='cuda:0'), covar=tensor([0.0075, 0.0065, 0.0055, 0.0098, 0.0058, 0.0169, 0.0072, 0.0098], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0064, 0.0069, 0.0089, 0.0070, 0.0098, 0.0081, 0.0083], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 09:33:57,582 INFO [train.py:898] (0/4) Epoch 17, batch 2350, loss[loss=0.1877, simple_loss=0.2802, pruned_loss=0.04754, over 18357.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2583, pruned_loss=0.04089, over 3595584.75 frames. ], batch size: 55, lr: 6.63e-03, grad_scale: 4.0 +2023-03-09 09:34:09,152 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60504.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:34:56,307 INFO [train.py:898] (0/4) Epoch 17, batch 2400, loss[loss=0.1616, simple_loss=0.2587, pruned_loss=0.03228, over 18349.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2588, pruned_loss=0.04119, over 3586266.05 frames. ], batch size: 55, lr: 6.63e-03, grad_scale: 8.0 +2023-03-09 09:34:56,563 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60545.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:35:05,015 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=60552.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:35:07,282 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=60554.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:35:20,245 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.011e+02 2.942e+02 3.643e+02 4.208e+02 9.657e+02, threshold=7.287e+02, percent-clipped=4.0 +2023-03-09 09:35:51,986 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=60593.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:35:54,157 INFO [train.py:898] (0/4) Epoch 17, batch 2450, loss[loss=0.1531, simple_loss=0.241, pruned_loss=0.03264, over 18407.00 frames. ], tot_loss[loss=0.1715, simple_loss=0.2599, pruned_loss=0.04154, over 3587680.73 frames. ], batch size: 48, lr: 6.63e-03, grad_scale: 8.0 +2023-03-09 09:36:23,273 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9622, 4.9945, 5.0345, 4.7781, 4.7958, 4.8085, 5.1314, 5.1398], + device='cuda:0'), covar=tensor([0.0068, 0.0064, 0.0063, 0.0100, 0.0057, 0.0148, 0.0064, 0.0088], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0065, 0.0069, 0.0089, 0.0071, 0.0099, 0.0082, 0.0084], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 09:36:26,482 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5353, 6.0366, 5.5198, 5.7705, 5.5925, 5.4581, 6.0840, 6.0224], + device='cuda:0'), covar=tensor([0.1317, 0.0697, 0.0477, 0.0769, 0.1508, 0.0729, 0.0603, 0.0660], + device='cuda:0'), in_proj_covar=tensor([0.0588, 0.0502, 0.0367, 0.0522, 0.0718, 0.0525, 0.0707, 0.0530], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 09:36:45,534 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5409, 3.3720, 2.3248, 4.2743, 3.1796, 4.1961, 2.3552, 3.9771], + device='cuda:0'), covar=tensor([0.0604, 0.0873, 0.1459, 0.0538, 0.0841, 0.0353, 0.1302, 0.0386], + device='cuda:0'), in_proj_covar=tensor([0.0209, 0.0219, 0.0189, 0.0274, 0.0189, 0.0263, 0.0201, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 09:36:52,594 INFO [train.py:898] (0/4) Epoch 17, batch 2500, loss[loss=0.171, simple_loss=0.2647, pruned_loss=0.03862, over 18634.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.259, pruned_loss=0.04085, over 3591014.52 frames. ], batch size: 52, lr: 6.63e-03, grad_scale: 8.0 +2023-03-09 09:37:17,216 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.797e+02 2.798e+02 3.205e+02 3.786e+02 6.512e+02, threshold=6.411e+02, percent-clipped=0.0 +2023-03-09 09:37:26,830 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-03-09 09:37:51,493 INFO [train.py:898] (0/4) Epoch 17, batch 2550, loss[loss=0.1702, simple_loss=0.2632, pruned_loss=0.03854, over 18500.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2594, pruned_loss=0.04088, over 3577035.61 frames. ], batch size: 51, lr: 6.62e-03, grad_scale: 8.0 +2023-03-09 09:37:53,017 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60696.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:38:06,201 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60708.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:38:18,154 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=60718.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:38:21,252 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60721.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:38:36,064 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60734.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:38:38,435 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6520, 2.4981, 2.4949, 2.6805, 3.0449, 3.7687, 3.6791, 3.1061], + device='cuda:0'), covar=tensor([0.1473, 0.1960, 0.2673, 0.1589, 0.1853, 0.0384, 0.0591, 0.0694], + device='cuda:0'), in_proj_covar=tensor([0.0283, 0.0336, 0.0366, 0.0269, 0.0384, 0.0226, 0.0290, 0.0241], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 09:38:47,257 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=60744.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:38:48,802 INFO [train.py:898] (0/4) Epoch 17, batch 2600, loss[loss=0.1872, simple_loss=0.2771, pruned_loss=0.04865, over 18259.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.259, pruned_loss=0.04088, over 3577021.12 frames. ], batch size: 60, lr: 6.62e-03, grad_scale: 8.0 +2023-03-09 09:38:59,940 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6174, 2.9787, 2.3266, 3.0913, 3.6581, 3.5563, 3.2366, 3.0070], + device='cuda:0'), covar=tensor([0.0235, 0.0237, 0.0669, 0.0284, 0.0154, 0.0133, 0.0321, 0.0335], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0129, 0.0161, 0.0153, 0.0122, 0.0108, 0.0151, 0.0149], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 09:39:01,921 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=60756.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:39:13,462 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.006e+02 2.726e+02 3.213e+02 3.687e+02 6.855e+02, threshold=6.427e+02, percent-clipped=2.0 +2023-03-09 09:39:16,909 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=60769.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:39:28,270 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=60779.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:39:31,132 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=60782.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:39:45,262 INFO [train.py:898] (0/4) Epoch 17, batch 2650, loss[loss=0.183, simple_loss=0.2713, pruned_loss=0.04738, over 18326.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2589, pruned_loss=0.04087, over 3584547.84 frames. ], batch size: 57, lr: 6.62e-03, grad_scale: 8.0 +2023-03-09 09:39:51,035 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2755, 5.8591, 5.3925, 5.5883, 5.4420, 5.2307, 5.8777, 5.8957], + device='cuda:0'), covar=tensor([0.1108, 0.0689, 0.0489, 0.0702, 0.1332, 0.0677, 0.0544, 0.0578], + device='cuda:0'), in_proj_covar=tensor([0.0581, 0.0502, 0.0362, 0.0519, 0.0712, 0.0519, 0.0704, 0.0526], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 09:40:27,349 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.12 vs. limit=5.0 +2023-03-09 09:40:43,824 INFO [train.py:898] (0/4) Epoch 17, batch 2700, loss[loss=0.2071, simple_loss=0.2871, pruned_loss=0.06356, over 12364.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2588, pruned_loss=0.04092, over 3589171.35 frames. ], batch size: 130, lr: 6.61e-03, grad_scale: 8.0 +2023-03-09 09:40:54,805 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=60854.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:41:08,389 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.851e+02 2.950e+02 3.306e+02 4.033e+02 9.458e+02, threshold=6.612e+02, percent-clipped=5.0 +2023-03-09 09:41:42,538 INFO [train.py:898] (0/4) Epoch 17, batch 2750, loss[loss=0.1685, simple_loss=0.2614, pruned_loss=0.03782, over 18565.00 frames. ], tot_loss[loss=0.1705, simple_loss=0.2591, pruned_loss=0.04098, over 3578445.19 frames. ], batch size: 54, lr: 6.61e-03, grad_scale: 8.0 +2023-03-09 09:41:50,479 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=60902.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:42:39,346 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7188, 5.2761, 5.2573, 5.2518, 4.7739, 5.1333, 4.5509, 5.1306], + device='cuda:0'), covar=tensor([0.0264, 0.0303, 0.0186, 0.0467, 0.0397, 0.0220, 0.1133, 0.0363], + device='cuda:0'), in_proj_covar=tensor([0.0204, 0.0250, 0.0242, 0.0306, 0.0258, 0.0253, 0.0297, 0.0247], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 09:42:41,194 INFO [train.py:898] (0/4) Epoch 17, batch 2800, loss[loss=0.1528, simple_loss=0.2415, pruned_loss=0.03206, over 18365.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2587, pruned_loss=0.0409, over 3584147.56 frames. ], batch size: 50, lr: 6.61e-03, grad_scale: 8.0 +2023-03-09 09:43:06,400 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.038e+02 2.738e+02 3.235e+02 3.720e+02 7.893e+02, threshold=6.471e+02, percent-clipped=2.0 +2023-03-09 09:43:16,083 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-09 09:43:40,112 INFO [train.py:898] (0/4) Epoch 17, batch 2850, loss[loss=0.1749, simple_loss=0.2614, pruned_loss=0.04416, over 18495.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.258, pruned_loss=0.04072, over 3582815.52 frames. ], batch size: 53, lr: 6.61e-03, grad_scale: 8.0 +2023-03-09 09:43:53,328 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5422, 3.0935, 3.8568, 3.6441, 3.0138, 2.8825, 3.5504, 3.9830], + device='cuda:0'), covar=tensor([0.0774, 0.1145, 0.0252, 0.0407, 0.0869, 0.1036, 0.0455, 0.0300], + device='cuda:0'), in_proj_covar=tensor([0.0145, 0.0265, 0.0137, 0.0176, 0.0186, 0.0187, 0.0188, 0.0181], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 09:44:27,372 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7077, 3.6509, 3.5327, 3.0941, 3.4104, 2.8320, 2.6625, 3.6925], + device='cuda:0'), covar=tensor([0.0063, 0.0086, 0.0075, 0.0132, 0.0085, 0.0164, 0.0196, 0.0065], + device='cuda:0'), in_proj_covar=tensor([0.0127, 0.0150, 0.0129, 0.0180, 0.0133, 0.0172, 0.0176, 0.0111], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 09:44:39,449 INFO [train.py:898] (0/4) Epoch 17, batch 2900, loss[loss=0.1535, simple_loss=0.2414, pruned_loss=0.03275, over 18309.00 frames. ], tot_loss[loss=0.1706, simple_loss=0.2592, pruned_loss=0.04106, over 3565980.21 frames. ], batch size: 49, lr: 6.60e-03, grad_scale: 8.0 +2023-03-09 09:44:57,711 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61061.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:45:04,193 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.725e+02 2.779e+02 3.469e+02 4.268e+02 9.879e+02, threshold=6.938e+02, percent-clipped=3.0 +2023-03-09 09:45:13,417 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=61074.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:45:34,618 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1358, 3.0069, 3.0201, 2.7502, 2.9319, 2.4273, 2.4587, 3.1356], + device='cuda:0'), covar=tensor([0.0068, 0.0102, 0.0078, 0.0133, 0.0096, 0.0178, 0.0181, 0.0059], + device='cuda:0'), in_proj_covar=tensor([0.0126, 0.0149, 0.0128, 0.0179, 0.0132, 0.0172, 0.0175, 0.0111], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 09:45:37,772 INFO [train.py:898] (0/4) Epoch 17, batch 2950, loss[loss=0.1766, simple_loss=0.2712, pruned_loss=0.04104, over 18215.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2592, pruned_loss=0.0412, over 3555759.21 frames. ], batch size: 60, lr: 6.60e-03, grad_scale: 4.0 +2023-03-09 09:45:59,918 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0972, 5.1229, 5.3069, 5.3790, 5.0872, 5.9107, 5.5428, 5.1642], + device='cuda:0'), covar=tensor([0.1135, 0.0690, 0.0818, 0.0780, 0.1614, 0.0731, 0.0678, 0.1572], + device='cuda:0'), in_proj_covar=tensor([0.0332, 0.0266, 0.0289, 0.0285, 0.0314, 0.0396, 0.0258, 0.0385], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 09:46:09,463 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61122.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:46:36,107 INFO [train.py:898] (0/4) Epoch 17, batch 3000, loss[loss=0.1485, simple_loss=0.2298, pruned_loss=0.03357, over 17614.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2592, pruned_loss=0.04119, over 3568507.11 frames. ], batch size: 39, lr: 6.60e-03, grad_scale: 4.0 +2023-03-09 09:46:36,109 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 09:46:48,252 INFO [train.py:932] (0/4) Epoch 17, validation: loss=0.1521, simple_loss=0.2525, pruned_loss=0.02589, over 944034.00 frames. +2023-03-09 09:46:48,253 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 09:47:08,031 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5608, 2.1896, 2.5444, 2.6583, 3.1677, 4.9394, 4.6266, 3.5148], + device='cuda:0'), covar=tensor([0.1627, 0.2407, 0.2780, 0.1717, 0.2212, 0.0172, 0.0400, 0.0807], + device='cuda:0'), in_proj_covar=tensor([0.0284, 0.0336, 0.0368, 0.0269, 0.0385, 0.0226, 0.0289, 0.0241], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 09:47:14,275 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.190e+02 3.156e+02 3.810e+02 4.470e+02 9.676e+02, threshold=7.619e+02, percent-clipped=4.0 +2023-03-09 09:47:46,763 INFO [train.py:898] (0/4) Epoch 17, batch 3050, loss[loss=0.1868, simple_loss=0.2803, pruned_loss=0.04663, over 17972.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2588, pruned_loss=0.04096, over 3583082.05 frames. ], batch size: 65, lr: 6.60e-03, grad_scale: 4.0 +2023-03-09 09:48:16,253 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0972, 5.5660, 2.9811, 5.3908, 5.2507, 5.5659, 5.4708, 2.8660], + device='cuda:0'), covar=tensor([0.0176, 0.0052, 0.0653, 0.0060, 0.0062, 0.0065, 0.0065, 0.0818], + device='cuda:0'), in_proj_covar=tensor([0.0083, 0.0077, 0.0094, 0.0090, 0.0083, 0.0072, 0.0083, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 09:48:44,416 INFO [train.py:898] (0/4) Epoch 17, batch 3100, loss[loss=0.1968, simple_loss=0.2865, pruned_loss=0.05358, over 18239.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2585, pruned_loss=0.04103, over 3585180.71 frames. ], batch size: 60, lr: 6.59e-03, grad_scale: 4.0 +2023-03-09 09:49:09,870 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.926e+02 2.710e+02 3.175e+02 4.043e+02 1.016e+03, threshold=6.350e+02, percent-clipped=3.0 +2023-03-09 09:49:42,486 INFO [train.py:898] (0/4) Epoch 17, batch 3150, loss[loss=0.1873, simple_loss=0.2784, pruned_loss=0.04814, over 16987.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2582, pruned_loss=0.04096, over 3586109.48 frames. ], batch size: 78, lr: 6.59e-03, grad_scale: 4.0 +2023-03-09 09:50:40,781 INFO [train.py:898] (0/4) Epoch 17, batch 3200, loss[loss=0.1692, simple_loss=0.2557, pruned_loss=0.04135, over 18240.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2579, pruned_loss=0.04078, over 3598206.53 frames. ], batch size: 60, lr: 6.59e-03, grad_scale: 8.0 +2023-03-09 09:51:06,761 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.872e+02 2.721e+02 3.197e+02 3.884e+02 6.601e+02, threshold=6.394e+02, percent-clipped=2.0 +2023-03-09 09:51:15,048 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=61374.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:51:36,090 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-09 09:51:38,940 INFO [train.py:898] (0/4) Epoch 17, batch 3250, loss[loss=0.1486, simple_loss=0.2266, pruned_loss=0.03528, over 18424.00 frames. ], tot_loss[loss=0.1698, simple_loss=0.258, pruned_loss=0.04078, over 3587236.70 frames. ], batch size: 42, lr: 6.59e-03, grad_scale: 8.0 +2023-03-09 09:51:42,697 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2576, 5.2126, 4.8511, 5.1693, 5.1885, 4.5185, 5.0424, 4.8707], + device='cuda:0'), covar=tensor([0.0421, 0.0463, 0.1408, 0.0878, 0.0617, 0.0503, 0.0449, 0.1043], + device='cuda:0'), in_proj_covar=tensor([0.0468, 0.0532, 0.0686, 0.0420, 0.0424, 0.0493, 0.0521, 0.0648], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 09:51:57,906 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7527, 3.7664, 4.9894, 2.8517, 4.3672, 2.6075, 3.1732, 1.8143], + device='cuda:0'), covar=tensor([0.1145, 0.0798, 0.0115, 0.0868, 0.0557, 0.2432, 0.2469, 0.2050], + device='cuda:0'), in_proj_covar=tensor([0.0215, 0.0237, 0.0165, 0.0188, 0.0248, 0.0264, 0.0316, 0.0228], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 09:52:04,739 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=61417.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:52:04,927 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6649, 2.7407, 2.6989, 2.9771, 3.6742, 3.6037, 3.1700, 3.0244], + device='cuda:0'), covar=tensor([0.0168, 0.0286, 0.0504, 0.0303, 0.0164, 0.0143, 0.0356, 0.0328], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0127, 0.0159, 0.0153, 0.0120, 0.0108, 0.0149, 0.0145], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 09:52:10,906 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=61422.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:52:37,388 INFO [train.py:898] (0/4) Epoch 17, batch 3300, loss[loss=0.153, simple_loss=0.2442, pruned_loss=0.03092, over 18551.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2581, pruned_loss=0.04091, over 3586284.82 frames. ], batch size: 49, lr: 6.58e-03, grad_scale: 8.0 +2023-03-09 09:52:47,680 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1848, 5.2935, 4.3962, 5.1129, 5.2937, 4.6900, 4.9657, 4.7126], + device='cuda:0'), covar=tensor([0.0798, 0.0669, 0.2584, 0.1268, 0.0697, 0.0638, 0.0968, 0.1307], + device='cuda:0'), in_proj_covar=tensor([0.0463, 0.0528, 0.0677, 0.0417, 0.0420, 0.0488, 0.0517, 0.0642], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 09:52:50,327 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.63 vs. limit=5.0 +2023-03-09 09:53:01,029 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5085, 2.7109, 2.5106, 2.8412, 3.5765, 3.4555, 2.9963, 2.9111], + device='cuda:0'), covar=tensor([0.0195, 0.0303, 0.0574, 0.0367, 0.0163, 0.0180, 0.0409, 0.0366], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0126, 0.0158, 0.0151, 0.0118, 0.0107, 0.0148, 0.0144], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 09:53:02,169 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.987e+02 2.924e+02 3.407e+02 4.178e+02 8.081e+02, threshold=6.814e+02, percent-clipped=6.0 +2023-03-09 09:53:27,704 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2608, 3.1619, 1.9466, 4.0025, 2.7455, 3.8090, 2.1730, 3.4730], + device='cuda:0'), covar=tensor([0.0587, 0.0799, 0.1429, 0.0497, 0.0890, 0.0264, 0.1255, 0.0399], + device='cuda:0'), in_proj_covar=tensor([0.0206, 0.0219, 0.0186, 0.0270, 0.0190, 0.0260, 0.0200, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 09:53:35,247 INFO [train.py:898] (0/4) Epoch 17, batch 3350, loss[loss=0.1675, simple_loss=0.2616, pruned_loss=0.03672, over 15773.00 frames. ], tot_loss[loss=0.1702, simple_loss=0.2584, pruned_loss=0.04094, over 3578942.07 frames. ], batch size: 94, lr: 6.58e-03, grad_scale: 8.0 +2023-03-09 09:54:33,204 INFO [train.py:898] (0/4) Epoch 17, batch 3400, loss[loss=0.1755, simple_loss=0.2649, pruned_loss=0.04306, over 18354.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2587, pruned_loss=0.04105, over 3568757.26 frames. ], batch size: 56, lr: 6.58e-03, grad_scale: 8.0 +2023-03-09 09:54:58,506 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.971e+02 2.720e+02 3.056e+02 3.431e+02 6.999e+02, threshold=6.113e+02, percent-clipped=1.0 +2023-03-09 09:55:31,067 INFO [train.py:898] (0/4) Epoch 17, batch 3450, loss[loss=0.1681, simple_loss=0.2574, pruned_loss=0.03946, over 18613.00 frames. ], tot_loss[loss=0.1704, simple_loss=0.2583, pruned_loss=0.04124, over 3568812.65 frames. ], batch size: 52, lr: 6.57e-03, grad_scale: 8.0 +2023-03-09 09:56:30,011 INFO [train.py:898] (0/4) Epoch 17, batch 3500, loss[loss=0.186, simple_loss=0.2839, pruned_loss=0.04402, over 17192.00 frames. ], tot_loss[loss=0.171, simple_loss=0.2588, pruned_loss=0.04158, over 3557022.22 frames. ], batch size: 78, lr: 6.57e-03, grad_scale: 8.0 +2023-03-09 09:56:38,558 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7075, 2.3326, 4.2954, 3.9846, 2.4024, 4.4655, 3.8443, 2.7550], + device='cuda:0'), covar=tensor([0.0471, 0.2204, 0.0289, 0.0282, 0.2211, 0.0323, 0.0562, 0.1352], + device='cuda:0'), in_proj_covar=tensor([0.0208, 0.0235, 0.0195, 0.0154, 0.0223, 0.0207, 0.0239, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 09:56:56,115 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.780e+02 2.828e+02 3.319e+02 4.046e+02 7.174e+02, threshold=6.638e+02, percent-clipped=2.0 +2023-03-09 09:57:25,970 INFO [train.py:898] (0/4) Epoch 17, batch 3550, loss[loss=0.1732, simple_loss=0.2661, pruned_loss=0.04012, over 18568.00 frames. ], tot_loss[loss=0.171, simple_loss=0.259, pruned_loss=0.0415, over 3565526.75 frames. ], batch size: 54, lr: 6.57e-03, grad_scale: 8.0 +2023-03-09 09:57:27,405 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7274, 2.8875, 2.6657, 3.0798, 3.6779, 3.6246, 3.2310, 3.0739], + device='cuda:0'), covar=tensor([0.0217, 0.0283, 0.0551, 0.0345, 0.0194, 0.0172, 0.0393, 0.0374], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0126, 0.0159, 0.0151, 0.0119, 0.0107, 0.0148, 0.0146], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 09:57:36,210 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1397, 3.8174, 5.1276, 3.0015, 4.5303, 2.6993, 3.1356, 1.8449], + device='cuda:0'), covar=tensor([0.0995, 0.0829, 0.0114, 0.0836, 0.0507, 0.2466, 0.2621, 0.2116], + device='cuda:0'), in_proj_covar=tensor([0.0218, 0.0239, 0.0168, 0.0190, 0.0251, 0.0266, 0.0319, 0.0230], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 09:57:50,665 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=61717.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:58:07,026 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61732.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:58:18,857 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61743.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:58:20,833 INFO [train.py:898] (0/4) Epoch 17, batch 3600, loss[loss=0.1431, simple_loss=0.2253, pruned_loss=0.03044, over 18439.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2589, pruned_loss=0.04139, over 3562710.14 frames. ], batch size: 43, lr: 6.57e-03, grad_scale: 8.0 +2023-03-09 09:58:42,479 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=61765.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 09:58:44,426 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.882e+02 3.033e+02 3.500e+02 4.086e+02 7.199e+02, threshold=7.000e+02, percent-clipped=1.0 +2023-03-09 09:58:56,300 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-17.pt +2023-03-09 09:59:23,037 INFO [train.py:898] (0/4) Epoch 18, batch 0, loss[loss=0.1661, simple_loss=0.2545, pruned_loss=0.03887, over 18505.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2545, pruned_loss=0.03887, over 18505.00 frames. ], batch size: 51, lr: 6.38e-03, grad_scale: 8.0 +2023-03-09 09:59:23,039 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 09:59:34,864 INFO [train.py:932] (0/4) Epoch 18, validation: loss=0.1526, simple_loss=0.2531, pruned_loss=0.0261, over 944034.00 frames. +2023-03-09 09:59:34,865 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 09:59:50,853 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61793.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:00:05,777 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61804.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:00:34,728 INFO [train.py:898] (0/4) Epoch 18, batch 50, loss[loss=0.1602, simple_loss=0.2517, pruned_loss=0.03428, over 18629.00 frames. ], tot_loss[loss=0.1726, simple_loss=0.2616, pruned_loss=0.04174, over 818145.93 frames. ], batch size: 52, lr: 6.37e-03, grad_scale: 8.0 +2023-03-09 10:01:20,167 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.961e+02 2.892e+02 3.419e+02 4.081e+02 7.756e+02, threshold=6.838e+02, percent-clipped=1.0 +2023-03-09 10:01:33,967 INFO [train.py:898] (0/4) Epoch 18, batch 100, loss[loss=0.1693, simple_loss=0.2484, pruned_loss=0.04504, over 17607.00 frames. ], tot_loss[loss=0.1719, simple_loss=0.2609, pruned_loss=0.04143, over 1427838.99 frames. ], batch size: 39, lr: 6.37e-03, grad_scale: 8.0 +2023-03-09 10:02:02,558 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-09 10:02:15,356 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=61914.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:02:32,572 INFO [train.py:898] (0/4) Epoch 18, batch 150, loss[loss=0.1739, simple_loss=0.2579, pruned_loss=0.04491, over 18266.00 frames. ], tot_loss[loss=0.1711, simple_loss=0.2602, pruned_loss=0.04097, over 1916402.68 frames. ], batch size: 49, lr: 6.37e-03, grad_scale: 8.0 +2023-03-09 10:02:49,400 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.90 vs. limit=5.0 +2023-03-09 10:03:17,466 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.774e+02 2.985e+02 3.348e+02 3.997e+02 9.242e+02, threshold=6.695e+02, percent-clipped=1.0 +2023-03-09 10:03:27,601 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=61975.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:03:32,033 INFO [train.py:898] (0/4) Epoch 18, batch 200, loss[loss=0.1547, simple_loss=0.2467, pruned_loss=0.03135, over 18511.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2584, pruned_loss=0.04027, over 2292045.77 frames. ], batch size: 47, lr: 6.37e-03, grad_scale: 8.0 +2023-03-09 10:03:56,040 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-62000.pt +2023-03-09 10:04:35,242 INFO [train.py:898] (0/4) Epoch 18, batch 250, loss[loss=0.1571, simple_loss=0.2364, pruned_loss=0.03887, over 18049.00 frames. ], tot_loss[loss=0.1697, simple_loss=0.2586, pruned_loss=0.04044, over 2583999.66 frames. ], batch size: 40, lr: 6.36e-03, grad_scale: 8.0 +2023-03-09 10:04:38,921 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9893, 5.2373, 2.4878, 5.1008, 5.0453, 5.2925, 5.0331, 2.4912], + device='cuda:0'), covar=tensor([0.0190, 0.0078, 0.0879, 0.0097, 0.0076, 0.0075, 0.0103, 0.1097], + device='cuda:0'), in_proj_covar=tensor([0.0084, 0.0078, 0.0094, 0.0091, 0.0083, 0.0072, 0.0083, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 10:04:40,186 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8008, 3.6322, 5.0483, 4.4890, 3.4379, 3.0437, 4.5186, 5.2627], + device='cuda:0'), covar=tensor([0.0757, 0.1576, 0.0157, 0.0340, 0.0811, 0.1142, 0.0320, 0.0159], + device='cuda:0'), in_proj_covar=tensor([0.0145, 0.0268, 0.0138, 0.0177, 0.0187, 0.0188, 0.0190, 0.0184], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 10:04:45,101 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-09 10:05:14,850 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5738, 3.3181, 2.0232, 4.3923, 3.2520, 4.0152, 2.1403, 3.8538], + device='cuda:0'), covar=tensor([0.0594, 0.0809, 0.1498, 0.0486, 0.0736, 0.0247, 0.1465, 0.0446], + device='cuda:0'), in_proj_covar=tensor([0.0207, 0.0221, 0.0187, 0.0273, 0.0193, 0.0262, 0.0200, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 10:05:19,646 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.960e+02 2.782e+02 3.298e+02 4.084e+02 9.101e+02, threshold=6.597e+02, percent-clipped=2.0 +2023-03-09 10:05:34,836 INFO [train.py:898] (0/4) Epoch 18, batch 300, loss[loss=0.1497, simple_loss=0.2302, pruned_loss=0.03458, over 18388.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2572, pruned_loss=0.04007, over 2805299.57 frames. ], batch size: 42, lr: 6.36e-03, grad_scale: 8.0 +2023-03-09 10:05:45,148 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62088.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:05:57,525 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62099.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:06:24,887 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62122.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:06:33,271 INFO [train.py:898] (0/4) Epoch 18, batch 350, loss[loss=0.1775, simple_loss=0.2665, pruned_loss=0.04424, over 18217.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.257, pruned_loss=0.04, over 2975586.25 frames. ], batch size: 60, lr: 6.36e-03, grad_scale: 8.0 +2023-03-09 10:06:35,465 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-09 10:07:17,194 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.168e+02 2.722e+02 3.288e+02 4.143e+02 7.377e+02, threshold=6.576e+02, percent-clipped=3.0 +2023-03-09 10:07:32,439 INFO [train.py:898] (0/4) Epoch 18, batch 400, loss[loss=0.1679, simple_loss=0.2542, pruned_loss=0.04074, over 18528.00 frames. ], tot_loss[loss=0.1675, simple_loss=0.2561, pruned_loss=0.03943, over 3120862.09 frames. ], batch size: 47, lr: 6.36e-03, grad_scale: 8.0 +2023-03-09 10:07:37,202 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62183.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:08:26,315 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.75 vs. limit=5.0 +2023-03-09 10:08:30,215 INFO [train.py:898] (0/4) Epoch 18, batch 450, loss[loss=0.1591, simple_loss=0.2481, pruned_loss=0.03506, over 18562.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2572, pruned_loss=0.03982, over 3226510.83 frames. ], batch size: 49, lr: 6.35e-03, grad_scale: 8.0 +2023-03-09 10:08:32,188 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-09 10:09:15,160 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.738e+02 2.782e+02 3.170e+02 3.804e+02 6.822e+02, threshold=6.341e+02, percent-clipped=2.0 +2023-03-09 10:09:18,863 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62270.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:09:22,352 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62273.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:09:26,322 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8977, 4.6155, 4.7858, 3.7095, 3.8722, 3.6527, 2.7724, 2.5986], + device='cuda:0'), covar=tensor([0.0225, 0.0170, 0.0058, 0.0238, 0.0305, 0.0186, 0.0690, 0.0817], + device='cuda:0'), in_proj_covar=tensor([0.0067, 0.0056, 0.0058, 0.0066, 0.0087, 0.0064, 0.0076, 0.0082], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 10:09:29,823 INFO [train.py:898] (0/4) Epoch 18, batch 500, loss[loss=0.1687, simple_loss=0.2653, pruned_loss=0.03611, over 18550.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2564, pruned_loss=0.03948, over 3301536.98 frames. ], batch size: 54, lr: 6.35e-03, grad_scale: 8.0 +2023-03-09 10:10:28,251 INFO [train.py:898] (0/4) Epoch 18, batch 550, loss[loss=0.1972, simple_loss=0.2933, pruned_loss=0.05057, over 18298.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2575, pruned_loss=0.04003, over 3354826.33 frames. ], batch size: 57, lr: 6.35e-03, grad_scale: 8.0 +2023-03-09 10:10:34,851 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62334.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:11:13,350 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.182e+02 2.712e+02 3.173e+02 3.739e+02 5.690e+02, threshold=6.346e+02, percent-clipped=0.0 +2023-03-09 10:11:27,632 INFO [train.py:898] (0/4) Epoch 18, batch 600, loss[loss=0.1851, simple_loss=0.278, pruned_loss=0.0461, over 18000.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2566, pruned_loss=0.03965, over 3409137.84 frames. ], batch size: 65, lr: 6.35e-03, grad_scale: 8.0 +2023-03-09 10:11:38,852 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62388.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:11:42,962 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62391.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:11:52,668 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62399.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:12:26,536 INFO [train.py:898] (0/4) Epoch 18, batch 650, loss[loss=0.1703, simple_loss=0.2688, pruned_loss=0.03586, over 18347.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.257, pruned_loss=0.03999, over 3457678.92 frames. ], batch size: 55, lr: 6.34e-03, grad_scale: 8.0 +2023-03-09 10:12:35,142 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=62436.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:12:48,549 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=62447.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:12:55,126 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62452.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:13:12,081 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.651e+02 2.765e+02 3.266e+02 3.962e+02 6.380e+02, threshold=6.532e+02, percent-clipped=1.0 +2023-03-09 10:13:24,733 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62478.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:13:25,699 INFO [train.py:898] (0/4) Epoch 18, batch 700, loss[loss=0.158, simple_loss=0.2625, pruned_loss=0.02679, over 17750.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2572, pruned_loss=0.03988, over 3497211.12 frames. ], batch size: 70, lr: 6.34e-03, grad_scale: 8.0 +2023-03-09 10:13:47,672 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-09 10:14:21,569 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3681, 5.8765, 5.4502, 5.6620, 5.4354, 5.3757, 5.9365, 5.8724], + device='cuda:0'), covar=tensor([0.1097, 0.0771, 0.0453, 0.0712, 0.1441, 0.0656, 0.0583, 0.0736], + device='cuda:0'), in_proj_covar=tensor([0.0583, 0.0499, 0.0366, 0.0525, 0.0721, 0.0522, 0.0706, 0.0535], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 10:14:24,851 INFO [train.py:898] (0/4) Epoch 18, batch 750, loss[loss=0.1787, simple_loss=0.2701, pruned_loss=0.04366, over 18262.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.257, pruned_loss=0.03992, over 3518404.11 frames. ], batch size: 57, lr: 6.34e-03, grad_scale: 4.0 +2023-03-09 10:14:44,106 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62545.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:14:57,282 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0822, 4.2584, 2.5121, 4.2271, 5.3519, 2.7937, 3.9880, 3.9888], + device='cuda:0'), covar=tensor([0.0122, 0.1016, 0.1531, 0.0536, 0.0055, 0.1075, 0.0573, 0.0724], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0256, 0.0198, 0.0192, 0.0111, 0.0177, 0.0209, 0.0218], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 10:15:10,947 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.733e+02 2.727e+02 3.281e+02 3.903e+02 8.552e+02, threshold=6.561e+02, percent-clipped=5.0 +2023-03-09 10:15:13,395 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62570.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:15:23,437 INFO [train.py:898] (0/4) Epoch 18, batch 800, loss[loss=0.1561, simple_loss=0.2353, pruned_loss=0.03849, over 18427.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2572, pruned_loss=0.04005, over 3537125.46 frames. ], batch size: 43, lr: 6.34e-03, grad_scale: 8.0 +2023-03-09 10:15:30,574 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62585.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:15:56,916 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62606.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:16:10,946 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=62618.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:16:23,375 INFO [train.py:898] (0/4) Epoch 18, batch 850, loss[loss=0.1531, simple_loss=0.2398, pruned_loss=0.03318, over 18248.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2576, pruned_loss=0.03976, over 3556230.88 frames. ], batch size: 47, lr: 6.33e-03, grad_scale: 8.0 +2023-03-09 10:16:23,587 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62629.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:16:44,069 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62646.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:16:50,821 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8925, 5.1087, 2.6189, 4.9452, 4.8568, 5.1415, 4.8760, 2.4943], + device='cuda:0'), covar=tensor([0.0215, 0.0058, 0.0839, 0.0095, 0.0071, 0.0067, 0.0099, 0.1143], + device='cuda:0'), in_proj_covar=tensor([0.0084, 0.0078, 0.0093, 0.0091, 0.0083, 0.0073, 0.0083, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 10:17:07,928 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62666.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:17:09,781 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.769e+02 2.717e+02 3.249e+02 3.987e+02 1.142e+03, threshold=6.498e+02, percent-clipped=2.0 +2023-03-09 10:17:22,614 INFO [train.py:898] (0/4) Epoch 18, batch 900, loss[loss=0.1695, simple_loss=0.2566, pruned_loss=0.04123, over 18372.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2574, pruned_loss=0.03989, over 3558293.10 frames. ], batch size: 50, lr: 6.33e-03, grad_scale: 8.0 +2023-03-09 10:17:46,552 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8639, 3.5513, 5.0291, 4.5141, 3.3640, 3.0931, 4.4422, 5.2075], + device='cuda:0'), covar=tensor([0.0760, 0.1694, 0.0137, 0.0341, 0.0864, 0.1070, 0.0362, 0.0225], + device='cuda:0'), in_proj_covar=tensor([0.0147, 0.0273, 0.0140, 0.0179, 0.0188, 0.0189, 0.0191, 0.0187], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 10:17:56,786 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-09 10:17:58,492 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5386, 3.3528, 2.3523, 4.3819, 3.0217, 4.2497, 2.4546, 3.8337], + device='cuda:0'), covar=tensor([0.0629, 0.0838, 0.1393, 0.0430, 0.0861, 0.0296, 0.1142, 0.0386], + device='cuda:0'), in_proj_covar=tensor([0.0207, 0.0220, 0.0185, 0.0271, 0.0191, 0.0262, 0.0199, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 10:18:19,752 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62727.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:18:21,831 INFO [train.py:898] (0/4) Epoch 18, batch 950, loss[loss=0.1746, simple_loss=0.2675, pruned_loss=0.04081, over 17129.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2569, pruned_loss=0.03958, over 3570914.63 frames. ], batch size: 78, lr: 6.33e-03, grad_scale: 8.0 +2023-03-09 10:18:22,214 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6477, 3.4415, 2.4526, 4.4666, 3.1452, 4.3851, 2.6826, 3.9731], + device='cuda:0'), covar=tensor([0.0587, 0.0775, 0.1325, 0.0481, 0.0816, 0.0321, 0.1068, 0.0382], + device='cuda:0'), in_proj_covar=tensor([0.0207, 0.0220, 0.0185, 0.0271, 0.0191, 0.0262, 0.0199, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 10:18:43,077 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62747.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:19:07,974 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.629e+02 2.750e+02 3.313e+02 3.832e+02 8.841e+02, threshold=6.625e+02, percent-clipped=2.0 +2023-03-09 10:19:14,015 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=62773.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:19:20,207 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62778.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:19:21,041 INFO [train.py:898] (0/4) Epoch 18, batch 1000, loss[loss=0.1384, simple_loss=0.221, pruned_loss=0.02794, over 18460.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.257, pruned_loss=0.03961, over 3574515.64 frames. ], batch size: 44, lr: 6.33e-03, grad_scale: 8.0 +2023-03-09 10:20:16,368 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=62826.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:20:19,782 INFO [train.py:898] (0/4) Epoch 18, batch 1050, loss[loss=0.1486, simple_loss=0.2377, pruned_loss=0.02972, over 18546.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2566, pruned_loss=0.03951, over 3588369.83 frames. ], batch size: 49, lr: 6.32e-03, grad_scale: 8.0 +2023-03-09 10:20:26,337 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=62834.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:21:05,603 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.820e+02 2.730e+02 3.063e+02 3.769e+02 8.307e+02, threshold=6.126e+02, percent-clipped=1.0 +2023-03-09 10:21:11,490 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2260, 5.1972, 5.5190, 5.5887, 5.1464, 6.0606, 5.6910, 5.2756], + device='cuda:0'), covar=tensor([0.1080, 0.0596, 0.0619, 0.0698, 0.1358, 0.0733, 0.0567, 0.1593], + device='cuda:0'), in_proj_covar=tensor([0.0341, 0.0273, 0.0296, 0.0294, 0.0322, 0.0404, 0.0264, 0.0400], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 10:21:18,728 INFO [train.py:898] (0/4) Epoch 18, batch 1100, loss[loss=0.2027, simple_loss=0.2985, pruned_loss=0.05346, over 18463.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2569, pruned_loss=0.03962, over 3597584.28 frames. ], batch size: 59, lr: 6.32e-03, grad_scale: 8.0 +2023-03-09 10:21:44,700 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62901.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:22:17,695 INFO [train.py:898] (0/4) Epoch 18, batch 1150, loss[loss=0.1788, simple_loss=0.2717, pruned_loss=0.04295, over 16104.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2566, pruned_loss=0.03944, over 3601735.61 frames. ], batch size: 94, lr: 6.32e-03, grad_scale: 8.0 +2023-03-09 10:22:17,931 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=62929.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:22:32,070 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=62941.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:22:40,477 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2275, 5.2281, 5.3759, 5.4187, 5.1381, 5.9802, 5.5948, 5.2825], + device='cuda:0'), covar=tensor([0.1011, 0.0642, 0.0728, 0.0810, 0.1441, 0.0755, 0.0720, 0.1596], + device='cuda:0'), in_proj_covar=tensor([0.0344, 0.0275, 0.0298, 0.0297, 0.0324, 0.0407, 0.0267, 0.0401], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 10:23:03,231 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.305e+02 2.742e+02 3.290e+02 3.886e+02 6.757e+02, threshold=6.580e+02, percent-clipped=2.0 +2023-03-09 10:23:07,164 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.74 vs. limit=5.0 +2023-03-09 10:23:12,998 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0727, 4.3511, 2.3838, 4.1606, 5.3618, 2.6800, 3.9773, 4.0641], + device='cuda:0'), covar=tensor([0.0149, 0.1004, 0.1804, 0.0662, 0.0071, 0.1301, 0.0636, 0.0693], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0258, 0.0200, 0.0194, 0.0113, 0.0178, 0.0211, 0.0219], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0001, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 10:23:13,838 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=62977.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:23:15,924 INFO [train.py:898] (0/4) Epoch 18, batch 1200, loss[loss=0.1598, simple_loss=0.252, pruned_loss=0.03381, over 18563.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2566, pruned_loss=0.03946, over 3607681.56 frames. ], batch size: 54, lr: 6.32e-03, grad_scale: 8.0 +2023-03-09 10:23:48,121 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8727, 3.5913, 4.8665, 2.8985, 4.2000, 2.5845, 3.0830, 1.8080], + device='cuda:0'), covar=tensor([0.1106, 0.0879, 0.0168, 0.0896, 0.0586, 0.2528, 0.2528, 0.2115], + device='cuda:0'), in_proj_covar=tensor([0.0215, 0.0239, 0.0168, 0.0189, 0.0249, 0.0265, 0.0316, 0.0227], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 10:24:06,505 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63022.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:24:14,842 INFO [train.py:898] (0/4) Epoch 18, batch 1250, loss[loss=0.1521, simple_loss=0.2416, pruned_loss=0.03128, over 18380.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2567, pruned_loss=0.03965, over 3599195.31 frames. ], batch size: 46, lr: 6.31e-03, grad_scale: 8.0 +2023-03-09 10:24:35,994 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63047.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:24:44,718 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6146, 3.4245, 5.0145, 3.1007, 4.2823, 2.6051, 3.0469, 1.7606], + device='cuda:0'), covar=tensor([0.1207, 0.0983, 0.0122, 0.0740, 0.0558, 0.2412, 0.2428, 0.2038], + device='cuda:0'), in_proj_covar=tensor([0.0217, 0.0241, 0.0170, 0.0191, 0.0251, 0.0267, 0.0319, 0.0228], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 10:25:00,073 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.748e+02 2.862e+02 3.402e+02 4.233e+02 9.717e+02, threshold=6.805e+02, percent-clipped=8.0 +2023-03-09 10:25:13,511 INFO [train.py:898] (0/4) Epoch 18, batch 1300, loss[loss=0.1737, simple_loss=0.2646, pruned_loss=0.04137, over 18277.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2576, pruned_loss=0.04016, over 3579187.46 frames. ], batch size: 57, lr: 6.31e-03, grad_scale: 8.0 +2023-03-09 10:25:32,640 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=63095.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:26:12,334 INFO [train.py:898] (0/4) Epoch 18, batch 1350, loss[loss=0.1382, simple_loss=0.2256, pruned_loss=0.02545, over 18257.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2581, pruned_loss=0.04032, over 3575000.75 frames. ], batch size: 45, lr: 6.31e-03, grad_scale: 8.0 +2023-03-09 10:26:12,465 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63129.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:26:58,486 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.671e+02 2.877e+02 3.417e+02 4.263e+02 7.577e+02, threshold=6.833e+02, percent-clipped=1.0 +2023-03-09 10:27:00,082 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8524, 3.1595, 4.5148, 3.9685, 3.2264, 4.9015, 4.1783, 3.3407], + device='cuda:0'), covar=tensor([0.0481, 0.1396, 0.0248, 0.0422, 0.1368, 0.0165, 0.0455, 0.0848], + device='cuda:0'), in_proj_covar=tensor([0.0209, 0.0236, 0.0197, 0.0157, 0.0225, 0.0207, 0.0241, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 10:27:11,774 INFO [train.py:898] (0/4) Epoch 18, batch 1400, loss[loss=0.1964, simple_loss=0.2822, pruned_loss=0.05524, over 18148.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2583, pruned_loss=0.04042, over 3580045.09 frames. ], batch size: 62, lr: 6.31e-03, grad_scale: 8.0 +2023-03-09 10:27:39,020 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63201.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:28:02,583 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63221.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:28:11,962 INFO [train.py:898] (0/4) Epoch 18, batch 1450, loss[loss=0.1777, simple_loss=0.2711, pruned_loss=0.0422, over 18633.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2585, pruned_loss=0.04011, over 3592151.85 frames. ], batch size: 52, lr: 6.30e-03, grad_scale: 8.0 +2023-03-09 10:28:26,200 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63241.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:28:35,842 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=63249.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:28:57,797 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.907e+02 2.745e+02 3.338e+02 4.107e+02 1.231e+03, threshold=6.676e+02, percent-clipped=2.0 +2023-03-09 10:29:10,024 INFO [train.py:898] (0/4) Epoch 18, batch 1500, loss[loss=0.1637, simple_loss=0.2526, pruned_loss=0.0374, over 18369.00 frames. ], tot_loss[loss=0.1708, simple_loss=0.2597, pruned_loss=0.04094, over 3576509.91 frames. ], batch size: 46, lr: 6.30e-03, grad_scale: 8.0 +2023-03-09 10:29:14,266 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63282.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:29:22,512 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=63289.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:29:31,339 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1969, 4.3800, 2.7898, 4.3650, 5.4604, 2.7657, 4.1195, 4.2917], + device='cuda:0'), covar=tensor([0.0137, 0.1018, 0.1382, 0.0568, 0.0070, 0.1144, 0.0564, 0.0613], + device='cuda:0'), in_proj_covar=tensor([0.0155, 0.0260, 0.0201, 0.0195, 0.0114, 0.0180, 0.0213, 0.0221], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 10:30:01,225 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63322.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:30:09,223 INFO [train.py:898] (0/4) Epoch 18, batch 1550, loss[loss=0.1727, simple_loss=0.2665, pruned_loss=0.03942, over 17841.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.2582, pruned_loss=0.04028, over 3593042.02 frames. ], batch size: 70, lr: 6.30e-03, grad_scale: 8.0 +2023-03-09 10:30:55,753 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.939e+02 2.883e+02 3.563e+02 4.171e+02 7.864e+02, threshold=7.125e+02, percent-clipped=2.0 +2023-03-09 10:30:58,246 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=63370.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:31:08,378 INFO [train.py:898] (0/4) Epoch 18, batch 1600, loss[loss=0.1735, simple_loss=0.2657, pruned_loss=0.04064, over 16880.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2588, pruned_loss=0.04046, over 3589876.99 frames. ], batch size: 78, lr: 6.30e-03, grad_scale: 8.0 +2023-03-09 10:31:57,980 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63420.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:32:00,266 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63422.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:32:08,055 INFO [train.py:898] (0/4) Epoch 18, batch 1650, loss[loss=0.1762, simple_loss=0.2642, pruned_loss=0.04412, over 18243.00 frames. ], tot_loss[loss=0.1701, simple_loss=0.2593, pruned_loss=0.04046, over 3587456.98 frames. ], batch size: 60, lr: 6.29e-03, grad_scale: 8.0 +2023-03-09 10:32:08,350 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63429.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:32:54,720 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.082e+02 2.837e+02 3.509e+02 4.099e+02 8.239e+02, threshold=7.017e+02, percent-clipped=1.0 +2023-03-09 10:32:57,733 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-09 10:33:05,379 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=63477.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:33:07,398 INFO [train.py:898] (0/4) Epoch 18, batch 1700, loss[loss=0.1681, simple_loss=0.2568, pruned_loss=0.03969, over 18398.00 frames. ], tot_loss[loss=0.1703, simple_loss=0.2595, pruned_loss=0.04057, over 3601129.75 frames. ], batch size: 48, lr: 6.29e-03, grad_scale: 8.0 +2023-03-09 10:33:10,108 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63481.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:33:12,263 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63483.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:34:06,914 INFO [train.py:898] (0/4) Epoch 18, batch 1750, loss[loss=0.1745, simple_loss=0.2661, pruned_loss=0.04143, over 18580.00 frames. ], tot_loss[loss=0.17, simple_loss=0.2591, pruned_loss=0.0405, over 3602031.78 frames. ], batch size: 54, lr: 6.29e-03, grad_scale: 8.0 +2023-03-09 10:34:26,189 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.59 vs. limit=2.0 +2023-03-09 10:34:52,617 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.976e+02 2.714e+02 3.242e+02 3.831e+02 6.997e+02, threshold=6.484e+02, percent-clipped=0.0 +2023-03-09 10:34:54,065 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63569.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:35:03,542 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63577.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:35:05,724 INFO [train.py:898] (0/4) Epoch 18, batch 1800, loss[loss=0.1704, simple_loss=0.2633, pruned_loss=0.03878, over 18499.00 frames. ], tot_loss[loss=0.1699, simple_loss=0.2593, pruned_loss=0.04025, over 3600323.18 frames. ], batch size: 51, lr: 6.29e-03, grad_scale: 8.0 +2023-03-09 10:35:13,895 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63586.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:36:05,192 INFO [train.py:898] (0/4) Epoch 18, batch 1850, loss[loss=0.2003, simple_loss=0.2815, pruned_loss=0.05953, over 13058.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2583, pruned_loss=0.04028, over 3589244.35 frames. ], batch size: 130, lr: 6.28e-03, grad_scale: 8.0 +2023-03-09 10:36:06,660 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63630.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:36:14,546 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8552, 5.3647, 5.3278, 5.3477, 4.8420, 5.2749, 4.6745, 5.2083], + device='cuda:0'), covar=tensor([0.0244, 0.0270, 0.0190, 0.0366, 0.0420, 0.0217, 0.1049, 0.0293], + device='cuda:0'), in_proj_covar=tensor([0.0207, 0.0251, 0.0242, 0.0309, 0.0262, 0.0257, 0.0298, 0.0249], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 10:36:26,784 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63647.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:36:51,538 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.913e+02 2.867e+02 3.371e+02 3.791e+02 7.715e+02, threshold=6.743e+02, percent-clipped=3.0 +2023-03-09 10:37:04,237 INFO [train.py:898] (0/4) Epoch 18, batch 1900, loss[loss=0.1471, simple_loss=0.2353, pruned_loss=0.02947, over 18401.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2572, pruned_loss=0.03991, over 3591673.19 frames. ], batch size: 42, lr: 6.28e-03, grad_scale: 8.0 +2023-03-09 10:37:19,032 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6666, 3.4458, 5.0338, 3.0335, 4.4482, 2.5177, 2.9407, 1.8798], + device='cuda:0'), covar=tensor([0.1249, 0.0989, 0.0139, 0.0731, 0.0491, 0.2647, 0.2855, 0.2153], + device='cuda:0'), in_proj_covar=tensor([0.0216, 0.0239, 0.0170, 0.0191, 0.0251, 0.0265, 0.0317, 0.0227], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 10:37:32,256 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.49 vs. limit=5.0 +2023-03-09 10:37:51,115 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6443, 2.8384, 2.6020, 2.9474, 3.7419, 3.5424, 3.1902, 2.9987], + device='cuda:0'), covar=tensor([0.0181, 0.0315, 0.0551, 0.0379, 0.0158, 0.0203, 0.0359, 0.0364], + device='cuda:0'), in_proj_covar=tensor([0.0133, 0.0130, 0.0162, 0.0152, 0.0123, 0.0110, 0.0151, 0.0150], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 10:38:02,775 INFO [train.py:898] (0/4) Epoch 18, batch 1950, loss[loss=0.2144, simple_loss=0.2948, pruned_loss=0.06696, over 17108.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2582, pruned_loss=0.04044, over 3589473.21 frames. ], batch size: 78, lr: 6.28e-03, grad_scale: 8.0 +2023-03-09 10:38:19,602 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8602, 4.5396, 4.5810, 3.4290, 3.7084, 3.4600, 2.6689, 2.3084], + device='cuda:0'), covar=tensor([0.0205, 0.0143, 0.0084, 0.0293, 0.0323, 0.0216, 0.0684, 0.0914], + device='cuda:0'), in_proj_covar=tensor([0.0067, 0.0056, 0.0059, 0.0066, 0.0086, 0.0064, 0.0075, 0.0082], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 10:38:50,309 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.821e+02 2.688e+02 3.189e+02 3.862e+02 9.985e+02, threshold=6.379e+02, percent-clipped=3.0 +2023-03-09 10:38:59,647 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63776.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:39:01,976 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63778.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:39:02,965 INFO [train.py:898] (0/4) Epoch 18, batch 2000, loss[loss=0.1443, simple_loss=0.2249, pruned_loss=0.03185, over 18495.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2581, pruned_loss=0.04044, over 3585425.83 frames. ], batch size: 44, lr: 6.28e-03, grad_scale: 8.0 +2023-03-09 10:39:40,020 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63810.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:40:02,092 INFO [train.py:898] (0/4) Epoch 18, batch 2050, loss[loss=0.1811, simple_loss=0.2744, pruned_loss=0.04392, over 18053.00 frames. ], tot_loss[loss=0.1694, simple_loss=0.258, pruned_loss=0.04039, over 3595191.05 frames. ], batch size: 65, lr: 6.27e-03, grad_scale: 8.0 +2023-03-09 10:40:14,289 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6301, 2.6161, 2.6035, 2.4268, 2.5428, 2.1641, 2.2448, 2.6291], + device='cuda:0'), covar=tensor([0.0069, 0.0087, 0.0073, 0.0108, 0.0105, 0.0169, 0.0171, 0.0075], + device='cuda:0'), in_proj_covar=tensor([0.0127, 0.0151, 0.0127, 0.0179, 0.0133, 0.0171, 0.0176, 0.0112], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 10:40:21,055 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63845.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:40:48,571 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.076e+02 2.804e+02 3.273e+02 4.132e+02 7.162e+02, threshold=6.546e+02, percent-clipped=3.0 +2023-03-09 10:40:53,077 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63871.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:40:59,888 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=63877.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:41:01,890 INFO [train.py:898] (0/4) Epoch 18, batch 2100, loss[loss=0.175, simple_loss=0.2662, pruned_loss=0.04187, over 18136.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2576, pruned_loss=0.04037, over 3583790.63 frames. ], batch size: 62, lr: 6.27e-03, grad_scale: 8.0 +2023-03-09 10:41:26,447 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4527, 3.2582, 2.1942, 4.1843, 3.0055, 4.0960, 2.1929, 3.6669], + device='cuda:0'), covar=tensor([0.0569, 0.0856, 0.1298, 0.0476, 0.0827, 0.0300, 0.1299, 0.0417], + device='cuda:0'), in_proj_covar=tensor([0.0205, 0.0219, 0.0183, 0.0271, 0.0188, 0.0258, 0.0197, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 10:41:33,598 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63906.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:41:45,659 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=63916.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:41:56,441 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=63925.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:41:56,462 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63925.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:42:00,982 INFO [train.py:898] (0/4) Epoch 18, batch 2150, loss[loss=0.1948, simple_loss=0.2836, pruned_loss=0.05303, over 18497.00 frames. ], tot_loss[loss=0.1695, simple_loss=0.2579, pruned_loss=0.04056, over 3561182.91 frames. ], batch size: 59, lr: 6.27e-03, grad_scale: 8.0 +2023-03-09 10:42:16,525 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=63942.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:42:47,025 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.000e+02 2.608e+02 3.106e+02 3.507e+02 7.149e+02, threshold=6.212e+02, percent-clipped=2.0 +2023-03-09 10:42:57,366 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5096, 5.3557, 5.0940, 5.2132, 4.6848, 5.1378, 5.5071, 5.3608], + device='cuda:0'), covar=tensor([0.2635, 0.1282, 0.0956, 0.1328, 0.2731, 0.1233, 0.1117, 0.1250], + device='cuda:0'), in_proj_covar=tensor([0.0601, 0.0509, 0.0369, 0.0542, 0.0736, 0.0536, 0.0727, 0.0553], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 10:42:57,513 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=63977.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:42:59,800 INFO [train.py:898] (0/4) Epoch 18, batch 2200, loss[loss=0.1614, simple_loss=0.2525, pruned_loss=0.03518, over 18384.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2575, pruned_loss=0.04003, over 3568044.43 frames. ], batch size: 52, lr: 6.27e-03, grad_scale: 8.0 +2023-03-09 10:43:24,446 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-64000.pt +2023-03-09 10:43:44,402 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8099, 3.8422, 5.0956, 4.5144, 3.3524, 3.2674, 4.7160, 5.4495], + device='cuda:0'), covar=tensor([0.0768, 0.1473, 0.0180, 0.0379, 0.0883, 0.1002, 0.0302, 0.0154], + device='cuda:0'), in_proj_covar=tensor([0.0146, 0.0270, 0.0141, 0.0180, 0.0187, 0.0186, 0.0191, 0.0186], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 10:43:51,567 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.7041, 6.2359, 5.5882, 6.0586, 5.8759, 5.6829, 6.2963, 6.2238], + device='cuda:0'), covar=tensor([0.1136, 0.0663, 0.0425, 0.0639, 0.1329, 0.0651, 0.0518, 0.0610], + device='cuda:0'), in_proj_covar=tensor([0.0598, 0.0506, 0.0369, 0.0538, 0.0734, 0.0535, 0.0724, 0.0551], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 10:43:53,798 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3086, 5.7893, 5.3540, 5.5721, 5.4057, 5.2162, 5.8452, 5.7823], + device='cuda:0'), covar=tensor([0.1099, 0.0764, 0.0547, 0.0741, 0.1397, 0.0717, 0.0563, 0.0706], + device='cuda:0'), in_proj_covar=tensor([0.0597, 0.0506, 0.0369, 0.0537, 0.0734, 0.0534, 0.0723, 0.0551], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 10:43:56,109 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64023.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:44:02,720 INFO [train.py:898] (0/4) Epoch 18, batch 2250, loss[loss=0.1588, simple_loss=0.2518, pruned_loss=0.03294, over 18635.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2573, pruned_loss=0.04007, over 3567944.85 frames. ], batch size: 52, lr: 6.26e-03, grad_scale: 8.0 +2023-03-09 10:44:14,961 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1552, 5.6743, 3.3934, 5.4218, 5.3519, 5.6852, 5.4982, 3.2758], + device='cuda:0'), covar=tensor([0.0174, 0.0046, 0.0535, 0.0065, 0.0052, 0.0044, 0.0060, 0.0687], + device='cuda:0'), in_proj_covar=tensor([0.0085, 0.0078, 0.0094, 0.0091, 0.0083, 0.0073, 0.0083, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 10:44:47,978 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.732e+02 2.740e+02 3.229e+02 3.665e+02 1.396e+03, threshold=6.458e+02, percent-clipped=4.0 +2023-03-09 10:44:57,919 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64076.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:45:00,181 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64078.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:45:01,195 INFO [train.py:898] (0/4) Epoch 18, batch 2300, loss[loss=0.1697, simple_loss=0.2607, pruned_loss=0.03932, over 18355.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2569, pruned_loss=0.04017, over 3568459.02 frames. ], batch size: 56, lr: 6.26e-03, grad_scale: 8.0 +2023-03-09 10:45:07,196 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64084.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:45:51,499 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2272, 4.2083, 3.9828, 4.1380, 4.1770, 3.5812, 4.1513, 3.9565], + device='cuda:0'), covar=tensor([0.0476, 0.0585, 0.1180, 0.0752, 0.0604, 0.0502, 0.0466, 0.1000], + device='cuda:0'), in_proj_covar=tensor([0.0464, 0.0526, 0.0677, 0.0417, 0.0426, 0.0486, 0.0517, 0.0648], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 10:45:54,225 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=64124.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:45:56,456 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=64126.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:45:59,675 INFO [train.py:898] (0/4) Epoch 18, batch 2350, loss[loss=0.1949, simple_loss=0.281, pruned_loss=0.05444, over 12939.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2563, pruned_loss=0.0399, over 3561155.95 frames. ], batch size: 131, lr: 6.26e-03, grad_scale: 8.0 +2023-03-09 10:46:43,626 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64166.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:46:45,504 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.074e+02 2.798e+02 3.262e+02 4.005e+02 1.022e+03, threshold=6.524e+02, percent-clipped=3.0 +2023-03-09 10:46:58,408 INFO [train.py:898] (0/4) Epoch 18, batch 2400, loss[loss=0.1562, simple_loss=0.244, pruned_loss=0.03416, over 18506.00 frames. ], tot_loss[loss=0.1681, simple_loss=0.2568, pruned_loss=0.03975, over 3565126.25 frames. ], batch size: 47, lr: 6.26e-03, grad_scale: 8.0 +2023-03-09 10:47:24,689 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64201.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:47:53,537 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64225.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:47:58,506 INFO [train.py:898] (0/4) Epoch 18, batch 2450, loss[loss=0.1852, simple_loss=0.2688, pruned_loss=0.05082, over 18247.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2565, pruned_loss=0.03967, over 3567004.05 frames. ], batch size: 60, lr: 6.26e-03, grad_scale: 8.0 +2023-03-09 10:48:02,200 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64232.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:48:14,178 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64242.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:48:44,840 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.830e+02 2.733e+02 3.243e+02 4.040e+02 1.212e+03, threshold=6.487e+02, percent-clipped=5.0 +2023-03-09 10:48:49,126 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64272.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:48:50,272 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=64273.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:48:57,008 INFO [train.py:898] (0/4) Epoch 18, batch 2500, loss[loss=0.1602, simple_loss=0.2413, pruned_loss=0.03954, over 18492.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2569, pruned_loss=0.03973, over 3567155.29 frames. ], batch size: 47, lr: 6.25e-03, grad_scale: 4.0 +2023-03-09 10:49:10,366 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=64290.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:49:14,620 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64293.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:49:30,851 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.24 vs. limit=5.0 +2023-03-09 10:49:56,110 INFO [train.py:898] (0/4) Epoch 18, batch 2550, loss[loss=0.1702, simple_loss=0.2632, pruned_loss=0.03867, over 17830.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.257, pruned_loss=0.03975, over 3567494.60 frames. ], batch size: 70, lr: 6.25e-03, grad_scale: 4.0 +2023-03-09 10:50:00,604 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2530, 4.4933, 2.7886, 4.2204, 5.4995, 2.9585, 4.3146, 4.3508], + device='cuda:0'), covar=tensor([0.0131, 0.1006, 0.1425, 0.0568, 0.0058, 0.1057, 0.0494, 0.0540], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0261, 0.0201, 0.0194, 0.0114, 0.0180, 0.0212, 0.0220], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 10:50:10,180 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8067, 5.4011, 5.3441, 5.3692, 4.8668, 5.2490, 4.6588, 5.2507], + device='cuda:0'), covar=tensor([0.0267, 0.0269, 0.0195, 0.0345, 0.0425, 0.0228, 0.1149, 0.0262], + device='cuda:0'), in_proj_covar=tensor([0.0208, 0.0254, 0.0244, 0.0310, 0.0262, 0.0257, 0.0302, 0.0250], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 10:50:11,419 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7780, 4.4724, 4.5832, 3.4031, 3.7010, 3.4503, 2.6430, 2.5231], + device='cuda:0'), covar=tensor([0.0249, 0.0160, 0.0070, 0.0304, 0.0359, 0.0232, 0.0743, 0.0874], + device='cuda:0'), in_proj_covar=tensor([0.0068, 0.0057, 0.0060, 0.0067, 0.0088, 0.0065, 0.0076, 0.0084], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 10:50:13,617 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9269, 5.1705, 2.5346, 4.9489, 4.8949, 5.2127, 4.9475, 2.5861], + device='cuda:0'), covar=tensor([0.0215, 0.0061, 0.0846, 0.0097, 0.0069, 0.0064, 0.0092, 0.1057], + device='cuda:0'), in_proj_covar=tensor([0.0086, 0.0079, 0.0095, 0.0093, 0.0083, 0.0074, 0.0083, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 10:50:15,813 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64345.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:50:43,210 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.795e+02 2.856e+02 3.469e+02 4.229e+02 8.675e+02, threshold=6.938e+02, percent-clipped=2.0 +2023-03-09 10:50:54,971 INFO [train.py:898] (0/4) Epoch 18, batch 2600, loss[loss=0.1628, simple_loss=0.2512, pruned_loss=0.03723, over 18273.00 frames. ], tot_loss[loss=0.168, simple_loss=0.2569, pruned_loss=0.03961, over 3574027.32 frames. ], batch size: 47, lr: 6.25e-03, grad_scale: 4.0 +2023-03-09 10:50:55,164 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64379.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:51:27,164 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64406.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:51:27,549 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-09 10:51:41,003 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4028, 5.9994, 5.5691, 5.7241, 5.5672, 5.3837, 6.0214, 5.9874], + device='cuda:0'), covar=tensor([0.1194, 0.0756, 0.0396, 0.0730, 0.1367, 0.0653, 0.0516, 0.0686], + device='cuda:0'), in_proj_covar=tensor([0.0596, 0.0508, 0.0369, 0.0530, 0.0732, 0.0531, 0.0718, 0.0549], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 10:51:53,368 INFO [train.py:898] (0/4) Epoch 18, batch 2650, loss[loss=0.1715, simple_loss=0.269, pruned_loss=0.037, over 18503.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2577, pruned_loss=0.04002, over 3570571.20 frames. ], batch size: 47, lr: 6.25e-03, grad_scale: 4.0 +2023-03-09 10:52:35,740 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64465.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:52:36,750 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64466.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:52:39,836 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.908e+02 2.670e+02 3.183e+02 3.846e+02 7.889e+02, threshold=6.366e+02, percent-clipped=2.0 +2023-03-09 10:52:52,153 INFO [train.py:898] (0/4) Epoch 18, batch 2700, loss[loss=0.1686, simple_loss=0.2616, pruned_loss=0.03779, over 18572.00 frames. ], tot_loss[loss=0.1693, simple_loss=0.2581, pruned_loss=0.04024, over 3575872.30 frames. ], batch size: 54, lr: 6.24e-03, grad_scale: 4.0 +2023-03-09 10:53:04,815 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4119, 6.0253, 5.5428, 5.7786, 5.6119, 5.4312, 6.0652, 5.9721], + device='cuda:0'), covar=tensor([0.1305, 0.0694, 0.0517, 0.0695, 0.1392, 0.0682, 0.0566, 0.0736], + device='cuda:0'), in_proj_covar=tensor([0.0600, 0.0510, 0.0371, 0.0534, 0.0733, 0.0533, 0.0726, 0.0551], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 10:53:18,427 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64501.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:53:32,914 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=64514.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:53:47,604 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64526.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:53:50,696 INFO [train.py:898] (0/4) Epoch 18, batch 2750, loss[loss=0.156, simple_loss=0.2502, pruned_loss=0.03091, over 18495.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2577, pruned_loss=0.03997, over 3590443.53 frames. ], batch size: 51, lr: 6.24e-03, grad_scale: 4.0 +2023-03-09 10:54:14,654 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=64549.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:54:37,407 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.946e+02 2.775e+02 3.277e+02 4.138e+02 9.183e+02, threshold=6.554e+02, percent-clipped=4.0 +2023-03-09 10:54:41,141 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64572.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:54:49,211 INFO [train.py:898] (0/4) Epoch 18, batch 2800, loss[loss=0.1405, simple_loss=0.2207, pruned_loss=0.03015, over 18395.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2577, pruned_loss=0.04001, over 3588152.58 frames. ], batch size: 42, lr: 6.24e-03, grad_scale: 8.0 +2023-03-09 10:55:00,298 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64588.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:55:01,566 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64589.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:55:37,326 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=64620.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:55:47,426 INFO [train.py:898] (0/4) Epoch 18, batch 2850, loss[loss=0.1607, simple_loss=0.2451, pruned_loss=0.03811, over 18499.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2578, pruned_loss=0.04013, over 3597546.95 frames. ], batch size: 47, lr: 6.24e-03, grad_scale: 8.0 +2023-03-09 10:56:10,253 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64648.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:56:13,211 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64650.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:56:34,846 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.051e+02 2.756e+02 3.294e+02 3.793e+02 6.493e+02, threshold=6.588e+02, percent-clipped=0.0 +2023-03-09 10:56:46,011 INFO [train.py:898] (0/4) Epoch 18, batch 2900, loss[loss=0.1602, simple_loss=0.246, pruned_loss=0.03718, over 18286.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2579, pruned_loss=0.03991, over 3604712.82 frames. ], batch size: 49, lr: 6.23e-03, grad_scale: 8.0 +2023-03-09 10:56:46,367 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64679.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:57:12,495 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64701.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:57:14,890 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64703.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:57:21,887 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64709.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:57:39,622 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-09 10:57:42,511 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=64727.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:57:44,567 INFO [train.py:898] (0/4) Epoch 18, batch 2950, loss[loss=0.1701, simple_loss=0.2552, pruned_loss=0.04252, over 18519.00 frames. ], tot_loss[loss=0.1689, simple_loss=0.2579, pruned_loss=0.03998, over 3600253.76 frames. ], batch size: 44, lr: 6.23e-03, grad_scale: 8.0 +2023-03-09 10:58:03,593 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64745.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:58:12,599 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64752.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:58:14,247 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.78 vs. limit=2.0 +2023-03-09 10:58:26,517 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64764.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:58:31,864 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.931e+02 2.790e+02 3.324e+02 3.890e+02 1.048e+03, threshold=6.648e+02, percent-clipped=3.0 +2023-03-09 10:58:43,037 INFO [train.py:898] (0/4) Epoch 18, batch 3000, loss[loss=0.1585, simple_loss=0.2461, pruned_loss=0.03542, over 18324.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.257, pruned_loss=0.04, over 3582810.54 frames. ], batch size: 49, lr: 6.23e-03, grad_scale: 8.0 +2023-03-09 10:58:43,039 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 10:58:55,126 INFO [train.py:932] (0/4) Epoch 18, validation: loss=0.1513, simple_loss=0.2515, pruned_loss=0.02557, over 944034.00 frames. +2023-03-09 10:58:55,127 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 10:59:19,035 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-09 10:59:28,046 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64806.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:59:36,008 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=64813.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:59:44,853 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64821.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 10:59:53,893 INFO [train.py:898] (0/4) Epoch 18, batch 3050, loss[loss=0.1715, simple_loss=0.2596, pruned_loss=0.0417, over 18545.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2578, pruned_loss=0.04014, over 3588311.45 frames. ], batch size: 49, lr: 6.23e-03, grad_scale: 8.0 +2023-03-09 11:00:41,792 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.847e+02 2.582e+02 2.978e+02 3.625e+02 9.287e+02, threshold=5.956e+02, percent-clipped=2.0 +2023-03-09 11:00:52,627 INFO [train.py:898] (0/4) Epoch 18, batch 3100, loss[loss=0.2064, simple_loss=0.2944, pruned_loss=0.05921, over 18295.00 frames. ], tot_loss[loss=0.1692, simple_loss=0.2578, pruned_loss=0.04031, over 3596631.62 frames. ], batch size: 57, lr: 6.22e-03, grad_scale: 8.0 +2023-03-09 11:01:03,667 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=64888.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:01:25,399 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.43 vs. limit=2.0 +2023-03-09 11:01:51,079 INFO [train.py:898] (0/4) Epoch 18, batch 3150, loss[loss=0.1766, simple_loss=0.2669, pruned_loss=0.04318, over 18628.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2576, pruned_loss=0.04018, over 3606199.28 frames. ], batch size: 52, lr: 6.22e-03, grad_scale: 8.0 +2023-03-09 11:01:58,995 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=64936.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:02:10,192 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=64945.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:02:38,246 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.946e+02 2.751e+02 3.227e+02 4.235e+02 1.394e+03, threshold=6.453e+02, percent-clipped=7.0 +2023-03-09 11:02:46,503 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=64976.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:02:49,635 INFO [train.py:898] (0/4) Epoch 18, batch 3200, loss[loss=0.1491, simple_loss=0.2321, pruned_loss=0.03307, over 18243.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2567, pruned_loss=0.04001, over 3599444.69 frames. ], batch size: 45, lr: 6.22e-03, grad_scale: 8.0 +2023-03-09 11:03:16,774 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65001.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:03:20,114 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65004.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:03:48,367 INFO [train.py:898] (0/4) Epoch 18, batch 3250, loss[loss=0.1651, simple_loss=0.2529, pruned_loss=0.03866, over 18405.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2571, pruned_loss=0.04017, over 3593921.05 frames. ], batch size: 52, lr: 6.22e-03, grad_scale: 8.0 +2023-03-09 11:03:57,985 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65037.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 11:04:12,026 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=65049.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:04:18,531 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5650, 3.8947, 5.1492, 4.3530, 2.7701, 2.7040, 4.3300, 5.3144], + device='cuda:0'), covar=tensor([0.0865, 0.1454, 0.0139, 0.0432, 0.1175, 0.1333, 0.0437, 0.0170], + device='cuda:0'), in_proj_covar=tensor([0.0146, 0.0271, 0.0140, 0.0179, 0.0188, 0.0188, 0.0191, 0.0186], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 11:04:19,578 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65055.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:04:24,025 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65059.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:04:28,065 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-09 11:04:35,617 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.699e+02 2.719e+02 3.291e+02 4.073e+02 7.543e+02, threshold=6.581e+02, percent-clipped=1.0 +2023-03-09 11:04:46,973 INFO [train.py:898] (0/4) Epoch 18, batch 3300, loss[loss=0.1889, simple_loss=0.2791, pruned_loss=0.04936, over 16224.00 frames. ], tot_loss[loss=0.169, simple_loss=0.2576, pruned_loss=0.04017, over 3592226.95 frames. ], batch size: 94, lr: 6.21e-03, grad_scale: 4.0 +2023-03-09 11:05:13,339 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65101.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:05:21,795 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65108.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:05:31,461 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65116.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:05:32,642 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8922, 4.6399, 4.6835, 3.4268, 3.8250, 3.4662, 2.6759, 2.5866], + device='cuda:0'), covar=tensor([0.0220, 0.0174, 0.0071, 0.0311, 0.0343, 0.0256, 0.0749, 0.0851], + device='cuda:0'), in_proj_covar=tensor([0.0068, 0.0057, 0.0060, 0.0067, 0.0087, 0.0066, 0.0076, 0.0083], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 11:05:37,634 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65121.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:05:46,560 INFO [train.py:898] (0/4) Epoch 18, batch 3350, loss[loss=0.1437, simple_loss=0.2216, pruned_loss=0.03289, over 18387.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.257, pruned_loss=0.03984, over 3591165.77 frames. ], batch size: 42, lr: 6.21e-03, grad_scale: 4.0 +2023-03-09 11:06:33,068 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=65169.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:06:34,019 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.758e+02 2.720e+02 3.418e+02 4.125e+02 2.086e+03, threshold=6.835e+02, percent-clipped=5.0 +2023-03-09 11:06:44,720 INFO [train.py:898] (0/4) Epoch 18, batch 3400, loss[loss=0.1571, simple_loss=0.2444, pruned_loss=0.03496, over 18269.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.257, pruned_loss=0.03994, over 3580638.29 frames. ], batch size: 49, lr: 6.21e-03, grad_scale: 4.0 +2023-03-09 11:07:43,149 INFO [train.py:898] (0/4) Epoch 18, batch 3450, loss[loss=0.1676, simple_loss=0.2578, pruned_loss=0.03872, over 16950.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2574, pruned_loss=0.04015, over 3590170.36 frames. ], batch size: 78, lr: 6.21e-03, grad_scale: 4.0 +2023-03-09 11:07:43,921 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.51 vs. limit=5.0 +2023-03-09 11:08:01,631 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65245.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:08:24,756 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1049, 5.1088, 5.1937, 4.9349, 4.8963, 4.9723, 5.2622, 5.2742], + device='cuda:0'), covar=tensor([0.0059, 0.0049, 0.0044, 0.0083, 0.0058, 0.0120, 0.0062, 0.0066], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0067, 0.0071, 0.0090, 0.0073, 0.0101, 0.0085, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 11:08:31,045 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.103e+02 2.885e+02 3.344e+02 3.937e+02 8.073e+02, threshold=6.688e+02, percent-clipped=3.0 +2023-03-09 11:08:37,235 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7134, 3.5769, 4.9605, 4.3110, 3.1022, 2.9050, 4.2434, 5.1367], + device='cuda:0'), covar=tensor([0.0893, 0.1452, 0.0187, 0.0417, 0.1113, 0.1323, 0.0453, 0.0203], + device='cuda:0'), in_proj_covar=tensor([0.0147, 0.0273, 0.0142, 0.0181, 0.0191, 0.0190, 0.0193, 0.0189], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 11:08:41,985 INFO [train.py:898] (0/4) Epoch 18, batch 3500, loss[loss=0.1895, simple_loss=0.2691, pruned_loss=0.05496, over 12763.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2573, pruned_loss=0.0401, over 3588435.57 frames. ], batch size: 129, lr: 6.20e-03, grad_scale: 4.0 +2023-03-09 11:08:48,919 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65285.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:08:58,095 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=65293.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:09:10,737 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65304.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:09:17,642 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6564, 2.3062, 2.6641, 2.7461, 3.3529, 5.0421, 4.7473, 3.4575], + device='cuda:0'), covar=tensor([0.1742, 0.2313, 0.2980, 0.1753, 0.2145, 0.0174, 0.0376, 0.0883], + device='cuda:0'), in_proj_covar=tensor([0.0291, 0.0343, 0.0373, 0.0274, 0.0391, 0.0234, 0.0295, 0.0250], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 11:09:37,553 INFO [train.py:898] (0/4) Epoch 18, batch 3550, loss[loss=0.1492, simple_loss=0.2349, pruned_loss=0.03172, over 18437.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2575, pruned_loss=0.04007, over 3582672.75 frames. ], batch size: 43, lr: 6.20e-03, grad_scale: 4.0 +2023-03-09 11:09:41,520 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65332.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 11:09:57,198 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65346.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:10:03,342 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=65352.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:10:10,897 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65359.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:10:23,169 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.008e+02 2.567e+02 2.931e+02 3.482e+02 6.619e+02, threshold=5.862e+02, percent-clipped=0.0 +2023-03-09 11:10:31,678 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.85 vs. limit=2.0 +2023-03-09 11:10:33,169 INFO [train.py:898] (0/4) Epoch 18, batch 3600, loss[loss=0.1589, simple_loss=0.25, pruned_loss=0.03392, over 18388.00 frames. ], tot_loss[loss=0.1688, simple_loss=0.2576, pruned_loss=0.04004, over 3589604.21 frames. ], batch size: 50, lr: 6.20e-03, grad_scale: 8.0 +2023-03-09 11:10:53,603 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-09 11:10:57,524 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65401.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:11:03,635 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=65407.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:11:04,724 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65408.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:11:07,296 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65411.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:11:09,332 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-18.pt +2023-03-09 11:11:38,110 INFO [train.py:898] (0/4) Epoch 19, batch 0, loss[loss=0.1732, simple_loss=0.2655, pruned_loss=0.04043, over 18110.00 frames. ], tot_loss[loss=0.1732, simple_loss=0.2655, pruned_loss=0.04043, over 18110.00 frames. ], batch size: 62, lr: 6.03e-03, grad_scale: 8.0 +2023-03-09 11:11:38,112 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 11:11:49,830 INFO [train.py:932] (0/4) Epoch 19, validation: loss=0.1513, simple_loss=0.2518, pruned_loss=0.02538, over 944034.00 frames. +2023-03-09 11:11:49,830 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 11:11:55,693 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5590, 5.2963, 5.7972, 5.7254, 5.4582, 6.2736, 5.9187, 5.6174], + device='cuda:0'), covar=tensor([0.1049, 0.0531, 0.0658, 0.0705, 0.1322, 0.0603, 0.0587, 0.1542], + device='cuda:0'), in_proj_covar=tensor([0.0348, 0.0273, 0.0297, 0.0297, 0.0323, 0.0408, 0.0271, 0.0397], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 11:12:32,043 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=65449.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:12:40,038 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=65456.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:12:47,636 INFO [train.py:898] (0/4) Epoch 19, batch 50, loss[loss=0.1347, simple_loss=0.2196, pruned_loss=0.02492, over 18504.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2563, pruned_loss=0.03947, over 814094.00 frames. ], batch size: 44, lr: 6.03e-03, grad_scale: 8.0 +2023-03-09 11:12:55,631 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.183e+02 3.168e+02 3.745e+02 4.486e+02 9.575e+02, threshold=7.490e+02, percent-clipped=6.0 +2023-03-09 11:12:58,291 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5841, 2.4786, 4.2958, 3.9737, 2.2680, 4.4885, 3.8106, 2.7171], + device='cuda:0'), covar=tensor([0.0462, 0.1971, 0.0241, 0.0281, 0.2146, 0.0265, 0.0625, 0.1299], + device='cuda:0'), in_proj_covar=tensor([0.0198, 0.0225, 0.0193, 0.0151, 0.0214, 0.0197, 0.0232, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 11:13:46,481 INFO [train.py:898] (0/4) Epoch 19, batch 100, loss[loss=0.1698, simple_loss=0.2584, pruned_loss=0.04056, over 18211.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.256, pruned_loss=0.03964, over 1418454.50 frames. ], batch size: 60, lr: 6.03e-03, grad_scale: 8.0 +2023-03-09 11:14:08,280 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65532.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 11:14:44,711 INFO [train.py:898] (0/4) Epoch 19, batch 150, loss[loss=0.1903, simple_loss=0.2818, pruned_loss=0.0494, over 18480.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2556, pruned_loss=0.03902, over 1915247.57 frames. ], batch size: 53, lr: 6.02e-03, grad_scale: 4.0 +2023-03-09 11:14:53,873 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.910e+02 2.761e+02 3.150e+02 3.838e+02 8.694e+02, threshold=6.300e+02, percent-clipped=3.0 +2023-03-09 11:15:20,677 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65593.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 11:15:44,540 INFO [train.py:898] (0/4) Epoch 19, batch 200, loss[loss=0.1451, simple_loss=0.2356, pruned_loss=0.02724, over 18374.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2548, pruned_loss=0.03839, over 2297213.88 frames. ], batch size: 50, lr: 6.02e-03, grad_scale: 4.0 +2023-03-09 11:15:57,227 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6703, 4.0098, 2.4306, 3.8861, 4.9300, 2.4900, 3.5079, 3.8676], + device='cuda:0'), covar=tensor([0.0187, 0.1110, 0.1581, 0.0609, 0.0089, 0.1247, 0.0786, 0.0670], + device='cuda:0'), in_proj_covar=tensor([0.0157, 0.0264, 0.0202, 0.0195, 0.0117, 0.0181, 0.0213, 0.0222], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 11:16:06,570 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65632.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:16:17,200 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65641.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:16:39,507 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-09 11:16:43,290 INFO [train.py:898] (0/4) Epoch 19, batch 250, loss[loss=0.1541, simple_loss=0.2388, pruned_loss=0.03471, over 18279.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2549, pruned_loss=0.03862, over 2592289.63 frames. ], batch size: 49, lr: 6.02e-03, grad_scale: 4.0 +2023-03-09 11:16:51,829 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7465, 2.2822, 2.6940, 2.6441, 3.1780, 4.9339, 4.5949, 3.5171], + device='cuda:0'), covar=tensor([0.1709, 0.2514, 0.2919, 0.1936, 0.2395, 0.0217, 0.0433, 0.0853], + device='cuda:0'), in_proj_covar=tensor([0.0290, 0.0341, 0.0371, 0.0272, 0.0387, 0.0233, 0.0293, 0.0247], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 11:16:52,480 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.860e+02 2.820e+02 3.379e+02 4.029e+02 8.035e+02, threshold=6.759e+02, percent-clipped=4.0 +2023-03-09 11:17:02,880 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=65680.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:17:31,763 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1395, 4.4039, 2.6027, 4.2281, 5.3979, 2.7163, 4.0225, 4.2343], + device='cuda:0'), covar=tensor([0.0162, 0.1083, 0.1633, 0.0675, 0.0073, 0.1337, 0.0643, 0.0624], + device='cuda:0'), in_proj_covar=tensor([0.0156, 0.0262, 0.0200, 0.0194, 0.0116, 0.0179, 0.0212, 0.0220], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 11:17:40,395 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65711.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:17:42,322 INFO [train.py:898] (0/4) Epoch 19, batch 300, loss[loss=0.1765, simple_loss=0.2638, pruned_loss=0.04459, over 18337.00 frames. ], tot_loss[loss=0.1682, simple_loss=0.2573, pruned_loss=0.03957, over 2791806.29 frames. ], batch size: 56, lr: 6.02e-03, grad_scale: 4.0 +2023-03-09 11:18:24,063 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-09 11:18:35,642 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=65759.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:18:40,556 INFO [train.py:898] (0/4) Epoch 19, batch 350, loss[loss=0.1892, simple_loss=0.2769, pruned_loss=0.05071, over 18104.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2573, pruned_loss=0.04006, over 2968539.33 frames. ], batch size: 62, lr: 6.01e-03, grad_scale: 4.0 +2023-03-09 11:18:49,493 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.996e+02 2.649e+02 2.976e+02 3.551e+02 5.920e+02, threshold=5.952e+02, percent-clipped=0.0 +2023-03-09 11:18:51,273 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.18 vs. limit=5.0 +2023-03-09 11:19:38,371 INFO [train.py:898] (0/4) Epoch 19, batch 400, loss[loss=0.1562, simple_loss=0.2495, pruned_loss=0.03145, over 18286.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.2582, pruned_loss=0.04051, over 3093717.13 frames. ], batch size: 49, lr: 6.01e-03, grad_scale: 8.0 +2023-03-09 11:20:07,704 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=65838.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 11:20:08,241 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-03-09 11:20:12,443 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-09 11:20:37,148 INFO [train.py:898] (0/4) Epoch 19, batch 450, loss[loss=0.1597, simple_loss=0.2369, pruned_loss=0.0412, over 18243.00 frames. ], tot_loss[loss=0.1696, simple_loss=0.258, pruned_loss=0.0406, over 3202156.82 frames. ], batch size: 45, lr: 6.01e-03, grad_scale: 8.0 +2023-03-09 11:20:45,871 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4285, 5.9900, 5.4469, 5.7612, 5.5529, 5.3951, 6.0173, 5.9841], + device='cuda:0'), covar=tensor([0.1147, 0.0625, 0.0548, 0.0679, 0.1371, 0.0630, 0.0563, 0.0607], + device='cuda:0'), in_proj_covar=tensor([0.0595, 0.0512, 0.0368, 0.0535, 0.0729, 0.0532, 0.0726, 0.0547], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 11:20:46,703 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.864e+02 3.029e+02 3.525e+02 4.091e+02 6.836e+02, threshold=7.049e+02, percent-clipped=6.0 +2023-03-09 11:21:05,981 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=65888.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 11:21:18,693 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=65899.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 11:21:35,174 INFO [train.py:898] (0/4) Epoch 19, batch 500, loss[loss=0.1604, simple_loss=0.254, pruned_loss=0.03337, over 18506.00 frames. ], tot_loss[loss=0.1686, simple_loss=0.2568, pruned_loss=0.04019, over 3278792.66 frames. ], batch size: 53, lr: 6.01e-03, grad_scale: 8.0 +2023-03-09 11:22:05,185 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6780, 2.2648, 2.6529, 2.6899, 3.1910, 4.8195, 4.6206, 3.2862], + device='cuda:0'), covar=tensor([0.1771, 0.2401, 0.2906, 0.1855, 0.2295, 0.0221, 0.0380, 0.0946], + device='cuda:0'), in_proj_covar=tensor([0.0291, 0.0341, 0.0372, 0.0272, 0.0389, 0.0234, 0.0294, 0.0248], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 11:22:08,332 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=65941.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:22:33,524 INFO [train.py:898] (0/4) Epoch 19, batch 550, loss[loss=0.1605, simple_loss=0.2455, pruned_loss=0.03775, over 18297.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2568, pruned_loss=0.03996, over 3351696.09 frames. ], batch size: 49, lr: 6.01e-03, grad_scale: 8.0 +2023-03-09 11:22:42,912 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.825e+02 2.542e+02 3.090e+02 3.456e+02 5.520e+02, threshold=6.179e+02, percent-clipped=0.0 +2023-03-09 11:23:04,291 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=65989.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:23:16,542 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-66000.pt +2023-03-09 11:23:35,291 INFO [train.py:898] (0/4) Epoch 19, batch 600, loss[loss=0.185, simple_loss=0.2721, pruned_loss=0.04892, over 17793.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.257, pruned_loss=0.03989, over 3409789.24 frames. ], batch size: 70, lr: 6.00e-03, grad_scale: 8.0 +2023-03-09 11:24:14,083 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2121, 5.0681, 5.4339, 5.4493, 5.1509, 5.9569, 5.6133, 5.2678], + device='cuda:0'), covar=tensor([0.1083, 0.0673, 0.0778, 0.0675, 0.1483, 0.0717, 0.0660, 0.1590], + device='cuda:0'), in_proj_covar=tensor([0.0352, 0.0275, 0.0301, 0.0301, 0.0326, 0.0414, 0.0273, 0.0401], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 11:24:25,578 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7240, 3.6397, 4.9076, 4.3850, 3.0995, 3.0066, 4.4717, 5.2040], + device='cuda:0'), covar=tensor([0.0809, 0.1392, 0.0194, 0.0353, 0.1000, 0.1119, 0.0337, 0.0235], + device='cuda:0'), in_proj_covar=tensor([0.0147, 0.0272, 0.0143, 0.0179, 0.0190, 0.0188, 0.0192, 0.0189], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 11:24:34,014 INFO [train.py:898] (0/4) Epoch 19, batch 650, loss[loss=0.1602, simple_loss=0.2554, pruned_loss=0.03247, over 18002.00 frames. ], tot_loss[loss=0.1683, simple_loss=0.2573, pruned_loss=0.03962, over 3446551.93 frames. ], batch size: 65, lr: 6.00e-03, grad_scale: 8.0 +2023-03-09 11:24:42,669 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.920e+02 2.591e+02 2.961e+02 3.650e+02 8.631e+02, threshold=5.923e+02, percent-clipped=2.0 +2023-03-09 11:24:58,435 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5558, 6.1325, 5.6384, 5.9126, 5.7276, 5.5235, 6.1915, 6.1225], + device='cuda:0'), covar=tensor([0.1178, 0.0662, 0.0383, 0.0699, 0.1343, 0.0723, 0.0537, 0.0647], + device='cuda:0'), in_proj_covar=tensor([0.0601, 0.0517, 0.0372, 0.0541, 0.0737, 0.0538, 0.0733, 0.0551], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 11:25:32,784 INFO [train.py:898] (0/4) Epoch 19, batch 700, loss[loss=0.1593, simple_loss=0.2559, pruned_loss=0.03131, over 18383.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2555, pruned_loss=0.03927, over 3481190.39 frames. ], batch size: 52, lr: 6.00e-03, grad_scale: 8.0 +2023-03-09 11:26:04,184 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3125, 5.2874, 4.9113, 5.2078, 5.2089, 4.6307, 5.1239, 4.9183], + device='cuda:0'), covar=tensor([0.0346, 0.0416, 0.1189, 0.0706, 0.0530, 0.0382, 0.0372, 0.0910], + device='cuda:0'), in_proj_covar=tensor([0.0460, 0.0533, 0.0680, 0.0417, 0.0427, 0.0486, 0.0523, 0.0655], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 11:26:09,880 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66144.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:26:27,957 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66160.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 11:26:31,031 INFO [train.py:898] (0/4) Epoch 19, batch 750, loss[loss=0.1585, simple_loss=0.2361, pruned_loss=0.04046, over 18235.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2555, pruned_loss=0.03935, over 3505103.99 frames. ], batch size: 45, lr: 6.00e-03, grad_scale: 8.0 +2023-03-09 11:26:40,055 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.858e+02 2.715e+02 3.373e+02 4.119e+02 8.892e+02, threshold=6.747e+02, percent-clipped=6.0 +2023-03-09 11:27:01,488 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66188.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 11:27:08,701 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66194.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 11:27:21,497 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66205.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 11:27:21,616 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5845, 2.1416, 2.5600, 2.6511, 3.0196, 4.9100, 4.6507, 3.6740], + device='cuda:0'), covar=tensor([0.1829, 0.2644, 0.3145, 0.1908, 0.2580, 0.0196, 0.0376, 0.0723], + device='cuda:0'), in_proj_covar=tensor([0.0293, 0.0343, 0.0373, 0.0273, 0.0390, 0.0234, 0.0296, 0.0249], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 11:27:30,196 INFO [train.py:898] (0/4) Epoch 19, batch 800, loss[loss=0.1642, simple_loss=0.2496, pruned_loss=0.03938, over 18265.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2557, pruned_loss=0.03957, over 3521888.58 frames. ], batch size: 47, lr: 5.99e-03, grad_scale: 8.0 +2023-03-09 11:27:39,571 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66221.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 11:27:58,023 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=66236.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 11:28:23,777 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66258.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:28:29,019 INFO [train.py:898] (0/4) Epoch 19, batch 850, loss[loss=0.152, simple_loss=0.2271, pruned_loss=0.03842, over 18425.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2566, pruned_loss=0.03936, over 3536005.52 frames. ], batch size: 43, lr: 5.99e-03, grad_scale: 8.0 +2023-03-09 11:28:34,463 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-09 11:28:37,652 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8759, 3.7339, 4.9265, 4.4341, 3.2454, 3.0900, 4.3552, 5.2344], + device='cuda:0'), covar=tensor([0.0776, 0.1342, 0.0243, 0.0353, 0.0902, 0.1090, 0.0395, 0.0159], + device='cuda:0'), in_proj_covar=tensor([0.0145, 0.0269, 0.0141, 0.0178, 0.0188, 0.0186, 0.0190, 0.0186], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 11:28:38,315 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.829e+02 2.659e+02 3.037e+02 3.502e+02 7.094e+02, threshold=6.073e+02, percent-clipped=1.0 +2023-03-09 11:29:27,611 INFO [train.py:898] (0/4) Epoch 19, batch 900, loss[loss=0.1492, simple_loss=0.2391, pruned_loss=0.02968, over 18501.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2565, pruned_loss=0.03918, over 3551838.55 frames. ], batch size: 51, lr: 5.99e-03, grad_scale: 8.0 +2023-03-09 11:29:34,874 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66319.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:29:52,722 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0955, 4.2171, 2.5462, 4.1462, 5.3476, 2.6182, 4.1805, 4.3652], + device='cuda:0'), covar=tensor([0.0135, 0.0925, 0.1407, 0.0543, 0.0064, 0.1092, 0.0483, 0.0495], + device='cuda:0'), in_proj_covar=tensor([0.0155, 0.0262, 0.0199, 0.0193, 0.0117, 0.0179, 0.0212, 0.0220], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 11:30:26,351 INFO [train.py:898] (0/4) Epoch 19, batch 950, loss[loss=0.1701, simple_loss=0.2551, pruned_loss=0.0425, over 18544.00 frames. ], tot_loss[loss=0.1679, simple_loss=0.2569, pruned_loss=0.03949, over 3555454.50 frames. ], batch size: 49, lr: 5.99e-03, grad_scale: 8.0 +2023-03-09 11:30:35,531 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.870e+02 2.786e+02 3.276e+02 3.823e+02 6.453e+02, threshold=6.553e+02, percent-clipped=1.0 +2023-03-09 11:31:01,951 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8854, 4.6361, 4.6597, 3.6934, 3.9819, 3.5982, 2.9944, 2.6281], + device='cuda:0'), covar=tensor([0.0176, 0.0103, 0.0068, 0.0239, 0.0298, 0.0199, 0.0586, 0.0746], + device='cuda:0'), in_proj_covar=tensor([0.0069, 0.0058, 0.0060, 0.0068, 0.0087, 0.0066, 0.0076, 0.0084], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 11:31:15,173 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.59 vs. limit=5.0 +2023-03-09 11:31:24,775 INFO [train.py:898] (0/4) Epoch 19, batch 1000, loss[loss=0.1726, simple_loss=0.262, pruned_loss=0.04159, over 18343.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2562, pruned_loss=0.03917, over 3565171.05 frames. ], batch size: 55, lr: 5.99e-03, grad_scale: 8.0 +2023-03-09 11:31:49,948 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9201, 4.5677, 4.6146, 3.3919, 3.8281, 3.4921, 3.0540, 2.5579], + device='cuda:0'), covar=tensor([0.0183, 0.0114, 0.0072, 0.0309, 0.0284, 0.0221, 0.0561, 0.0792], + device='cuda:0'), in_proj_covar=tensor([0.0069, 0.0058, 0.0060, 0.0068, 0.0087, 0.0066, 0.0076, 0.0084], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 11:32:06,507 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3596, 2.6845, 2.3302, 2.7912, 3.4680, 3.3121, 2.9035, 2.7915], + device='cuda:0'), covar=tensor([0.0192, 0.0260, 0.0637, 0.0402, 0.0188, 0.0173, 0.0415, 0.0377], + device='cuda:0'), in_proj_covar=tensor([0.0135, 0.0131, 0.0163, 0.0153, 0.0126, 0.0111, 0.0151, 0.0152], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 11:32:16,067 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.96 vs. limit=5.0 +2023-03-09 11:32:20,708 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-09 11:32:23,329 INFO [train.py:898] (0/4) Epoch 19, batch 1050, loss[loss=0.1379, simple_loss=0.2236, pruned_loss=0.02608, over 18449.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.256, pruned_loss=0.03927, over 3576187.39 frames. ], batch size: 43, lr: 5.98e-03, grad_scale: 8.0 +2023-03-09 11:32:32,545 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.909e+02 2.933e+02 3.279e+02 4.197e+02 8.258e+02, threshold=6.558e+02, percent-clipped=3.0 +2023-03-09 11:32:59,542 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66494.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 11:33:06,844 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66500.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 11:33:22,931 INFO [train.py:898] (0/4) Epoch 19, batch 1100, loss[loss=0.1795, simple_loss=0.2743, pruned_loss=0.04238, over 17804.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.256, pruned_loss=0.03923, over 3577454.80 frames. ], batch size: 70, lr: 5.98e-03, grad_scale: 8.0 +2023-03-09 11:33:26,425 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66516.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 11:33:29,929 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9164, 5.3549, 5.2890, 5.3422, 4.8584, 5.2781, 4.5582, 5.2529], + device='cuda:0'), covar=tensor([0.0228, 0.0351, 0.0254, 0.0378, 0.0374, 0.0221, 0.1286, 0.0304], + device='cuda:0'), in_proj_covar=tensor([0.0210, 0.0256, 0.0245, 0.0320, 0.0263, 0.0262, 0.0305, 0.0252], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 11:33:31,997 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66521.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 11:33:39,996 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6665, 3.5203, 4.8388, 2.6404, 4.1612, 2.6074, 2.9463, 1.7673], + device='cuda:0'), covar=tensor([0.1276, 0.0953, 0.0154, 0.1019, 0.0670, 0.2569, 0.2640, 0.2210], + device='cuda:0'), in_proj_covar=tensor([0.0215, 0.0238, 0.0175, 0.0191, 0.0250, 0.0266, 0.0316, 0.0229], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 11:33:42,109 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0297, 5.0516, 5.0902, 4.8199, 4.8230, 4.8225, 5.1796, 5.1587], + device='cuda:0'), covar=tensor([0.0067, 0.0064, 0.0056, 0.0099, 0.0067, 0.0147, 0.0068, 0.0083], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0067, 0.0071, 0.0090, 0.0073, 0.0101, 0.0084, 0.0084], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 11:33:51,691 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.83 vs. limit=5.0 +2023-03-09 11:33:56,080 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=66542.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 11:34:19,092 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5279, 5.2901, 5.7139, 5.7261, 5.4432, 6.2728, 5.9457, 5.4970], + device='cuda:0'), covar=tensor([0.1046, 0.0625, 0.0676, 0.0724, 0.1445, 0.0683, 0.0542, 0.1762], + device='cuda:0'), in_proj_covar=tensor([0.0351, 0.0275, 0.0299, 0.0301, 0.0324, 0.0414, 0.0270, 0.0400], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 11:34:20,978 INFO [train.py:898] (0/4) Epoch 19, batch 1150, loss[loss=0.1665, simple_loss=0.2405, pruned_loss=0.04627, over 17685.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2562, pruned_loss=0.03924, over 3587553.04 frames. ], batch size: 39, lr: 5.98e-03, grad_scale: 8.0 +2023-03-09 11:34:29,828 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.816e+02 2.725e+02 3.159e+02 3.763e+02 6.148e+02, threshold=6.318e+02, percent-clipped=0.0 +2023-03-09 11:34:36,263 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-09 11:34:42,792 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66582.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 11:35:10,289 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6376, 3.5624, 2.1914, 4.4950, 3.2286, 4.4683, 2.6658, 4.1546], + device='cuda:0'), covar=tensor([0.0616, 0.0811, 0.1475, 0.0579, 0.0796, 0.0297, 0.1131, 0.0377], + device='cuda:0'), in_proj_covar=tensor([0.0209, 0.0220, 0.0186, 0.0274, 0.0188, 0.0258, 0.0198, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 11:35:19,963 INFO [train.py:898] (0/4) Epoch 19, batch 1200, loss[loss=0.1526, simple_loss=0.2385, pruned_loss=0.03341, over 18483.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2562, pruned_loss=0.03932, over 3587912.22 frames. ], batch size: 44, lr: 5.98e-03, grad_scale: 8.0 +2023-03-09 11:35:21,254 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66614.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:36:18,955 INFO [train.py:898] (0/4) Epoch 19, batch 1250, loss[loss=0.18, simple_loss=0.2775, pruned_loss=0.0412, over 18478.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2547, pruned_loss=0.03898, over 3589624.44 frames. ], batch size: 51, lr: 5.97e-03, grad_scale: 8.0 +2023-03-09 11:36:27,893 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.899e+02 2.792e+02 3.362e+02 4.146e+02 6.699e+02, threshold=6.725e+02, percent-clipped=2.0 +2023-03-09 11:37:03,949 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1959, 4.4580, 2.6139, 4.2762, 5.4216, 2.8345, 4.0762, 4.2360], + device='cuda:0'), covar=tensor([0.0145, 0.1027, 0.1532, 0.0598, 0.0069, 0.1058, 0.0579, 0.0635], + device='cuda:0'), in_proj_covar=tensor([0.0157, 0.0266, 0.0200, 0.0194, 0.0118, 0.0180, 0.0214, 0.0223], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 11:37:16,537 INFO [train.py:898] (0/4) Epoch 19, batch 1300, loss[loss=0.1507, simple_loss=0.2371, pruned_loss=0.03217, over 18264.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.256, pruned_loss=0.03925, over 3587904.57 frames. ], batch size: 47, lr: 5.97e-03, grad_scale: 8.0 +2023-03-09 11:37:21,942 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6700, 2.8779, 2.7098, 2.9822, 3.7239, 3.6966, 3.1714, 2.9335], + device='cuda:0'), covar=tensor([0.0154, 0.0266, 0.0479, 0.0366, 0.0157, 0.0127, 0.0337, 0.0359], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0131, 0.0161, 0.0154, 0.0125, 0.0110, 0.0149, 0.0151], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 11:37:37,964 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66731.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 11:37:40,274 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3677, 2.7034, 2.4749, 2.7882, 3.4881, 3.4682, 2.9408, 2.7264], + device='cuda:0'), covar=tensor([0.0215, 0.0269, 0.0545, 0.0399, 0.0174, 0.0150, 0.0388, 0.0400], + device='cuda:0'), in_proj_covar=tensor([0.0134, 0.0130, 0.0161, 0.0153, 0.0125, 0.0109, 0.0149, 0.0150], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 11:38:14,575 INFO [train.py:898] (0/4) Epoch 19, batch 1350, loss[loss=0.1464, simple_loss=0.227, pruned_loss=0.03292, over 18144.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2559, pruned_loss=0.03904, over 3592223.39 frames. ], batch size: 44, lr: 5.97e-03, grad_scale: 8.0 +2023-03-09 11:38:24,644 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.905e+02 2.817e+02 3.306e+02 3.967e+02 8.245e+02, threshold=6.612e+02, percent-clipped=2.0 +2023-03-09 11:38:48,256 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=66792.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 11:38:57,342 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66800.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 11:39:12,421 INFO [train.py:898] (0/4) Epoch 19, batch 1400, loss[loss=0.169, simple_loss=0.257, pruned_loss=0.04047, over 18326.00 frames. ], tot_loss[loss=0.1668, simple_loss=0.2558, pruned_loss=0.03891, over 3597734.50 frames. ], batch size: 55, lr: 5.97e-03, grad_scale: 8.0 +2023-03-09 11:39:16,768 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66816.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 11:39:54,025 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=66848.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:40:11,318 INFO [train.py:898] (0/4) Epoch 19, batch 1450, loss[loss=0.1642, simple_loss=0.2501, pruned_loss=0.03915, over 18350.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2571, pruned_loss=0.03914, over 3586660.13 frames. ], batch size: 46, lr: 5.97e-03, grad_scale: 8.0 +2023-03-09 11:40:12,629 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=66864.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 11:40:21,462 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.935e+02 2.720e+02 3.294e+02 4.147e+02 8.956e+02, threshold=6.588e+02, percent-clipped=4.0 +2023-03-09 11:40:28,838 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=66877.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 11:41:10,286 INFO [train.py:898] (0/4) Epoch 19, batch 1500, loss[loss=0.1793, simple_loss=0.2688, pruned_loss=0.04487, over 18302.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2568, pruned_loss=0.039, over 3587444.23 frames. ], batch size: 57, lr: 5.96e-03, grad_scale: 8.0 +2023-03-09 11:41:11,737 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=66914.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:41:33,117 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7275, 2.4072, 2.7197, 2.8479, 3.3176, 5.1393, 4.9907, 3.3273], + device='cuda:0'), covar=tensor([0.1687, 0.2303, 0.2881, 0.1650, 0.2202, 0.0174, 0.0320, 0.0958], + device='cuda:0'), in_proj_covar=tensor([0.0291, 0.0339, 0.0370, 0.0271, 0.0387, 0.0232, 0.0292, 0.0246], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 11:42:02,675 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=66957.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:42:08,436 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=66962.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:42:09,393 INFO [train.py:898] (0/4) Epoch 19, batch 1550, loss[loss=0.1797, simple_loss=0.2688, pruned_loss=0.04529, over 18073.00 frames. ], tot_loss[loss=0.167, simple_loss=0.2563, pruned_loss=0.03879, over 3587026.32 frames. ], batch size: 62, lr: 5.96e-03, grad_scale: 8.0 +2023-03-09 11:42:18,929 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.827e+02 2.644e+02 3.102e+02 3.654e+02 6.872e+02, threshold=6.204e+02, percent-clipped=2.0 +2023-03-09 11:42:22,127 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-09 11:42:55,610 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-03-09 11:43:08,127 INFO [train.py:898] (0/4) Epoch 19, batch 1600, loss[loss=0.1364, simple_loss=0.2198, pruned_loss=0.0265, over 18156.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2565, pruned_loss=0.03899, over 3574578.12 frames. ], batch size: 44, lr: 5.96e-03, grad_scale: 8.0 +2023-03-09 11:43:14,595 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67018.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 11:43:27,936 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5952, 2.9199, 2.6418, 2.9675, 3.6882, 3.6221, 3.1916, 3.0071], + device='cuda:0'), covar=tensor([0.0174, 0.0271, 0.0525, 0.0363, 0.0183, 0.0136, 0.0326, 0.0335], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0131, 0.0163, 0.0154, 0.0127, 0.0111, 0.0150, 0.0152], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 11:44:06,911 INFO [train.py:898] (0/4) Epoch 19, batch 1650, loss[loss=0.1552, simple_loss=0.2423, pruned_loss=0.03409, over 18261.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2561, pruned_loss=0.03887, over 3569787.82 frames. ], batch size: 45, lr: 5.96e-03, grad_scale: 8.0 +2023-03-09 11:44:16,358 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.888e+02 2.665e+02 3.285e+02 3.842e+02 7.636e+02, threshold=6.571e+02, percent-clipped=3.0 +2023-03-09 11:44:31,222 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67083.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:44:36,612 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67087.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 11:45:05,335 INFO [train.py:898] (0/4) Epoch 19, batch 1700, loss[loss=0.1464, simple_loss=0.2273, pruned_loss=0.0327, over 18435.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2562, pruned_loss=0.03915, over 3572851.63 frames. ], batch size: 43, lr: 5.95e-03, grad_scale: 8.0 +2023-03-09 11:45:43,269 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67144.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:46:04,393 INFO [train.py:898] (0/4) Epoch 19, batch 1750, loss[loss=0.1682, simple_loss=0.2596, pruned_loss=0.03839, over 18261.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2558, pruned_loss=0.03899, over 3580572.22 frames. ], batch size: 47, lr: 5.95e-03, grad_scale: 8.0 +2023-03-09 11:46:10,477 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6624, 2.2558, 2.6065, 2.6303, 3.1942, 4.7112, 4.4544, 3.4141], + device='cuda:0'), covar=tensor([0.1665, 0.2395, 0.2705, 0.1781, 0.2252, 0.0230, 0.0419, 0.0874], + device='cuda:0'), in_proj_covar=tensor([0.0293, 0.0340, 0.0373, 0.0273, 0.0388, 0.0234, 0.0293, 0.0247], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 11:46:13,338 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.723e+02 2.728e+02 3.267e+02 4.070e+02 1.061e+03, threshold=6.534e+02, percent-clipped=5.0 +2023-03-09 11:46:16,523 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1706, 2.6676, 2.3815, 2.7109, 3.3677, 3.2767, 2.9915, 2.7701], + device='cuda:0'), covar=tensor([0.0177, 0.0256, 0.0557, 0.0325, 0.0185, 0.0131, 0.0321, 0.0312], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0131, 0.0163, 0.0154, 0.0127, 0.0111, 0.0150, 0.0151], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 11:46:21,064 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67177.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 11:46:52,688 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6559, 2.3832, 2.7722, 2.8175, 3.3681, 4.9920, 4.8023, 3.5439], + device='cuda:0'), covar=tensor([0.1719, 0.2326, 0.2901, 0.1665, 0.2139, 0.0204, 0.0351, 0.0827], + device='cuda:0'), in_proj_covar=tensor([0.0293, 0.0340, 0.0372, 0.0273, 0.0387, 0.0234, 0.0292, 0.0247], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 11:47:02,903 INFO [train.py:898] (0/4) Epoch 19, batch 1800, loss[loss=0.1962, simple_loss=0.2819, pruned_loss=0.05525, over 18351.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2558, pruned_loss=0.03919, over 3583678.91 frames. ], batch size: 56, lr: 5.95e-03, grad_scale: 8.0 +2023-03-09 11:47:17,310 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=67225.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 11:48:01,508 INFO [train.py:898] (0/4) Epoch 19, batch 1850, loss[loss=0.1858, simple_loss=0.2869, pruned_loss=0.04237, over 17135.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2558, pruned_loss=0.03901, over 3591550.25 frames. ], batch size: 78, lr: 5.95e-03, grad_scale: 8.0 +2023-03-09 11:48:10,591 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.962e+02 2.832e+02 3.331e+02 4.104e+02 8.145e+02, threshold=6.662e+02, percent-clipped=3.0 +2023-03-09 11:48:20,681 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8975, 4.5926, 4.6470, 3.7347, 3.8148, 3.4623, 3.0634, 2.6919], + device='cuda:0'), covar=tensor([0.0226, 0.0146, 0.0077, 0.0242, 0.0358, 0.0230, 0.0597, 0.0805], + device='cuda:0'), in_proj_covar=tensor([0.0068, 0.0057, 0.0060, 0.0067, 0.0087, 0.0066, 0.0075, 0.0083], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 11:48:37,625 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6144, 2.5640, 2.6038, 2.4857, 2.4838, 2.2021, 2.3071, 2.5843], + device='cuda:0'), covar=tensor([0.0072, 0.0103, 0.0073, 0.0103, 0.0097, 0.0155, 0.0158, 0.0076], + device='cuda:0'), in_proj_covar=tensor([0.0130, 0.0155, 0.0129, 0.0182, 0.0136, 0.0176, 0.0178, 0.0115], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 11:49:00,052 INFO [train.py:898] (0/4) Epoch 19, batch 1900, loss[loss=0.1739, simple_loss=0.264, pruned_loss=0.04193, over 18401.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2576, pruned_loss=0.03959, over 3583832.98 frames. ], batch size: 52, lr: 5.95e-03, grad_scale: 4.0 +2023-03-09 11:49:00,331 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67313.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 11:49:20,479 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.78 vs. limit=2.0 +2023-03-09 11:49:26,863 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8672, 5.3277, 5.2973, 5.2849, 4.8393, 5.2257, 4.6492, 5.1853], + device='cuda:0'), covar=tensor([0.0235, 0.0292, 0.0217, 0.0483, 0.0408, 0.0215, 0.1089, 0.0329], + device='cuda:0'), in_proj_covar=tensor([0.0210, 0.0257, 0.0247, 0.0320, 0.0263, 0.0263, 0.0304, 0.0251], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 11:49:58,272 INFO [train.py:898] (0/4) Epoch 19, batch 1950, loss[loss=0.1448, simple_loss=0.2392, pruned_loss=0.02524, over 18387.00 frames. ], tot_loss[loss=0.1687, simple_loss=0.2576, pruned_loss=0.03992, over 3577472.80 frames. ], batch size: 50, lr: 5.94e-03, grad_scale: 4.0 +2023-03-09 11:50:08,467 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.842e+02 2.816e+02 3.321e+02 4.112e+02 1.785e+03, threshold=6.643e+02, percent-clipped=3.0 +2023-03-09 11:50:22,572 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-09 11:50:26,320 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67387.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 11:50:31,405 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3212, 5.1608, 5.5477, 5.5909, 5.1755, 6.1336, 5.7554, 5.4179], + device='cuda:0'), covar=tensor([0.1006, 0.0649, 0.0702, 0.0731, 0.1488, 0.0699, 0.0562, 0.1550], + device='cuda:0'), in_proj_covar=tensor([0.0352, 0.0277, 0.0299, 0.0298, 0.0325, 0.0413, 0.0273, 0.0403], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 11:50:33,876 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5903, 5.5402, 5.1808, 5.5067, 5.4421, 4.9033, 5.3935, 5.1119], + device='cuda:0'), covar=tensor([0.0385, 0.0409, 0.1261, 0.0670, 0.0598, 0.0377, 0.0388, 0.1004], + device='cuda:0'), in_proj_covar=tensor([0.0468, 0.0540, 0.0681, 0.0420, 0.0436, 0.0487, 0.0527, 0.0663], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 11:50:48,582 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2152, 5.7380, 5.3483, 5.4910, 5.3287, 5.1670, 5.7882, 5.7329], + device='cuda:0'), covar=tensor([0.1294, 0.0751, 0.0602, 0.0815, 0.1484, 0.0734, 0.0601, 0.0711], + device='cuda:0'), in_proj_covar=tensor([0.0610, 0.0522, 0.0376, 0.0548, 0.0735, 0.0541, 0.0736, 0.0558], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 11:50:57,633 INFO [train.py:898] (0/4) Epoch 19, batch 2000, loss[loss=0.1753, simple_loss=0.2701, pruned_loss=0.04019, over 17001.00 frames. ], tot_loss[loss=0.1685, simple_loss=0.2574, pruned_loss=0.03978, over 3578338.44 frames. ], batch size: 78, lr: 5.94e-03, grad_scale: 8.0 +2023-03-09 11:51:22,887 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=67435.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 11:51:27,976 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67439.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:51:56,182 INFO [train.py:898] (0/4) Epoch 19, batch 2050, loss[loss=0.1475, simple_loss=0.2288, pruned_loss=0.03309, over 18440.00 frames. ], tot_loss[loss=0.168, simple_loss=0.257, pruned_loss=0.03945, over 3595923.21 frames. ], batch size: 43, lr: 5.94e-03, grad_scale: 8.0 +2023-03-09 11:52:06,256 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.056e+02 2.683e+02 3.184e+02 3.899e+02 7.354e+02, threshold=6.369e+02, percent-clipped=1.0 +2023-03-09 11:52:24,572 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.49 vs. limit=5.0 +2023-03-09 11:52:54,224 INFO [train.py:898] (0/4) Epoch 19, batch 2100, loss[loss=0.1507, simple_loss=0.2464, pruned_loss=0.02752, over 18385.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2567, pruned_loss=0.0394, over 3600819.07 frames. ], batch size: 50, lr: 5.94e-03, grad_scale: 8.0 +2023-03-09 11:52:54,520 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67513.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:53:22,517 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3092, 5.1741, 5.5137, 5.5102, 5.2174, 6.0270, 5.6883, 5.2885], + device='cuda:0'), covar=tensor([0.1002, 0.0665, 0.0679, 0.0653, 0.1674, 0.0760, 0.0607, 0.1667], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0277, 0.0300, 0.0297, 0.0327, 0.0414, 0.0272, 0.0404], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 11:53:52,384 INFO [train.py:898] (0/4) Epoch 19, batch 2150, loss[loss=0.1469, simple_loss=0.2278, pruned_loss=0.03296, over 18269.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2563, pruned_loss=0.0394, over 3593296.57 frames. ], batch size: 45, lr: 5.93e-03, grad_scale: 8.0 +2023-03-09 11:54:03,260 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.859e+02 2.589e+02 3.256e+02 4.119e+02 1.040e+03, threshold=6.512e+02, percent-clipped=4.0 +2023-03-09 11:54:05,925 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67574.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 11:54:27,976 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9978, 3.8384, 5.2928, 3.1468, 4.5244, 2.7747, 3.1673, 1.9990], + device='cuda:0'), covar=tensor([0.1013, 0.0835, 0.0103, 0.0762, 0.0484, 0.2378, 0.2523, 0.1899], + device='cuda:0'), in_proj_covar=tensor([0.0215, 0.0238, 0.0175, 0.0191, 0.0252, 0.0266, 0.0315, 0.0228], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 11:54:43,750 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.97 vs. limit=5.0 +2023-03-09 11:54:51,797 INFO [train.py:898] (0/4) Epoch 19, batch 2200, loss[loss=0.1616, simple_loss=0.2518, pruned_loss=0.03574, over 18398.00 frames. ], tot_loss[loss=0.1672, simple_loss=0.2558, pruned_loss=0.0393, over 3582367.15 frames. ], batch size: 50, lr: 5.93e-03, grad_scale: 8.0 +2023-03-09 11:54:52,037 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67613.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:55:48,947 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=67661.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:55:50,988 INFO [train.py:898] (0/4) Epoch 19, batch 2250, loss[loss=0.1654, simple_loss=0.2597, pruned_loss=0.03553, over 18474.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2553, pruned_loss=0.03907, over 3579628.00 frames. ], batch size: 51, lr: 5.93e-03, grad_scale: 8.0 +2023-03-09 11:56:01,704 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.067e+02 2.672e+02 3.118e+02 3.558e+02 7.247e+02, threshold=6.237e+02, percent-clipped=1.0 +2023-03-09 11:56:02,225 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6149, 2.2897, 2.7225, 2.8159, 3.2777, 4.9542, 4.7680, 3.4652], + device='cuda:0'), covar=tensor([0.1755, 0.2438, 0.2942, 0.1685, 0.2318, 0.0204, 0.0353, 0.0894], + device='cuda:0'), in_proj_covar=tensor([0.0293, 0.0340, 0.0371, 0.0273, 0.0387, 0.0234, 0.0292, 0.0249], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 11:56:11,212 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67680.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:56:24,885 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67692.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:56:50,294 INFO [train.py:898] (0/4) Epoch 19, batch 2300, loss[loss=0.1749, simple_loss=0.2718, pruned_loss=0.03901, over 16273.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2544, pruned_loss=0.03907, over 3569173.77 frames. ], batch size: 95, lr: 5.93e-03, grad_scale: 8.0 +2023-03-09 11:57:11,936 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8414, 3.8475, 3.6819, 3.4009, 3.5723, 2.9447, 3.0803, 3.8976], + device='cuda:0'), covar=tensor([0.0059, 0.0081, 0.0077, 0.0113, 0.0078, 0.0167, 0.0170, 0.0049], + device='cuda:0'), in_proj_covar=tensor([0.0129, 0.0152, 0.0128, 0.0181, 0.0134, 0.0172, 0.0175, 0.0113], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 11:57:15,118 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8059, 5.2775, 5.2694, 5.3206, 4.8280, 5.1817, 4.5576, 5.1669], + device='cuda:0'), covar=tensor([0.0241, 0.0320, 0.0205, 0.0364, 0.0364, 0.0233, 0.1075, 0.0300], + device='cuda:0'), in_proj_covar=tensor([0.0212, 0.0260, 0.0248, 0.0321, 0.0265, 0.0264, 0.0303, 0.0253], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 11:57:20,946 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=67739.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:57:23,437 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67741.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:57:37,138 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=67753.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 11:57:48,523 INFO [train.py:898] (0/4) Epoch 19, batch 2350, loss[loss=0.1653, simple_loss=0.2502, pruned_loss=0.0402, over 18349.00 frames. ], tot_loss[loss=0.1665, simple_loss=0.2548, pruned_loss=0.0391, over 3577978.53 frames. ], batch size: 46, lr: 5.93e-03, grad_scale: 8.0 +2023-03-09 11:57:57,285 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8450, 5.3661, 5.3648, 5.4009, 4.8771, 5.2548, 4.7299, 5.2719], + device='cuda:0'), covar=tensor([0.0258, 0.0284, 0.0181, 0.0364, 0.0372, 0.0238, 0.0976, 0.0288], + device='cuda:0'), in_proj_covar=tensor([0.0212, 0.0259, 0.0247, 0.0319, 0.0264, 0.0264, 0.0303, 0.0251], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 11:57:59,066 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.915e+02 2.613e+02 3.195e+02 3.840e+02 8.434e+02, threshold=6.389e+02, percent-clipped=1.0 +2023-03-09 11:58:16,648 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=67787.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 11:58:40,336 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0563, 4.0010, 5.3204, 2.9385, 4.6390, 2.7849, 3.2081, 2.0100], + device='cuda:0'), covar=tensor([0.0994, 0.0811, 0.0114, 0.0896, 0.0473, 0.2432, 0.2771, 0.2039], + device='cuda:0'), in_proj_covar=tensor([0.0218, 0.0240, 0.0178, 0.0193, 0.0255, 0.0268, 0.0319, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 11:58:47,097 INFO [train.py:898] (0/4) Epoch 19, batch 2400, loss[loss=0.1582, simple_loss=0.2377, pruned_loss=0.03935, over 18397.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2547, pruned_loss=0.03926, over 3573471.56 frames. ], batch size: 42, lr: 5.92e-03, grad_scale: 8.0 +2023-03-09 11:58:57,314 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7357, 2.4613, 2.7684, 2.9856, 3.3799, 5.0780, 5.0354, 3.5624], + device='cuda:0'), covar=tensor([0.1735, 0.2366, 0.2804, 0.1610, 0.2187, 0.0186, 0.0280, 0.0865], + device='cuda:0'), in_proj_covar=tensor([0.0296, 0.0343, 0.0375, 0.0275, 0.0391, 0.0237, 0.0294, 0.0251], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 11:59:03,095 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.42 vs. limit=2.0 +2023-03-09 11:59:45,735 INFO [train.py:898] (0/4) Epoch 19, batch 2450, loss[loss=0.1499, simple_loss=0.2375, pruned_loss=0.03111, over 18336.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2544, pruned_loss=0.03906, over 3569231.07 frames. ], batch size: 46, lr: 5.92e-03, grad_scale: 8.0 +2023-03-09 11:59:53,341 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=67869.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 11:59:57,182 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.025e+02 2.777e+02 3.368e+02 4.107e+02 6.941e+02, threshold=6.736e+02, percent-clipped=1.0 +2023-03-09 11:59:57,636 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2388, 2.7436, 2.4311, 2.7037, 3.3361, 3.1787, 2.9394, 2.6563], + device='cuda:0'), covar=tensor([0.0218, 0.0284, 0.0580, 0.0385, 0.0212, 0.0202, 0.0365, 0.0397], + device='cuda:0'), in_proj_covar=tensor([0.0137, 0.0131, 0.0163, 0.0155, 0.0128, 0.0113, 0.0150, 0.0153], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 12:00:44,887 INFO [train.py:898] (0/4) Epoch 19, batch 2500, loss[loss=0.1525, simple_loss=0.2347, pruned_loss=0.03518, over 18232.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2541, pruned_loss=0.03878, over 3571022.91 frames. ], batch size: 45, lr: 5.92e-03, grad_scale: 8.0 +2023-03-09 12:01:38,398 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2692, 2.7531, 2.3812, 2.7165, 3.4139, 3.2700, 2.9307, 2.6565], + device='cuda:0'), covar=tensor([0.0196, 0.0260, 0.0601, 0.0379, 0.0199, 0.0169, 0.0352, 0.0406], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0129, 0.0162, 0.0153, 0.0127, 0.0112, 0.0149, 0.0152], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 12:01:43,837 INFO [train.py:898] (0/4) Epoch 19, batch 2550, loss[loss=0.1522, simple_loss=0.2402, pruned_loss=0.03205, over 18255.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2541, pruned_loss=0.03863, over 3586873.92 frames. ], batch size: 47, lr: 5.92e-03, grad_scale: 4.0 +2023-03-09 12:01:46,378 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7445, 4.1657, 5.3600, 4.6260, 3.1365, 2.9957, 4.5529, 5.5816], + device='cuda:0'), covar=tensor([0.0774, 0.1214, 0.0139, 0.0340, 0.0956, 0.1096, 0.0376, 0.0123], + device='cuda:0'), in_proj_covar=tensor([0.0146, 0.0270, 0.0144, 0.0179, 0.0190, 0.0186, 0.0190, 0.0188], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 12:01:56,271 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.852e+02 2.690e+02 3.137e+02 3.674e+02 8.020e+02, threshold=6.273e+02, percent-clipped=2.0 +2023-03-09 12:02:05,962 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=67981.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:02:28,144 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-68000.pt +2023-03-09 12:02:48,100 INFO [train.py:898] (0/4) Epoch 19, batch 2600, loss[loss=0.2008, simple_loss=0.2813, pruned_loss=0.06015, over 18295.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2551, pruned_loss=0.03929, over 3573156.54 frames. ], batch size: 57, lr: 5.91e-03, grad_scale: 4.0 +2023-03-09 12:02:55,323 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.15 vs. limit=5.0 +2023-03-09 12:03:13,554 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2896, 4.3512, 2.5475, 4.1246, 5.4430, 2.9918, 4.1885, 4.1655], + device='cuda:0'), covar=tensor([0.0112, 0.0977, 0.1486, 0.0584, 0.0063, 0.0975, 0.0512, 0.0626], + device='cuda:0'), in_proj_covar=tensor([0.0157, 0.0263, 0.0201, 0.0193, 0.0120, 0.0180, 0.0214, 0.0224], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 12:03:16,165 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68036.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:03:23,038 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68042.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:03:29,709 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68048.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 12:03:46,991 INFO [train.py:898] (0/4) Epoch 19, batch 2650, loss[loss=0.2054, simple_loss=0.2871, pruned_loss=0.06184, over 12697.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2556, pruned_loss=0.03913, over 3575478.07 frames. ], batch size: 129, lr: 5.91e-03, grad_scale: 4.0 +2023-03-09 12:03:56,672 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.22 vs. limit=5.0 +2023-03-09 12:03:58,869 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.938e+02 2.796e+02 3.306e+02 3.948e+02 6.802e+02, threshold=6.612e+02, percent-clipped=2.0 +2023-03-09 12:04:16,012 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68087.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 12:04:23,634 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1476, 5.1304, 5.4005, 5.4495, 5.0903, 5.9688, 5.5940, 5.2433], + device='cuda:0'), covar=tensor([0.1137, 0.0657, 0.0743, 0.0936, 0.1504, 0.0842, 0.0662, 0.1798], + device='cuda:0'), in_proj_covar=tensor([0.0349, 0.0275, 0.0301, 0.0299, 0.0325, 0.0412, 0.0272, 0.0402], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 12:04:27,319 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7725, 3.9800, 2.4321, 3.7918, 4.9585, 2.4943, 3.7540, 3.8441], + device='cuda:0'), covar=tensor([0.0160, 0.1124, 0.1555, 0.0666, 0.0080, 0.1253, 0.0657, 0.0683], + device='cuda:0'), in_proj_covar=tensor([0.0159, 0.0266, 0.0202, 0.0195, 0.0121, 0.0181, 0.0215, 0.0226], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 12:04:28,676 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-09 12:04:44,746 INFO [train.py:898] (0/4) Epoch 19, batch 2700, loss[loss=0.1708, simple_loss=0.2586, pruned_loss=0.04157, over 18360.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.2561, pruned_loss=0.03935, over 3572138.45 frames. ], batch size: 56, lr: 5.91e-03, grad_scale: 4.0 +2023-03-09 12:05:18,976 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-09 12:05:27,142 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68148.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 12:05:43,845 INFO [train.py:898] (0/4) Epoch 19, batch 2750, loss[loss=0.176, simple_loss=0.2542, pruned_loss=0.04889, over 17630.00 frames. ], tot_loss[loss=0.1667, simple_loss=0.2557, pruned_loss=0.03887, over 3580087.19 frames. ], batch size: 39, lr: 5.91e-03, grad_scale: 4.0 +2023-03-09 12:05:51,510 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68169.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:05:55,852 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.532e+02 2.565e+02 3.131e+02 3.754e+02 7.044e+02, threshold=6.262e+02, percent-clipped=1.0 +2023-03-09 12:06:30,609 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2541, 5.7631, 5.3582, 5.5576, 5.3035, 5.2085, 5.8165, 5.7235], + device='cuda:0'), covar=tensor([0.1247, 0.0787, 0.0644, 0.0761, 0.1606, 0.0719, 0.0570, 0.0777], + device='cuda:0'), in_proj_covar=tensor([0.0603, 0.0514, 0.0375, 0.0540, 0.0733, 0.0535, 0.0730, 0.0558], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 12:06:42,968 INFO [train.py:898] (0/4) Epoch 19, batch 2800, loss[loss=0.1579, simple_loss=0.2355, pruned_loss=0.0402, over 18495.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2551, pruned_loss=0.03908, over 3569145.40 frames. ], batch size: 44, lr: 5.91e-03, grad_scale: 8.0 +2023-03-09 12:06:47,640 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=68217.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:06:52,306 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-09 12:07:26,101 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6787, 3.5432, 2.4124, 4.4273, 3.1652, 4.2980, 2.5711, 4.0430], + device='cuda:0'), covar=tensor([0.0575, 0.0735, 0.1322, 0.0447, 0.0783, 0.0328, 0.1103, 0.0351], + device='cuda:0'), in_proj_covar=tensor([0.0210, 0.0226, 0.0189, 0.0280, 0.0193, 0.0262, 0.0202, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 12:07:41,657 INFO [train.py:898] (0/4) Epoch 19, batch 2850, loss[loss=0.1432, simple_loss=0.2241, pruned_loss=0.03117, over 18506.00 frames. ], tot_loss[loss=0.1674, simple_loss=0.256, pruned_loss=0.03944, over 3562582.05 frames. ], batch size: 44, lr: 5.90e-03, grad_scale: 8.0 +2023-03-09 12:07:53,622 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.861e+02 2.563e+02 3.164e+02 3.682e+02 9.840e+02, threshold=6.328e+02, percent-clipped=2.0 +2023-03-09 12:08:15,427 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-09 12:08:41,528 INFO [train.py:898] (0/4) Epoch 19, batch 2900, loss[loss=0.1715, simple_loss=0.2638, pruned_loss=0.03964, over 17036.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2562, pruned_loss=0.03917, over 3567237.18 frames. ], batch size: 78, lr: 5.90e-03, grad_scale: 8.0 +2023-03-09 12:09:08,673 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68336.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:09:10,283 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68337.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:09:23,318 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68348.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:09:40,262 INFO [train.py:898] (0/4) Epoch 19, batch 2950, loss[loss=0.202, simple_loss=0.2835, pruned_loss=0.06027, over 18464.00 frames. ], tot_loss[loss=0.1671, simple_loss=0.2564, pruned_loss=0.03888, over 3584248.67 frames. ], batch size: 59, lr: 5.90e-03, grad_scale: 8.0 +2023-03-09 12:09:51,484 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.776e+02 2.615e+02 3.161e+02 3.773e+02 8.205e+02, threshold=6.323e+02, percent-clipped=2.0 +2023-03-09 12:10:04,933 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=68384.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:10:19,056 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=68396.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:10:23,122 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68399.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:10:39,459 INFO [train.py:898] (0/4) Epoch 19, batch 3000, loss[loss=0.2092, simple_loss=0.2953, pruned_loss=0.0615, over 18293.00 frames. ], tot_loss[loss=0.1673, simple_loss=0.2566, pruned_loss=0.03901, over 3578098.40 frames. ], batch size: 57, lr: 5.90e-03, grad_scale: 8.0 +2023-03-09 12:10:39,461 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 12:10:51,589 INFO [train.py:932] (0/4) Epoch 19, validation: loss=0.1511, simple_loss=0.2509, pruned_loss=0.02564, over 944034.00 frames. +2023-03-09 12:10:51,590 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 12:11:27,560 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68443.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 12:11:47,176 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68460.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:11:50,383 INFO [train.py:898] (0/4) Epoch 19, batch 3050, loss[loss=0.1749, simple_loss=0.2645, pruned_loss=0.04264, over 17826.00 frames. ], tot_loss[loss=0.1676, simple_loss=0.2568, pruned_loss=0.03922, over 3585056.90 frames. ], batch size: 70, lr: 5.90e-03, grad_scale: 8.0 +2023-03-09 12:12:02,237 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.030e+02 2.743e+02 3.169e+02 3.811e+02 7.810e+02, threshold=6.338e+02, percent-clipped=1.0 +2023-03-09 12:12:49,514 INFO [train.py:898] (0/4) Epoch 19, batch 3100, loss[loss=0.1997, simple_loss=0.2776, pruned_loss=0.06085, over 18360.00 frames. ], tot_loss[loss=0.1677, simple_loss=0.2567, pruned_loss=0.03936, over 3564603.40 frames. ], batch size: 50, lr: 5.89e-03, grad_scale: 8.0 +2023-03-09 12:12:57,025 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5439, 6.0959, 5.5939, 5.8904, 5.6454, 5.4914, 6.1471, 6.0882], + device='cuda:0'), covar=tensor([0.1215, 0.0719, 0.0463, 0.0745, 0.1504, 0.0699, 0.0552, 0.0683], + device='cuda:0'), in_proj_covar=tensor([0.0604, 0.0516, 0.0377, 0.0542, 0.0735, 0.0536, 0.0729, 0.0555], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 12:13:45,171 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6146, 3.5201, 3.4333, 2.9881, 3.2512, 2.6566, 2.6043, 3.5203], + device='cuda:0'), covar=tensor([0.0063, 0.0102, 0.0076, 0.0138, 0.0108, 0.0199, 0.0223, 0.0064], + device='cuda:0'), in_proj_covar=tensor([0.0131, 0.0153, 0.0127, 0.0181, 0.0136, 0.0173, 0.0176, 0.0113], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 12:13:48,204 INFO [train.py:898] (0/4) Epoch 19, batch 3150, loss[loss=0.1786, simple_loss=0.2643, pruned_loss=0.0465, over 18324.00 frames. ], tot_loss[loss=0.1684, simple_loss=0.2572, pruned_loss=0.03978, over 3555651.88 frames. ], batch size: 57, lr: 5.89e-03, grad_scale: 8.0 +2023-03-09 12:13:59,947 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.941e+02 2.741e+02 3.192e+02 3.844e+02 7.803e+02, threshold=6.385e+02, percent-clipped=3.0 +2023-03-09 12:14:05,845 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68578.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:14:14,124 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-09 12:14:46,957 INFO [train.py:898] (0/4) Epoch 19, batch 3200, loss[loss=0.1625, simple_loss=0.2357, pruned_loss=0.04464, over 18056.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2553, pruned_loss=0.03926, over 3561771.82 frames. ], batch size: 40, lr: 5.89e-03, grad_scale: 8.0 +2023-03-09 12:15:15,755 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68637.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:15:18,760 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68639.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:15:46,038 INFO [train.py:898] (0/4) Epoch 19, batch 3250, loss[loss=0.1577, simple_loss=0.2507, pruned_loss=0.03235, over 18253.00 frames. ], tot_loss[loss=0.1669, simple_loss=0.2558, pruned_loss=0.03898, over 3582909.41 frames. ], batch size: 60, lr: 5.89e-03, grad_scale: 8.0 +2023-03-09 12:15:57,333 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.558e+02 2.445e+02 3.074e+02 3.780e+02 1.190e+03, threshold=6.148e+02, percent-clipped=4.0 +2023-03-09 12:16:11,501 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=68685.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:16:11,647 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=68685.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:16:45,037 INFO [train.py:898] (0/4) Epoch 19, batch 3300, loss[loss=0.1661, simple_loss=0.2578, pruned_loss=0.03717, over 17236.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2552, pruned_loss=0.03879, over 3579585.41 frames. ], batch size: 78, lr: 5.88e-03, grad_scale: 8.0 +2023-03-09 12:17:20,695 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=68743.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 12:17:24,166 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=68746.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:17:34,031 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68755.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:17:43,468 INFO [train.py:898] (0/4) Epoch 19, batch 3350, loss[loss=0.1532, simple_loss=0.2311, pruned_loss=0.03763, over 17717.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.255, pruned_loss=0.03876, over 3576245.73 frames. ], batch size: 39, lr: 5.88e-03, grad_scale: 8.0 +2023-03-09 12:17:54,602 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.843e+02 2.741e+02 3.273e+02 4.091e+02 6.319e+02, threshold=6.545e+02, percent-clipped=1.0 +2023-03-09 12:18:16,569 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=68791.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 12:18:42,931 INFO [train.py:898] (0/4) Epoch 19, batch 3400, loss[loss=0.155, simple_loss=0.2372, pruned_loss=0.03642, over 18356.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2551, pruned_loss=0.03878, over 3579639.10 frames. ], batch size: 46, lr: 5.88e-03, grad_scale: 8.0 +2023-03-09 12:19:36,322 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5663, 5.2182, 5.7447, 5.7872, 5.5044, 6.3006, 6.0129, 5.5506], + device='cuda:0'), covar=tensor([0.1029, 0.0571, 0.0732, 0.0764, 0.1423, 0.0599, 0.0589, 0.1457], + device='cuda:0'), in_proj_covar=tensor([0.0352, 0.0280, 0.0304, 0.0303, 0.0328, 0.0416, 0.0277, 0.0406], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 12:19:41,742 INFO [train.py:898] (0/4) Epoch 19, batch 3450, loss[loss=0.1503, simple_loss=0.2377, pruned_loss=0.03149, over 18410.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2543, pruned_loss=0.03843, over 3593418.35 frames. ], batch size: 48, lr: 5.88e-03, grad_scale: 8.0 +2023-03-09 12:19:44,612 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4379, 3.2784, 2.0977, 4.1603, 2.8174, 3.9584, 2.1993, 3.6431], + device='cuda:0'), covar=tensor([0.0578, 0.0795, 0.1355, 0.0472, 0.0873, 0.0341, 0.1250, 0.0426], + device='cuda:0'), in_proj_covar=tensor([0.0210, 0.0223, 0.0186, 0.0279, 0.0191, 0.0260, 0.0201, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 12:19:53,563 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.993e+02 2.660e+02 3.004e+02 3.762e+02 7.210e+02, threshold=6.009e+02, percent-clipped=0.0 +2023-03-09 12:20:09,621 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=2.06 vs. limit=2.0 +2023-03-09 12:20:39,574 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-09 12:20:40,016 INFO [train.py:898] (0/4) Epoch 19, batch 3500, loss[loss=0.187, simple_loss=0.2721, pruned_loss=0.05101, over 18315.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2547, pruned_loss=0.03843, over 3590145.00 frames. ], batch size: 57, lr: 5.88e-03, grad_scale: 8.0 +2023-03-09 12:21:04,357 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=68934.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:21:35,920 INFO [train.py:898] (0/4) Epoch 19, batch 3550, loss[loss=0.169, simple_loss=0.2625, pruned_loss=0.03778, over 16330.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2552, pruned_loss=0.03845, over 3584294.04 frames. ], batch size: 95, lr: 5.87e-03, grad_scale: 8.0 +2023-03-09 12:21:46,788 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.911e+02 2.828e+02 3.238e+02 3.903e+02 1.381e+03, threshold=6.476e+02, percent-clipped=5.0 +2023-03-09 12:21:56,396 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5522, 2.7591, 2.5479, 2.9113, 3.6474, 3.5568, 3.1222, 2.8838], + device='cuda:0'), covar=tensor([0.0172, 0.0295, 0.0569, 0.0347, 0.0174, 0.0144, 0.0362, 0.0388], + device='cuda:0'), in_proj_covar=tensor([0.0138, 0.0130, 0.0161, 0.0153, 0.0127, 0.0113, 0.0150, 0.0152], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 12:22:15,060 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5484, 5.0447, 5.0005, 5.0323, 4.5784, 4.9267, 4.4233, 4.8795], + device='cuda:0'), covar=tensor([0.0259, 0.0268, 0.0221, 0.0424, 0.0396, 0.0219, 0.0995, 0.0342], + device='cuda:0'), in_proj_covar=tensor([0.0210, 0.0256, 0.0245, 0.0318, 0.0263, 0.0262, 0.0300, 0.0252], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 12:22:30,455 INFO [train.py:898] (0/4) Epoch 19, batch 3600, loss[loss=0.1514, simple_loss=0.231, pruned_loss=0.03588, over 18565.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2556, pruned_loss=0.03863, over 3588800.03 frames. ], batch size: 45, lr: 5.87e-03, grad_scale: 8.0 +2023-03-09 12:22:42,375 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8170, 3.8735, 5.1040, 2.9195, 4.4019, 2.6050, 3.0537, 1.7282], + device='cuda:0'), covar=tensor([0.1198, 0.0831, 0.0131, 0.0893, 0.0506, 0.2551, 0.2757, 0.2262], + device='cuda:0'), in_proj_covar=tensor([0.0215, 0.0237, 0.0176, 0.0190, 0.0251, 0.0264, 0.0315, 0.0227], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 12:22:59,531 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69041.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:23:01,621 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69043.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:23:05,340 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-19.pt +2023-03-09 12:23:34,667 INFO [train.py:898] (0/4) Epoch 20, batch 0, loss[loss=0.1655, simple_loss=0.2506, pruned_loss=0.04024, over 18270.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2506, pruned_loss=0.04024, over 18270.00 frames. ], batch size: 47, lr: 5.72e-03, grad_scale: 8.0 +2023-03-09 12:23:34,669 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 12:23:39,398 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7782, 3.4944, 4.7925, 4.3109, 3.3937, 2.9944, 4.2848, 4.9647], + device='cuda:0'), covar=tensor([0.0773, 0.1488, 0.0173, 0.0361, 0.0901, 0.1200, 0.0374, 0.0236], + device='cuda:0'), in_proj_covar=tensor([0.0145, 0.0268, 0.0145, 0.0178, 0.0189, 0.0185, 0.0189, 0.0189], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 12:23:46,326 INFO [train.py:932] (0/4) Epoch 20, validation: loss=0.1509, simple_loss=0.2512, pruned_loss=0.02534, over 944034.00 frames. +2023-03-09 12:23:46,327 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 12:23:47,733 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4298, 5.4041, 5.0475, 5.2944, 5.3567, 4.8054, 5.2688, 5.0263], + device='cuda:0'), covar=tensor([0.0398, 0.0457, 0.1253, 0.0864, 0.0586, 0.0381, 0.0419, 0.1103], + device='cuda:0'), in_proj_covar=tensor([0.0466, 0.0534, 0.0677, 0.0418, 0.0435, 0.0485, 0.0529, 0.0657], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 12:23:56,053 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69055.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:24:17,097 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.770e+02 2.674e+02 3.193e+02 4.211e+02 7.931e+02, threshold=6.386e+02, percent-clipped=3.0 +2023-03-09 12:24:18,664 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7493, 4.0739, 2.4691, 3.8904, 4.9955, 2.4442, 3.6028, 3.8177], + device='cuda:0'), covar=tensor([0.0190, 0.1097, 0.1714, 0.0660, 0.0092, 0.1341, 0.0791, 0.0763], + device='cuda:0'), in_proj_covar=tensor([0.0161, 0.0266, 0.0202, 0.0194, 0.0121, 0.0181, 0.0216, 0.0225], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 12:24:40,573 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-09 12:24:45,724 INFO [train.py:898] (0/4) Epoch 20, batch 50, loss[loss=0.1649, simple_loss=0.2596, pruned_loss=0.03507, over 18372.00 frames. ], tot_loss[loss=0.1678, simple_loss=0.2578, pruned_loss=0.03885, over 811306.34 frames. ], batch size: 50, lr: 5.72e-03, grad_scale: 4.0 +2023-03-09 12:24:48,911 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-09 12:24:50,051 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.64 vs. limit=2.0 +2023-03-09 12:24:52,713 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=69103.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:24:54,016 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69104.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:25:37,760 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.24 vs. limit=5.0 +2023-03-09 12:25:44,569 INFO [train.py:898] (0/4) Epoch 20, batch 100, loss[loss=0.1571, simple_loss=0.2557, pruned_loss=0.02931, over 18402.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2556, pruned_loss=0.03782, over 1434065.64 frames. ], batch size: 52, lr: 5.72e-03, grad_scale: 4.0 +2023-03-09 12:26:01,892 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69162.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:26:06,966 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9378, 3.3259, 4.3944, 3.9407, 2.9710, 4.8688, 4.1587, 3.2153], + device='cuda:0'), covar=tensor([0.0462, 0.1232, 0.0279, 0.0383, 0.1434, 0.0174, 0.0458, 0.0901], + device='cuda:0'), in_proj_covar=tensor([0.0209, 0.0238, 0.0202, 0.0157, 0.0222, 0.0207, 0.0241, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 12:26:09,318 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69168.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:26:15,826 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.993e+02 2.622e+02 3.145e+02 3.671e+02 8.108e+02, threshold=6.290e+02, percent-clipped=1.0 +2023-03-09 12:26:42,960 INFO [train.py:898] (0/4) Epoch 20, batch 150, loss[loss=0.1627, simple_loss=0.2553, pruned_loss=0.03509, over 18622.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2542, pruned_loss=0.0376, over 1925397.16 frames. ], batch size: 52, lr: 5.71e-03, grad_scale: 4.0 +2023-03-09 12:27:14,040 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69223.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:27:16,491 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69225.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:27:21,057 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69229.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:27:26,753 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69234.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:27:34,876 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3488, 2.7372, 3.7061, 3.3686, 2.4410, 3.9603, 3.6105, 2.6771], + device='cuda:0'), covar=tensor([0.0515, 0.1315, 0.0325, 0.0406, 0.1495, 0.0233, 0.0536, 0.0957], + device='cuda:0'), in_proj_covar=tensor([0.0209, 0.0237, 0.0200, 0.0157, 0.0221, 0.0206, 0.0240, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 12:27:40,453 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69245.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 12:27:42,405 INFO [train.py:898] (0/4) Epoch 20, batch 200, loss[loss=0.1769, simple_loss=0.2731, pruned_loss=0.04032, over 18492.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.255, pruned_loss=0.03815, over 2306713.59 frames. ], batch size: 53, lr: 5.71e-03, grad_scale: 4.0 +2023-03-09 12:27:46,224 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7950, 3.8066, 5.1378, 4.6064, 3.4691, 3.0622, 4.6673, 5.3743], + device='cuda:0'), covar=tensor([0.0763, 0.1484, 0.0176, 0.0325, 0.0886, 0.1112, 0.0320, 0.0219], + device='cuda:0'), in_proj_covar=tensor([0.0146, 0.0269, 0.0145, 0.0179, 0.0190, 0.0186, 0.0190, 0.0190], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 12:28:07,079 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9950, 5.0935, 5.1361, 4.7871, 4.9003, 4.8766, 5.2333, 5.2061], + device='cuda:0'), covar=tensor([0.0067, 0.0056, 0.0051, 0.0101, 0.0055, 0.0144, 0.0061, 0.0076], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0068, 0.0072, 0.0091, 0.0074, 0.0102, 0.0085, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 12:28:13,579 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69273.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:28:14,466 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.067e+02 2.703e+02 3.198e+02 3.752e+02 7.537e+02, threshold=6.397e+02, percent-clipped=4.0 +2023-03-09 12:28:23,438 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=69282.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:28:28,143 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69286.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:28:41,463 INFO [train.py:898] (0/4) Epoch 20, batch 250, loss[loss=0.1979, simple_loss=0.2774, pruned_loss=0.05919, over 12957.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2549, pruned_loss=0.03821, over 2596027.17 frames. ], batch size: 129, lr: 5.71e-03, grad_scale: 4.0 +2023-03-09 12:28:52,672 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69306.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 12:29:01,488 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4447, 5.9658, 5.5166, 5.7831, 5.5685, 5.3981, 6.0568, 5.9691], + device='cuda:0'), covar=tensor([0.1161, 0.0844, 0.0531, 0.0759, 0.1476, 0.0765, 0.0574, 0.0752], + device='cuda:0'), in_proj_covar=tensor([0.0605, 0.0517, 0.0380, 0.0543, 0.0731, 0.0536, 0.0733, 0.0558], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 12:29:08,286 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69320.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:29:24,410 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69334.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 12:29:32,122 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69341.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:29:39,805 INFO [train.py:898] (0/4) Epoch 20, batch 300, loss[loss=0.1659, simple_loss=0.2604, pruned_loss=0.03573, over 18125.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2544, pruned_loss=0.03802, over 2832501.13 frames. ], batch size: 62, lr: 5.71e-03, grad_scale: 4.0 +2023-03-09 12:30:04,274 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69368.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:30:10,764 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.810e+02 2.793e+02 3.155e+02 4.197e+02 1.182e+03, threshold=6.310e+02, percent-clipped=4.0 +2023-03-09 12:30:13,420 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9421, 5.1749, 2.6576, 4.9443, 4.8232, 5.1768, 4.9453, 2.5260], + device='cuda:0'), covar=tensor([0.0207, 0.0071, 0.0823, 0.0096, 0.0092, 0.0072, 0.0093, 0.1149], + device='cuda:0'), in_proj_covar=tensor([0.0087, 0.0081, 0.0096, 0.0096, 0.0085, 0.0075, 0.0084, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 12:30:19,685 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69381.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:30:28,846 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=69389.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:30:34,527 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69394.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:30:37,561 INFO [train.py:898] (0/4) Epoch 20, batch 350, loss[loss=0.1494, simple_loss=0.2388, pruned_loss=0.03007, over 18362.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2543, pruned_loss=0.03818, over 3007109.53 frames. ], batch size: 46, lr: 5.71e-03, grad_scale: 4.0 +2023-03-09 12:30:40,081 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69399.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:31:16,069 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69429.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:31:36,507 INFO [train.py:898] (0/4) Epoch 20, batch 400, loss[loss=0.1598, simple_loss=0.2457, pruned_loss=0.03693, over 18502.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2545, pruned_loss=0.03844, over 3129681.46 frames. ], batch size: 47, lr: 5.70e-03, grad_scale: 8.0 +2023-03-09 12:31:47,153 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69455.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:32:00,528 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5199, 3.4937, 2.1964, 4.3787, 3.0601, 4.2597, 2.2978, 3.9843], + device='cuda:0'), covar=tensor([0.0761, 0.0838, 0.1512, 0.0492, 0.0980, 0.0331, 0.1400, 0.0408], + device='cuda:0'), in_proj_covar=tensor([0.0212, 0.0223, 0.0188, 0.0281, 0.0193, 0.0262, 0.0203, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 12:32:08,918 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.671e+02 2.636e+02 3.143e+02 3.755e+02 6.655e+02, threshold=6.287e+02, percent-clipped=1.0 +2023-03-09 12:32:35,587 INFO [train.py:898] (0/4) Epoch 20, batch 450, loss[loss=0.1519, simple_loss=0.2403, pruned_loss=0.03178, over 18406.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2544, pruned_loss=0.03848, over 3222320.98 frames. ], batch size: 48, lr: 5.70e-03, grad_scale: 8.0 +2023-03-09 12:33:00,398 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69518.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:33:07,328 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69524.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:33:33,752 INFO [train.py:898] (0/4) Epoch 20, batch 500, loss[loss=0.1549, simple_loss=0.2454, pruned_loss=0.03219, over 18548.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2546, pruned_loss=0.03827, over 3300283.38 frames. ], batch size: 49, lr: 5.70e-03, grad_scale: 8.0 +2023-03-09 12:34:05,763 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.904e+02 2.776e+02 3.162e+02 3.774e+02 8.469e+02, threshold=6.325e+02, percent-clipped=1.0 +2023-03-09 12:34:14,369 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69581.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:34:24,418 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5715, 5.3438, 5.7636, 5.7811, 5.5284, 6.3589, 6.0157, 5.7313], + device='cuda:0'), covar=tensor([0.1038, 0.0661, 0.0841, 0.0754, 0.1551, 0.0689, 0.0576, 0.1621], + device='cuda:0'), in_proj_covar=tensor([0.0352, 0.0278, 0.0305, 0.0302, 0.0327, 0.0413, 0.0276, 0.0404], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 12:34:32,472 INFO [train.py:898] (0/4) Epoch 20, batch 550, loss[loss=0.1704, simple_loss=0.2654, pruned_loss=0.03767, over 18067.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2543, pruned_loss=0.03808, over 3362910.38 frames. ], batch size: 62, lr: 5.70e-03, grad_scale: 4.0 +2023-03-09 12:34:37,490 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69601.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 12:34:39,955 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7160, 3.6734, 4.9959, 4.2968, 3.2535, 2.8872, 4.4491, 5.2018], + device='cuda:0'), covar=tensor([0.0774, 0.1378, 0.0165, 0.0400, 0.0943, 0.1164, 0.0356, 0.0193], + device='cuda:0'), in_proj_covar=tensor([0.0147, 0.0272, 0.0146, 0.0180, 0.0191, 0.0188, 0.0192, 0.0191], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 12:34:48,722 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-09 12:35:09,940 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69629.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 12:35:15,456 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9940, 3.7667, 5.1491, 2.8192, 4.4629, 2.6823, 3.2364, 1.8158], + device='cuda:0'), covar=tensor([0.1086, 0.0909, 0.0140, 0.0991, 0.0481, 0.2454, 0.2527, 0.2171], + device='cuda:0'), in_proj_covar=tensor([0.0216, 0.0241, 0.0179, 0.0192, 0.0252, 0.0267, 0.0318, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 12:35:31,660 INFO [train.py:898] (0/4) Epoch 20, batch 600, loss[loss=0.1593, simple_loss=0.2485, pruned_loss=0.035, over 18247.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2537, pruned_loss=0.03784, over 3428265.08 frames. ], batch size: 47, lr: 5.69e-03, grad_scale: 4.0 +2023-03-09 12:36:04,103 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=69674.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 12:36:04,877 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.123e+02 2.772e+02 3.241e+02 3.888e+02 6.801e+02, threshold=6.482e+02, percent-clipped=2.0 +2023-03-09 12:36:06,257 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69676.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:36:14,565 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.71 vs. limit=5.0 +2023-03-09 12:36:29,578 INFO [train.py:898] (0/4) Epoch 20, batch 650, loss[loss=0.1703, simple_loss=0.2647, pruned_loss=0.03796, over 18087.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2541, pruned_loss=0.03787, over 3475720.84 frames. ], batch size: 62, lr: 5.69e-03, grad_scale: 4.0 +2023-03-09 12:36:32,774 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69699.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:36:35,289 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-09 12:36:41,921 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7932, 3.1328, 4.4370, 3.9157, 2.7989, 4.7472, 3.9276, 3.1712], + device='cuda:0'), covar=tensor([0.0467, 0.1290, 0.0248, 0.0376, 0.1417, 0.0174, 0.0591, 0.0855], + device='cuda:0'), in_proj_covar=tensor([0.0208, 0.0237, 0.0202, 0.0159, 0.0221, 0.0207, 0.0243, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 12:37:01,616 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69724.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:37:13,989 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=69735.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 12:37:28,029 INFO [train.py:898] (0/4) Epoch 20, batch 700, loss[loss=0.1598, simple_loss=0.2427, pruned_loss=0.0385, over 17687.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2545, pruned_loss=0.03798, over 3508317.89 frames. ], batch size: 39, lr: 5.69e-03, grad_scale: 4.0 +2023-03-09 12:37:28,151 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=69747.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:37:28,405 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7079, 2.9393, 2.6027, 2.9929, 3.6990, 3.5947, 3.2027, 3.0697], + device='cuda:0'), covar=tensor([0.0199, 0.0255, 0.0574, 0.0384, 0.0199, 0.0175, 0.0358, 0.0347], + device='cuda:0'), in_proj_covar=tensor([0.0138, 0.0131, 0.0161, 0.0155, 0.0128, 0.0114, 0.0151, 0.0153], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 12:37:31,614 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=69750.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:37:41,081 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5220, 5.5007, 5.1168, 5.4740, 5.5011, 4.8301, 5.3496, 5.1061], + device='cuda:0'), covar=tensor([0.0431, 0.0448, 0.1309, 0.0765, 0.0494, 0.0393, 0.0451, 0.1011], + device='cuda:0'), in_proj_covar=tensor([0.0478, 0.0544, 0.0689, 0.0425, 0.0439, 0.0491, 0.0539, 0.0667], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 12:37:52,761 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.77 vs. limit=2.0 +2023-03-09 12:38:00,237 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.022e+02 2.614e+02 3.096e+02 3.753e+02 7.667e+02, threshold=6.192e+02, percent-clipped=2.0 +2023-03-09 12:38:26,159 INFO [train.py:898] (0/4) Epoch 20, batch 750, loss[loss=0.1614, simple_loss=0.2521, pruned_loss=0.03538, over 18356.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2549, pruned_loss=0.03804, over 3526479.77 frames. ], batch size: 55, lr: 5.69e-03, grad_scale: 4.0 +2023-03-09 12:38:50,771 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69818.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:38:51,021 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6814, 2.5084, 2.4768, 2.7070, 2.9974, 3.7967, 3.7171, 3.2076], + device='cuda:0'), covar=tensor([0.1588, 0.2142, 0.2686, 0.1687, 0.2088, 0.0464, 0.0574, 0.0750], + device='cuda:0'), in_proj_covar=tensor([0.0298, 0.0344, 0.0378, 0.0276, 0.0390, 0.0237, 0.0293, 0.0251], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 12:38:58,071 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69824.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:39:25,613 INFO [train.py:898] (0/4) Epoch 20, batch 800, loss[loss=0.1574, simple_loss=0.2421, pruned_loss=0.03632, over 18255.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2538, pruned_loss=0.03784, over 3522577.48 frames. ], batch size: 47, lr: 5.69e-03, grad_scale: 8.0 +2023-03-09 12:39:48,073 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=69866.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:39:55,109 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=69872.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:39:58,462 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.014e+02 2.750e+02 3.144e+02 3.939e+02 8.664e+02, threshold=6.288e+02, percent-clipped=4.0 +2023-03-09 12:40:05,913 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69881.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:40:23,200 INFO [train.py:898] (0/4) Epoch 20, batch 850, loss[loss=0.1575, simple_loss=0.2376, pruned_loss=0.03872, over 18164.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2539, pruned_loss=0.03798, over 3535338.18 frames. ], batch size: 44, lr: 5.68e-03, grad_scale: 8.0 +2023-03-09 12:40:28,223 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69901.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 12:41:01,252 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=69929.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:41:01,310 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69929.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 12:41:21,120 INFO [train.py:898] (0/4) Epoch 20, batch 900, loss[loss=0.1768, simple_loss=0.2678, pruned_loss=0.04291, over 18450.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.2533, pruned_loss=0.03765, over 3554814.56 frames. ], batch size: 59, lr: 5.68e-03, grad_scale: 8.0 +2023-03-09 12:41:23,564 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=69949.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 12:41:54,371 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.882e+02 2.754e+02 3.235e+02 3.952e+02 9.067e+02, threshold=6.470e+02, percent-clipped=4.0 +2023-03-09 12:41:55,810 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=69976.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:41:56,853 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=69977.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:42:19,880 INFO [train.py:898] (0/4) Epoch 20, batch 950, loss[loss=0.203, simple_loss=0.285, pruned_loss=0.06048, over 12783.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2541, pruned_loss=0.03783, over 3549535.17 frames. ], batch size: 129, lr: 5.68e-03, grad_scale: 8.0 +2023-03-09 12:42:23,618 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-70000.pt +2023-03-09 12:42:56,934 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=70024.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:42:57,025 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70024.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:43:04,206 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=70030.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 12:43:23,745 INFO [train.py:898] (0/4) Epoch 20, batch 1000, loss[loss=0.1575, simple_loss=0.2518, pruned_loss=0.03167, over 16201.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.254, pruned_loss=0.03794, over 3549022.87 frames. ], batch size: 94, lr: 5.68e-03, grad_scale: 8.0 +2023-03-09 12:43:27,317 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70050.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:43:52,877 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=70072.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:43:56,168 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.914e+02 2.640e+02 3.091e+02 3.737e+02 6.684e+02, threshold=6.182e+02, percent-clipped=1.0 +2023-03-09 12:44:08,238 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8284, 2.9242, 2.7617, 3.1016, 3.8010, 3.6890, 3.3315, 3.0347], + device='cuda:0'), covar=tensor([0.0198, 0.0307, 0.0568, 0.0389, 0.0176, 0.0168, 0.0364, 0.0397], + device='cuda:0'), in_proj_covar=tensor([0.0141, 0.0134, 0.0164, 0.0156, 0.0130, 0.0116, 0.0153, 0.0156], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 12:44:21,991 INFO [train.py:898] (0/4) Epoch 20, batch 1050, loss[loss=0.1817, simple_loss=0.2721, pruned_loss=0.04562, over 17105.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2549, pruned_loss=0.0382, over 3555306.03 frames. ], batch size: 78, lr: 5.68e-03, grad_scale: 8.0 +2023-03-09 12:44:23,276 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=70098.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:44:37,154 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=70110.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:45:20,004 INFO [train.py:898] (0/4) Epoch 20, batch 1100, loss[loss=0.182, simple_loss=0.2731, pruned_loss=0.04539, over 18203.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2541, pruned_loss=0.03788, over 3564981.20 frames. ], batch size: 60, lr: 5.67e-03, grad_scale: 8.0 +2023-03-09 12:45:22,967 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.72 vs. limit=2.0 +2023-03-09 12:45:48,150 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=70171.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:45:50,778 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-09 12:45:52,308 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.767e+02 2.847e+02 3.446e+02 4.067e+02 6.954e+02, threshold=6.891e+02, percent-clipped=3.0 +2023-03-09 12:46:17,937 INFO [train.py:898] (0/4) Epoch 20, batch 1150, loss[loss=0.176, simple_loss=0.2643, pruned_loss=0.04386, over 18570.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.2545, pruned_loss=0.03832, over 3559112.59 frames. ], batch size: 54, lr: 5.67e-03, grad_scale: 8.0 +2023-03-09 12:47:15,963 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9062, 5.0868, 2.6610, 4.8835, 4.7876, 5.1150, 4.9287, 2.6662], + device='cuda:0'), covar=tensor([0.0189, 0.0052, 0.0736, 0.0088, 0.0074, 0.0049, 0.0074, 0.0956], + device='cuda:0'), in_proj_covar=tensor([0.0086, 0.0079, 0.0094, 0.0094, 0.0083, 0.0074, 0.0083, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 12:47:16,772 INFO [train.py:898] (0/4) Epoch 20, batch 1200, loss[loss=0.1814, simple_loss=0.274, pruned_loss=0.04437, over 17816.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2547, pruned_loss=0.03839, over 3573888.62 frames. ], batch size: 70, lr: 5.67e-03, grad_scale: 8.0 +2023-03-09 12:47:31,591 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=70260.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:47:49,158 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.495e+02 2.757e+02 3.193e+02 3.861e+02 7.458e+02, threshold=6.386e+02, percent-clipped=1.0 +2023-03-09 12:47:52,914 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8093, 4.7998, 4.8869, 4.6836, 4.6842, 4.6852, 5.0077, 5.0459], + device='cuda:0'), covar=tensor([0.0075, 0.0070, 0.0062, 0.0106, 0.0062, 0.0153, 0.0068, 0.0085], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0069, 0.0072, 0.0092, 0.0074, 0.0103, 0.0086, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 12:48:15,299 INFO [train.py:898] (0/4) Epoch 20, batch 1250, loss[loss=0.1529, simple_loss=0.241, pruned_loss=0.03241, over 18280.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2548, pruned_loss=0.03838, over 3583198.78 frames. ], batch size: 49, lr: 5.67e-03, grad_scale: 8.0 +2023-03-09 12:48:15,626 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8638, 4.8671, 4.9848, 4.7293, 4.6854, 4.7435, 5.0374, 5.0649], + device='cuda:0'), covar=tensor([0.0083, 0.0076, 0.0065, 0.0112, 0.0072, 0.0164, 0.0078, 0.0099], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0069, 0.0073, 0.0092, 0.0075, 0.0103, 0.0086, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 12:48:31,755 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-09 12:48:36,913 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8927, 3.0536, 2.7525, 3.2184, 3.8603, 3.7751, 3.3596, 3.2915], + device='cuda:0'), covar=tensor([0.0174, 0.0244, 0.0484, 0.0319, 0.0157, 0.0136, 0.0353, 0.0311], + device='cuda:0'), in_proj_covar=tensor([0.0139, 0.0132, 0.0161, 0.0155, 0.0130, 0.0114, 0.0152, 0.0154], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 12:48:42,557 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=70321.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:48:53,763 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70330.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 12:49:13,036 INFO [train.py:898] (0/4) Epoch 20, batch 1300, loss[loss=0.1642, simple_loss=0.2533, pruned_loss=0.03754, over 17046.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2544, pruned_loss=0.03856, over 3586408.34 frames. ], batch size: 78, lr: 5.67e-03, grad_scale: 8.0 +2023-03-09 12:49:41,909 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.31 vs. limit=5.0 +2023-03-09 12:49:44,657 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.016e+02 2.769e+02 3.249e+02 3.833e+02 9.851e+02, threshold=6.498e+02, percent-clipped=4.0 +2023-03-09 12:49:48,779 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=70378.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 12:50:10,106 INFO [train.py:898] (0/4) Epoch 20, batch 1350, loss[loss=0.176, simple_loss=0.2672, pruned_loss=0.04237, over 18637.00 frames. ], tot_loss[loss=0.1666, simple_loss=0.2553, pruned_loss=0.03889, over 3585751.31 frames. ], batch size: 52, lr: 5.66e-03, grad_scale: 8.0 +2023-03-09 12:50:40,009 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6932, 2.3732, 2.6254, 2.7442, 3.2732, 4.9021, 4.7046, 3.2143], + device='cuda:0'), covar=tensor([0.1748, 0.2370, 0.2914, 0.1777, 0.2276, 0.0220, 0.0381, 0.1013], + device='cuda:0'), in_proj_covar=tensor([0.0297, 0.0344, 0.0378, 0.0276, 0.0386, 0.0238, 0.0294, 0.0251], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 12:51:08,590 INFO [train.py:898] (0/4) Epoch 20, batch 1400, loss[loss=0.1605, simple_loss=0.2556, pruned_loss=0.03275, over 18349.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2541, pruned_loss=0.03836, over 3586222.05 frames. ], batch size: 55, lr: 5.66e-03, grad_scale: 8.0 +2023-03-09 12:51:31,205 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=70466.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:51:41,139 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.979e+02 2.850e+02 3.137e+02 3.892e+02 8.751e+02, threshold=6.275e+02, percent-clipped=3.0 +2023-03-09 12:52:06,376 INFO [train.py:898] (0/4) Epoch 20, batch 1450, loss[loss=0.1762, simple_loss=0.2732, pruned_loss=0.03963, over 18581.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2539, pruned_loss=0.03813, over 3589143.67 frames. ], batch size: 54, lr: 5.66e-03, grad_scale: 8.0 +2023-03-09 12:52:18,696 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6119, 4.1985, 4.1565, 3.1843, 3.5263, 3.2446, 2.5228, 2.4169], + device='cuda:0'), covar=tensor([0.0207, 0.0126, 0.0082, 0.0314, 0.0331, 0.0223, 0.0718, 0.0809], + device='cuda:0'), in_proj_covar=tensor([0.0069, 0.0058, 0.0061, 0.0068, 0.0089, 0.0066, 0.0077, 0.0083], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 12:53:04,437 INFO [train.py:898] (0/4) Epoch 20, batch 1500, loss[loss=0.1576, simple_loss=0.2504, pruned_loss=0.03238, over 18623.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2543, pruned_loss=0.03826, over 3596244.19 frames. ], batch size: 52, lr: 5.66e-03, grad_scale: 4.0 +2023-03-09 12:53:38,234 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.836e+02 2.609e+02 3.412e+02 4.380e+02 8.286e+02, threshold=6.824e+02, percent-clipped=3.0 +2023-03-09 12:53:42,454 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.08 vs. limit=5.0 +2023-03-09 12:54:03,312 INFO [train.py:898] (0/4) Epoch 20, batch 1550, loss[loss=0.1377, simple_loss=0.2202, pruned_loss=0.02759, over 18431.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.2539, pruned_loss=0.03763, over 3598778.68 frames. ], batch size: 43, lr: 5.66e-03, grad_scale: 4.0 +2023-03-09 12:54:25,707 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=70616.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:54:28,165 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6525, 3.5255, 4.7958, 4.2145, 3.3543, 2.8072, 4.3137, 5.1123], + device='cuda:0'), covar=tensor([0.0838, 0.1626, 0.0235, 0.0445, 0.0901, 0.1232, 0.0368, 0.0214], + device='cuda:0'), in_proj_covar=tensor([0.0149, 0.0274, 0.0150, 0.0182, 0.0192, 0.0190, 0.0195, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 12:55:02,178 INFO [train.py:898] (0/4) Epoch 20, batch 1600, loss[loss=0.14, simple_loss=0.2196, pruned_loss=0.03023, over 18559.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2546, pruned_loss=0.0378, over 3585450.94 frames. ], batch size: 45, lr: 5.65e-03, grad_scale: 8.0 +2023-03-09 12:55:35,594 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.948e+02 2.737e+02 3.118e+02 3.818e+02 6.822e+02, threshold=6.236e+02, percent-clipped=0.0 +2023-03-09 12:55:59,466 INFO [train.py:898] (0/4) Epoch 20, batch 1650, loss[loss=0.1695, simple_loss=0.261, pruned_loss=0.03897, over 17829.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2546, pruned_loss=0.0378, over 3574941.75 frames. ], batch size: 70, lr: 5.65e-03, grad_scale: 8.0 +2023-03-09 12:56:57,257 INFO [train.py:898] (0/4) Epoch 20, batch 1700, loss[loss=0.1664, simple_loss=0.2591, pruned_loss=0.0368, over 18613.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2546, pruned_loss=0.03773, over 3593002.40 frames. ], batch size: 52, lr: 5.65e-03, grad_scale: 8.0 +2023-03-09 12:57:20,516 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70766.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:57:31,226 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.987e+02 2.775e+02 3.401e+02 4.011e+02 8.447e+02, threshold=6.802e+02, percent-clipped=3.0 +2023-03-09 12:57:38,774 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2308, 5.2484, 4.8622, 5.1682, 5.1683, 4.6006, 5.0842, 4.8519], + device='cuda:0'), covar=tensor([0.0427, 0.0471, 0.1289, 0.0725, 0.0654, 0.0411, 0.0396, 0.1033], + device='cuda:0'), in_proj_covar=tensor([0.0477, 0.0544, 0.0691, 0.0429, 0.0440, 0.0495, 0.0536, 0.0666], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 12:57:55,544 INFO [train.py:898] (0/4) Epoch 20, batch 1750, loss[loss=0.1475, simple_loss=0.2365, pruned_loss=0.02931, over 18290.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.254, pruned_loss=0.03748, over 3594886.13 frames. ], batch size: 49, lr: 5.65e-03, grad_scale: 8.0 +2023-03-09 12:58:03,827 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6489, 2.7340, 2.5888, 2.9243, 3.6594, 3.5560, 3.1285, 2.9015], + device='cuda:0'), covar=tensor([0.0225, 0.0315, 0.0536, 0.0417, 0.0199, 0.0192, 0.0384, 0.0430], + device='cuda:0'), in_proj_covar=tensor([0.0139, 0.0131, 0.0160, 0.0155, 0.0128, 0.0114, 0.0150, 0.0153], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 12:58:16,245 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=70814.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 12:58:39,948 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4334, 5.2791, 5.6251, 5.5212, 5.3201, 6.1621, 5.8403, 5.4414], + device='cuda:0'), covar=tensor([0.1245, 0.0676, 0.0678, 0.0672, 0.1380, 0.0781, 0.0595, 0.1765], + device='cuda:0'), in_proj_covar=tensor([0.0355, 0.0282, 0.0305, 0.0303, 0.0327, 0.0418, 0.0277, 0.0411], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 12:58:41,329 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9581, 4.6699, 4.7791, 3.4907, 3.8785, 3.4834, 2.9755, 2.7559], + device='cuda:0'), covar=tensor([0.0179, 0.0132, 0.0068, 0.0294, 0.0303, 0.0236, 0.0670, 0.0813], + device='cuda:0'), in_proj_covar=tensor([0.0069, 0.0059, 0.0061, 0.0069, 0.0089, 0.0066, 0.0077, 0.0084], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 12:58:43,486 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7050, 3.7571, 3.6070, 3.1897, 3.4516, 2.9159, 2.7748, 3.6650], + device='cuda:0'), covar=tensor([0.0058, 0.0075, 0.0063, 0.0126, 0.0078, 0.0167, 0.0181, 0.0082], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0155, 0.0130, 0.0184, 0.0140, 0.0178, 0.0179, 0.0116], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 12:58:54,614 INFO [train.py:898] (0/4) Epoch 20, batch 1800, loss[loss=0.1739, simple_loss=0.264, pruned_loss=0.04194, over 16203.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2542, pruned_loss=0.03766, over 3582270.39 frames. ], batch size: 94, lr: 5.65e-03, grad_scale: 8.0 +2023-03-09 12:59:28,906 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.702e+02 2.663e+02 3.014e+02 3.631e+02 8.036e+02, threshold=6.028e+02, percent-clipped=1.0 +2023-03-09 12:59:36,207 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6074, 2.2942, 2.4898, 2.5830, 3.0426, 4.4545, 4.3094, 3.3410], + device='cuda:0'), covar=tensor([0.1770, 0.2383, 0.3022, 0.1926, 0.2367, 0.0295, 0.0404, 0.0849], + device='cuda:0'), in_proj_covar=tensor([0.0295, 0.0344, 0.0377, 0.0276, 0.0385, 0.0239, 0.0293, 0.0250], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 12:59:41,826 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-09 12:59:53,193 INFO [train.py:898] (0/4) Epoch 20, batch 1850, loss[loss=0.1588, simple_loss=0.2529, pruned_loss=0.03236, over 18495.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2544, pruned_loss=0.03775, over 3585908.95 frames. ], batch size: 53, lr: 5.64e-03, grad_scale: 8.0 +2023-03-09 13:00:16,573 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8440, 4.9197, 4.9947, 4.6584, 4.7059, 4.6890, 4.9901, 5.0635], + device='cuda:0'), covar=tensor([0.0070, 0.0066, 0.0060, 0.0109, 0.0057, 0.0156, 0.0072, 0.0084], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0070, 0.0073, 0.0092, 0.0075, 0.0104, 0.0086, 0.0086], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 13:00:16,587 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=70916.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:00:51,684 INFO [train.py:898] (0/4) Epoch 20, batch 1900, loss[loss=0.1376, simple_loss=0.2201, pruned_loss=0.02757, over 18595.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2542, pruned_loss=0.03785, over 3587265.08 frames. ], batch size: 45, lr: 5.64e-03, grad_scale: 8.0 +2023-03-09 13:00:59,067 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5765, 2.7325, 2.6004, 2.9496, 3.6019, 3.4525, 3.1345, 2.9102], + device='cuda:0'), covar=tensor([0.0226, 0.0340, 0.0545, 0.0377, 0.0223, 0.0233, 0.0381, 0.0457], + device='cuda:0'), in_proj_covar=tensor([0.0139, 0.0130, 0.0159, 0.0154, 0.0127, 0.0113, 0.0149, 0.0153], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 13:01:09,043 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.49 vs. limit=5.0 +2023-03-09 13:01:11,844 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=70964.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:01:26,017 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.590e+02 2.797e+02 3.325e+02 3.887e+02 8.370e+02, threshold=6.650e+02, percent-clipped=5.0 +2023-03-09 13:01:44,824 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=70992.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:01:50,081 INFO [train.py:898] (0/4) Epoch 20, batch 1950, loss[loss=0.1769, simple_loss=0.2632, pruned_loss=0.04534, over 18386.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2553, pruned_loss=0.03824, over 3591106.34 frames. ], batch size: 52, lr: 5.64e-03, grad_scale: 8.0 +2023-03-09 13:02:28,798 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=71030.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:02:47,807 INFO [train.py:898] (0/4) Epoch 20, batch 2000, loss[loss=0.1656, simple_loss=0.2638, pruned_loss=0.03367, over 18571.00 frames. ], tot_loss[loss=0.1664, simple_loss=0.2557, pruned_loss=0.03856, over 3591347.11 frames. ], batch size: 54, lr: 5.64e-03, grad_scale: 8.0 +2023-03-09 13:02:54,830 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71053.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:03:21,385 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.116e+02 2.863e+02 3.373e+02 3.932e+02 6.166e+02, threshold=6.746e+02, percent-clipped=0.0 +2023-03-09 13:03:39,899 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71091.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:03:46,422 INFO [train.py:898] (0/4) Epoch 20, batch 2050, loss[loss=0.1729, simple_loss=0.2706, pruned_loss=0.03757, over 18301.00 frames. ], tot_loss[loss=0.1663, simple_loss=0.2556, pruned_loss=0.03849, over 3586533.75 frames. ], batch size: 54, lr: 5.64e-03, grad_scale: 8.0 +2023-03-09 13:04:45,465 INFO [train.py:898] (0/4) Epoch 20, batch 2100, loss[loss=0.1776, simple_loss=0.2525, pruned_loss=0.05132, over 18398.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.255, pruned_loss=0.03816, over 3583723.00 frames. ], batch size: 43, lr: 5.63e-03, grad_scale: 8.0 +2023-03-09 13:04:47,092 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.4621, 2.0473, 2.0806, 2.1689, 2.4361, 2.4515, 2.3471, 2.1261], + device='cuda:0'), covar=tensor([0.0283, 0.0242, 0.0472, 0.0391, 0.0227, 0.0204, 0.0387, 0.0353], + device='cuda:0'), in_proj_covar=tensor([0.0137, 0.0130, 0.0159, 0.0153, 0.0127, 0.0113, 0.0149, 0.0153], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 13:05:19,430 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.825e+02 2.628e+02 3.039e+02 3.790e+02 1.100e+03, threshold=6.078e+02, percent-clipped=2.0 +2023-03-09 13:05:44,230 INFO [train.py:898] (0/4) Epoch 20, batch 2150, loss[loss=0.1558, simple_loss=0.2348, pruned_loss=0.0384, over 18504.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2541, pruned_loss=0.03762, over 3584573.83 frames. ], batch size: 44, lr: 5.63e-03, grad_scale: 8.0 +2023-03-09 13:06:09,721 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0477, 5.1124, 2.4796, 5.0105, 4.7729, 5.1307, 4.8562, 2.2363], + device='cuda:0'), covar=tensor([0.0223, 0.0141, 0.1076, 0.0113, 0.0117, 0.0133, 0.0176, 0.1741], + device='cuda:0'), in_proj_covar=tensor([0.0086, 0.0080, 0.0095, 0.0095, 0.0084, 0.0075, 0.0084, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 13:06:18,601 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8675, 4.8717, 4.8878, 4.6625, 4.7541, 4.7270, 4.9952, 5.0237], + device='cuda:0'), covar=tensor([0.0068, 0.0067, 0.0077, 0.0109, 0.0058, 0.0148, 0.0065, 0.0086], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0070, 0.0073, 0.0092, 0.0074, 0.0103, 0.0087, 0.0086], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 13:06:43,284 INFO [train.py:898] (0/4) Epoch 20, batch 2200, loss[loss=0.166, simple_loss=0.2553, pruned_loss=0.03835, over 17973.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2549, pruned_loss=0.03803, over 3584306.74 frames. ], batch size: 65, lr: 5.63e-03, grad_scale: 8.0 +2023-03-09 13:07:16,427 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.980e+02 2.743e+02 3.244e+02 3.938e+02 1.174e+03, threshold=6.489e+02, percent-clipped=4.0 +2023-03-09 13:07:41,311 INFO [train.py:898] (0/4) Epoch 20, batch 2250, loss[loss=0.1808, simple_loss=0.2799, pruned_loss=0.04079, over 18625.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2543, pruned_loss=0.03782, over 3582059.64 frames. ], batch size: 52, lr: 5.63e-03, grad_scale: 8.0 +2023-03-09 13:08:13,267 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-09 13:08:40,030 INFO [train.py:898] (0/4) Epoch 20, batch 2300, loss[loss=0.1907, simple_loss=0.2739, pruned_loss=0.05379, over 17994.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2551, pruned_loss=0.03828, over 3578278.37 frames. ], batch size: 65, lr: 5.63e-03, grad_scale: 8.0 +2023-03-09 13:08:41,360 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=71348.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:09:13,544 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.760e+02 2.586e+02 3.152e+02 3.675e+02 6.468e+02, threshold=6.303e+02, percent-clipped=0.0 +2023-03-09 13:09:25,899 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=71386.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:09:28,361 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4077, 3.2430, 2.0455, 4.1812, 2.7667, 4.0579, 2.2040, 3.6740], + device='cuda:0'), covar=tensor([0.0664, 0.0925, 0.1561, 0.0461, 0.1032, 0.0287, 0.1320, 0.0430], + device='cuda:0'), in_proj_covar=tensor([0.0211, 0.0223, 0.0188, 0.0280, 0.0191, 0.0261, 0.0200, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 13:09:38,247 INFO [train.py:898] (0/4) Epoch 20, batch 2350, loss[loss=0.1821, simple_loss=0.2779, pruned_loss=0.04316, over 18388.00 frames. ], tot_loss[loss=0.166, simple_loss=0.2555, pruned_loss=0.03827, over 3572357.43 frames. ], batch size: 52, lr: 5.62e-03, grad_scale: 8.0 +2023-03-09 13:09:39,737 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6937, 3.7660, 3.6390, 3.2211, 3.4606, 2.8998, 2.8994, 3.7975], + device='cuda:0'), covar=tensor([0.0064, 0.0080, 0.0078, 0.0142, 0.0108, 0.0188, 0.0188, 0.0056], + device='cuda:0'), in_proj_covar=tensor([0.0137, 0.0156, 0.0130, 0.0185, 0.0141, 0.0178, 0.0180, 0.0118], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 13:09:47,404 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7962, 3.2105, 4.0236, 3.0313, 3.7054, 2.6747, 2.7858, 2.3323], + device='cuda:0'), covar=tensor([0.0941, 0.0844, 0.0235, 0.0557, 0.0636, 0.2106, 0.2121, 0.1537], + device='cuda:0'), in_proj_covar=tensor([0.0217, 0.0241, 0.0182, 0.0193, 0.0255, 0.0268, 0.0318, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 13:10:23,218 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-03-09 13:10:37,016 INFO [train.py:898] (0/4) Epoch 20, batch 2400, loss[loss=0.1776, simple_loss=0.2668, pruned_loss=0.04417, over 18267.00 frames. ], tot_loss[loss=0.1661, simple_loss=0.2552, pruned_loss=0.03846, over 3569232.95 frames. ], batch size: 60, lr: 5.62e-03, grad_scale: 8.0 +2023-03-09 13:11:10,786 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.729e+02 2.680e+02 3.169e+02 3.609e+02 6.256e+02, threshold=6.338e+02, percent-clipped=0.0 +2023-03-09 13:11:35,512 INFO [train.py:898] (0/4) Epoch 20, batch 2450, loss[loss=0.1659, simple_loss=0.2568, pruned_loss=0.03745, over 18568.00 frames. ], tot_loss[loss=0.1657, simple_loss=0.2547, pruned_loss=0.03833, over 3578256.62 frames. ], batch size: 54, lr: 5.62e-03, grad_scale: 8.0 +2023-03-09 13:12:16,816 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6883, 5.2243, 5.1981, 5.2493, 4.6620, 5.1192, 4.5287, 5.0669], + device='cuda:0'), covar=tensor([0.0276, 0.0295, 0.0202, 0.0393, 0.0393, 0.0235, 0.1152, 0.0332], + device='cuda:0'), in_proj_covar=tensor([0.0212, 0.0258, 0.0247, 0.0323, 0.0265, 0.0267, 0.0306, 0.0257], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 13:12:33,799 INFO [train.py:898] (0/4) Epoch 20, batch 2500, loss[loss=0.1932, simple_loss=0.2844, pruned_loss=0.051, over 18197.00 frames. ], tot_loss[loss=0.1659, simple_loss=0.2547, pruned_loss=0.03853, over 3567189.76 frames. ], batch size: 60, lr: 5.62e-03, grad_scale: 8.0 +2023-03-09 13:13:07,752 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.564e+02 2.655e+02 3.143e+02 3.839e+02 8.943e+02, threshold=6.287e+02, percent-clipped=3.0 +2023-03-09 13:13:31,477 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9392, 2.5358, 2.0828, 2.5942, 3.0249, 2.9616, 2.8239, 2.6618], + device='cuda:0'), covar=tensor([0.0241, 0.0235, 0.0624, 0.0311, 0.0189, 0.0192, 0.0301, 0.0290], + device='cuda:0'), in_proj_covar=tensor([0.0140, 0.0131, 0.0160, 0.0154, 0.0127, 0.0114, 0.0150, 0.0152], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 13:13:32,181 INFO [train.py:898] (0/4) Epoch 20, batch 2550, loss[loss=0.1831, simple_loss=0.2725, pruned_loss=0.04686, over 15958.00 frames. ], tot_loss[loss=0.166, simple_loss=0.255, pruned_loss=0.03847, over 3581440.17 frames. ], batch size: 94, lr: 5.62e-03, grad_scale: 8.0 +2023-03-09 13:13:55,596 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=71616.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:14:31,305 INFO [train.py:898] (0/4) Epoch 20, batch 2600, loss[loss=0.2016, simple_loss=0.2838, pruned_loss=0.0597, over 13013.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2547, pruned_loss=0.03812, over 3577514.58 frames. ], batch size: 130, lr: 5.62e-03, grad_scale: 8.0 +2023-03-09 13:14:32,699 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=71648.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:15:05,045 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.778e+02 2.596e+02 2.940e+02 3.683e+02 1.082e+03, threshold=5.881e+02, percent-clipped=3.0 +2023-03-09 13:15:06,481 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71677.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:15:16,550 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=71686.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:15:28,065 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=71696.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:15:29,061 INFO [train.py:898] (0/4) Epoch 20, batch 2650, loss[loss=0.1708, simple_loss=0.2644, pruned_loss=0.0386, over 18494.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.2548, pruned_loss=0.03807, over 3578636.23 frames. ], batch size: 53, lr: 5.61e-03, grad_scale: 8.0 +2023-03-09 13:16:12,126 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=71734.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:16:27,158 INFO [train.py:898] (0/4) Epoch 20, batch 2700, loss[loss=0.1386, simple_loss=0.2244, pruned_loss=0.02635, over 18432.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2544, pruned_loss=0.03785, over 3586291.95 frames. ], batch size: 43, lr: 5.61e-03, grad_scale: 8.0 +2023-03-09 13:16:42,954 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-09 13:17:01,229 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.908e+02 2.794e+02 3.196e+02 3.940e+02 6.442e+02, threshold=6.393e+02, percent-clipped=2.0 +2023-03-09 13:17:02,027 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-09 13:17:20,530 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0722, 5.1392, 5.3767, 5.4483, 5.0710, 5.9557, 5.6361, 5.2580], + device='cuda:0'), covar=tensor([0.1164, 0.0686, 0.0711, 0.0775, 0.1484, 0.0769, 0.0673, 0.1543], + device='cuda:0'), in_proj_covar=tensor([0.0353, 0.0283, 0.0306, 0.0304, 0.0327, 0.0416, 0.0278, 0.0409], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 13:17:25,384 INFO [train.py:898] (0/4) Epoch 20, batch 2750, loss[loss=0.142, simple_loss=0.2336, pruned_loss=0.02525, over 18492.00 frames. ], tot_loss[loss=0.1656, simple_loss=0.255, pruned_loss=0.03809, over 3587250.77 frames. ], batch size: 47, lr: 5.61e-03, grad_scale: 8.0 +2023-03-09 13:18:23,265 INFO [train.py:898] (0/4) Epoch 20, batch 2800, loss[loss=0.181, simple_loss=0.277, pruned_loss=0.04248, over 18290.00 frames. ], tot_loss[loss=0.1654, simple_loss=0.255, pruned_loss=0.03788, over 3592897.56 frames. ], batch size: 57, lr: 5.61e-03, grad_scale: 8.0 +2023-03-09 13:18:29,924 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0681, 4.1654, 2.5835, 4.2797, 5.2829, 2.8261, 3.9621, 4.0360], + device='cuda:0'), covar=tensor([0.0155, 0.1388, 0.1549, 0.0541, 0.0077, 0.1109, 0.0596, 0.0724], + device='cuda:0'), in_proj_covar=tensor([0.0162, 0.0268, 0.0204, 0.0194, 0.0124, 0.0181, 0.0215, 0.0224], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 13:18:32,408 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-09 13:18:33,550 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-09 13:18:56,840 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.812e+02 2.512e+02 3.220e+02 3.992e+02 1.001e+03, threshold=6.440e+02, percent-clipped=3.0 +2023-03-09 13:19:00,235 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6562, 2.2392, 2.5357, 2.6779, 3.2194, 4.8165, 4.5704, 3.3222], + device='cuda:0'), covar=tensor([0.1837, 0.2497, 0.2987, 0.1872, 0.2242, 0.0216, 0.0396, 0.0938], + device='cuda:0'), in_proj_covar=tensor([0.0297, 0.0345, 0.0379, 0.0276, 0.0387, 0.0239, 0.0294, 0.0253], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 13:19:22,130 INFO [train.py:898] (0/4) Epoch 20, batch 2850, loss[loss=0.1698, simple_loss=0.2634, pruned_loss=0.03812, over 18393.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2542, pruned_loss=0.03763, over 3600726.21 frames. ], batch size: 52, lr: 5.61e-03, grad_scale: 8.0 +2023-03-09 13:19:41,239 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-09 13:19:49,001 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=71919.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:20:21,472 INFO [train.py:898] (0/4) Epoch 20, batch 2900, loss[loss=0.1615, simple_loss=0.2576, pruned_loss=0.03275, over 16006.00 frames. ], tot_loss[loss=0.1655, simple_loss=0.2551, pruned_loss=0.03795, over 3587435.35 frames. ], batch size: 94, lr: 5.60e-03, grad_scale: 8.0 +2023-03-09 13:20:51,382 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=71972.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:20:55,781 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.736e+02 2.545e+02 2.901e+02 3.371e+02 5.685e+02, threshold=5.802e+02, percent-clipped=0.0 +2023-03-09 13:21:00,729 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=71980.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:21:20,629 INFO [train.py:898] (0/4) Epoch 20, batch 2950, loss[loss=0.1529, simple_loss=0.2476, pruned_loss=0.02904, over 18546.00 frames. ], tot_loss[loss=0.1653, simple_loss=0.2549, pruned_loss=0.03785, over 3588249.85 frames. ], batch size: 49, lr: 5.60e-03, grad_scale: 8.0 +2023-03-09 13:21:24,238 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-72000.pt +2023-03-09 13:21:44,017 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5293, 2.2155, 2.4605, 2.5318, 2.9437, 4.5672, 4.3455, 3.3621], + device='cuda:0'), covar=tensor([0.1842, 0.2578, 0.3173, 0.1902, 0.2601, 0.0294, 0.0461, 0.0861], + device='cuda:0'), in_proj_covar=tensor([0.0298, 0.0346, 0.0380, 0.0277, 0.0388, 0.0240, 0.0295, 0.0254], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 13:22:17,076 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2377, 5.4617, 3.0250, 5.2934, 5.1080, 5.3764, 5.1771, 2.5539], + device='cuda:0'), covar=tensor([0.0183, 0.0079, 0.0773, 0.0102, 0.0089, 0.0129, 0.0141, 0.1439], + device='cuda:0'), in_proj_covar=tensor([0.0085, 0.0079, 0.0093, 0.0093, 0.0083, 0.0074, 0.0083, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 13:22:17,713 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-09 13:22:24,556 INFO [train.py:898] (0/4) Epoch 20, batch 3000, loss[loss=0.1716, simple_loss=0.2557, pruned_loss=0.04379, over 18277.00 frames. ], tot_loss[loss=0.1647, simple_loss=0.2542, pruned_loss=0.03758, over 3584869.32 frames. ], batch size: 57, lr: 5.60e-03, grad_scale: 8.0 +2023-03-09 13:22:24,558 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 13:22:36,474 INFO [train.py:932] (0/4) Epoch 20, validation: loss=0.1501, simple_loss=0.25, pruned_loss=0.02514, over 944034.00 frames. +2023-03-09 13:22:36,475 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 13:22:45,380 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7855, 3.6846, 5.0085, 4.5246, 3.3839, 2.9569, 4.4944, 5.2664], + device='cuda:0'), covar=tensor([0.0810, 0.1552, 0.0240, 0.0361, 0.0895, 0.1220, 0.0380, 0.0332], + device='cuda:0'), in_proj_covar=tensor([0.0148, 0.0272, 0.0149, 0.0180, 0.0190, 0.0191, 0.0194, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 13:23:10,304 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.792e+02 2.528e+02 3.006e+02 3.462e+02 4.966e+02, threshold=6.013e+02, percent-clipped=0.0 +2023-03-09 13:23:33,867 INFO [train.py:898] (0/4) Epoch 20, batch 3050, loss[loss=0.1885, simple_loss=0.2784, pruned_loss=0.04929, over 18562.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2547, pruned_loss=0.03774, over 3583884.36 frames. ], batch size: 54, lr: 5.60e-03, grad_scale: 8.0 +2023-03-09 13:24:11,857 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72129.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:24:31,886 INFO [train.py:898] (0/4) Epoch 20, batch 3100, loss[loss=0.1486, simple_loss=0.2356, pruned_loss=0.03076, over 18343.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2543, pruned_loss=0.03778, over 3575366.31 frames. ], batch size: 46, lr: 5.60e-03, grad_scale: 8.0 +2023-03-09 13:24:54,907 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.76 vs. limit=2.0 +2023-03-09 13:25:05,467 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.922e+02 2.879e+02 3.465e+02 4.058e+02 1.741e+03, threshold=6.931e+02, percent-clipped=3.0 +2023-03-09 13:25:21,345 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0084, 3.7108, 5.0647, 2.8568, 4.3512, 2.6083, 3.1793, 1.9011], + device='cuda:0'), covar=tensor([0.1071, 0.0919, 0.0126, 0.0985, 0.0524, 0.2471, 0.2476, 0.2130], + device='cuda:0'), in_proj_covar=tensor([0.0217, 0.0243, 0.0184, 0.0193, 0.0255, 0.0268, 0.0319, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 13:25:22,415 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=72190.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:25:29,912 INFO [train.py:898] (0/4) Epoch 20, batch 3150, loss[loss=0.1557, simple_loss=0.2491, pruned_loss=0.03116, over 18290.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.2541, pruned_loss=0.03761, over 3585635.51 frames. ], batch size: 49, lr: 5.59e-03, grad_scale: 8.0 +2023-03-09 13:26:28,292 INFO [train.py:898] (0/4) Epoch 20, batch 3200, loss[loss=0.1561, simple_loss=0.245, pruned_loss=0.03357, over 18244.00 frames. ], tot_loss[loss=0.165, simple_loss=0.254, pruned_loss=0.03799, over 3572286.66 frames. ], batch size: 47, lr: 5.59e-03, grad_scale: 8.0 +2023-03-09 13:26:58,252 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=72272.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:27:01,425 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=72275.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:27:02,248 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.897e+02 2.541e+02 3.039e+02 3.727e+02 6.894e+02, threshold=6.078e+02, percent-clipped=0.0 +2023-03-09 13:27:26,960 INFO [train.py:898] (0/4) Epoch 20, batch 3250, loss[loss=0.1715, simple_loss=0.2645, pruned_loss=0.03929, over 18318.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2541, pruned_loss=0.03806, over 3576226.91 frames. ], batch size: 54, lr: 5.59e-03, grad_scale: 8.0 +2023-03-09 13:27:54,550 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=72320.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:28:01,537 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72326.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:28:26,307 INFO [train.py:898] (0/4) Epoch 20, batch 3300, loss[loss=0.1699, simple_loss=0.2597, pruned_loss=0.04007, over 17710.00 frames. ], tot_loss[loss=0.1651, simple_loss=0.2543, pruned_loss=0.03794, over 3580756.92 frames. ], batch size: 70, lr: 5.59e-03, grad_scale: 8.0 +2023-03-09 13:28:59,547 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.671e+02 2.585e+02 3.093e+02 3.745e+02 6.095e+02, threshold=6.186e+02, percent-clipped=1.0 +2023-03-09 13:29:12,907 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=72387.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:29:24,267 INFO [train.py:898] (0/4) Epoch 20, batch 3350, loss[loss=0.1958, simple_loss=0.2871, pruned_loss=0.05227, over 16065.00 frames. ], tot_loss[loss=0.1652, simple_loss=0.2542, pruned_loss=0.03814, over 3574135.70 frames. ], batch size: 94, lr: 5.59e-03, grad_scale: 8.0 +2023-03-09 13:30:23,820 INFO [train.py:898] (0/4) Epoch 20, batch 3400, loss[loss=0.1421, simple_loss=0.2242, pruned_loss=0.03005, over 18415.00 frames. ], tot_loss[loss=0.1658, simple_loss=0.2547, pruned_loss=0.03847, over 3578617.60 frames. ], batch size: 42, lr: 5.58e-03, grad_scale: 8.0 +2023-03-09 13:30:34,242 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9272, 4.6134, 4.7008, 3.5786, 3.8140, 3.5182, 2.8746, 2.6036], + device='cuda:0'), covar=tensor([0.0190, 0.0171, 0.0079, 0.0278, 0.0319, 0.0227, 0.0626, 0.0775], + device='cuda:0'), in_proj_covar=tensor([0.0069, 0.0058, 0.0061, 0.0067, 0.0087, 0.0065, 0.0076, 0.0082], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 13:30:57,199 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.915e+02 2.809e+02 3.353e+02 3.945e+02 1.222e+03, threshold=6.706e+02, percent-clipped=5.0 +2023-03-09 13:31:07,727 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=72485.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:31:22,241 INFO [train.py:898] (0/4) Epoch 20, batch 3450, loss[loss=0.1395, simple_loss=0.2216, pruned_loss=0.02871, over 18403.00 frames. ], tot_loss[loss=0.1648, simple_loss=0.2536, pruned_loss=0.03798, over 3579429.01 frames. ], batch size: 43, lr: 5.58e-03, grad_scale: 8.0 +2023-03-09 13:32:08,485 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4592, 2.7241, 2.4327, 2.8331, 3.5778, 3.4076, 3.0549, 2.8852], + device='cuda:0'), covar=tensor([0.0207, 0.0291, 0.0626, 0.0386, 0.0217, 0.0193, 0.0388, 0.0377], + device='cuda:0'), in_proj_covar=tensor([0.0141, 0.0131, 0.0162, 0.0156, 0.0129, 0.0115, 0.0152, 0.0155], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 13:32:20,078 INFO [train.py:898] (0/4) Epoch 20, batch 3500, loss[loss=0.1523, simple_loss=0.2452, pruned_loss=0.0297, over 18494.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2533, pruned_loss=0.03784, over 3570445.51 frames. ], batch size: 51, lr: 5.58e-03, grad_scale: 16.0 +2023-03-09 13:32:52,106 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=72575.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:32:53,021 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.775e+02 2.575e+02 2.979e+02 3.509e+02 6.316e+02, threshold=5.957e+02, percent-clipped=0.0 +2023-03-09 13:33:16,608 INFO [train.py:898] (0/4) Epoch 20, batch 3550, loss[loss=0.1559, simple_loss=0.249, pruned_loss=0.03139, over 18473.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2534, pruned_loss=0.03777, over 3582115.43 frames. ], batch size: 53, lr: 5.58e-03, grad_scale: 16.0 +2023-03-09 13:33:45,203 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=72623.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:34:10,349 INFO [train.py:898] (0/4) Epoch 20, batch 3600, loss[loss=0.1692, simple_loss=0.265, pruned_loss=0.0367, over 15925.00 frames. ], tot_loss[loss=0.1649, simple_loss=0.2539, pruned_loss=0.03799, over 3576285.44 frames. ], batch size: 94, lr: 5.58e-03, grad_scale: 8.0 +2023-03-09 13:34:26,276 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.68 vs. limit=5.0 +2023-03-09 13:34:42,422 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.959e+02 2.531e+02 3.153e+02 3.638e+02 9.354e+02, threshold=6.307e+02, percent-clipped=0.0 +2023-03-09 13:34:46,628 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-20.pt +2023-03-09 13:35:15,796 INFO [train.py:898] (0/4) Epoch 21, batch 0, loss[loss=0.1357, simple_loss=0.2154, pruned_loss=0.028, over 18379.00 frames. ], tot_loss[loss=0.1357, simple_loss=0.2154, pruned_loss=0.028, over 18379.00 frames. ], batch size: 42, lr: 5.44e-03, grad_scale: 8.0 +2023-03-09 13:35:15,798 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 13:35:27,496 INFO [train.py:932] (0/4) Epoch 21, validation: loss=0.1511, simple_loss=0.2511, pruned_loss=0.02556, over 944034.00 frames. +2023-03-09 13:35:27,497 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 13:35:28,910 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=72682.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:36:23,114 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6751, 3.8660, 2.3655, 3.9882, 4.9522, 2.5590, 3.7220, 3.7594], + device='cuda:0'), covar=tensor([0.0189, 0.1320, 0.1689, 0.0562, 0.0088, 0.1257, 0.0666, 0.0748], + device='cuda:0'), in_proj_covar=tensor([0.0164, 0.0269, 0.0202, 0.0194, 0.0125, 0.0183, 0.0215, 0.0222], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 13:36:26,171 INFO [train.py:898] (0/4) Epoch 21, batch 50, loss[loss=0.1736, simple_loss=0.2619, pruned_loss=0.04272, over 17094.00 frames. ], tot_loss[loss=0.1628, simple_loss=0.2518, pruned_loss=0.0369, over 815465.94 frames. ], batch size: 78, lr: 5.44e-03, grad_scale: 8.0 +2023-03-09 13:37:20,700 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.036e+02 2.565e+02 3.180e+02 3.639e+02 9.362e+02, threshold=6.360e+02, percent-clipped=2.0 +2023-03-09 13:37:25,029 INFO [train.py:898] (0/4) Epoch 21, batch 100, loss[loss=0.1771, simple_loss=0.2654, pruned_loss=0.0444, over 18116.00 frames. ], tot_loss[loss=0.1622, simple_loss=0.2512, pruned_loss=0.03664, over 1430751.88 frames. ], batch size: 62, lr: 5.43e-03, grad_scale: 8.0 +2023-03-09 13:37:29,818 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=72785.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:37:49,302 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0170, 3.7401, 5.1441, 2.8635, 4.5834, 2.6257, 3.1463, 1.8426], + device='cuda:0'), covar=tensor([0.1114, 0.0942, 0.0132, 0.0952, 0.0516, 0.2610, 0.2696, 0.2180], + device='cuda:0'), in_proj_covar=tensor([0.0220, 0.0244, 0.0186, 0.0196, 0.0257, 0.0271, 0.0324, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 13:38:23,942 INFO [train.py:898] (0/4) Epoch 21, batch 150, loss[loss=0.1722, simple_loss=0.2707, pruned_loss=0.03679, over 18395.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2527, pruned_loss=0.03687, over 1906747.08 frames. ], batch size: 52, lr: 5.43e-03, grad_scale: 8.0 +2023-03-09 13:38:26,431 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=72833.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:39:17,914 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.843e+02 2.851e+02 3.537e+02 4.282e+02 1.325e+03, threshold=7.073e+02, percent-clipped=5.0 +2023-03-09 13:39:22,598 INFO [train.py:898] (0/4) Epoch 21, batch 200, loss[loss=0.1364, simple_loss=0.2208, pruned_loss=0.02602, over 18490.00 frames. ], tot_loss[loss=0.1623, simple_loss=0.2513, pruned_loss=0.03666, over 2271018.32 frames. ], batch size: 44, lr: 5.43e-03, grad_scale: 8.0 +2023-03-09 13:39:37,505 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72894.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:40:04,453 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8213, 4.1701, 4.1513, 4.1786, 3.8572, 4.1002, 3.7700, 4.0907], + device='cuda:0'), covar=tensor([0.0289, 0.0371, 0.0255, 0.0545, 0.0344, 0.0269, 0.0905, 0.0373], + device='cuda:0'), in_proj_covar=tensor([0.0213, 0.0261, 0.0251, 0.0326, 0.0265, 0.0268, 0.0305, 0.0257], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 13:40:20,985 INFO [train.py:898] (0/4) Epoch 21, batch 250, loss[loss=0.1699, simple_loss=0.2622, pruned_loss=0.03882, over 18487.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2521, pruned_loss=0.03693, over 2564067.41 frames. ], batch size: 53, lr: 5.43e-03, grad_scale: 8.0 +2023-03-09 13:40:48,957 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=72955.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:40:59,839 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=72964.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 13:41:14,357 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.784e+02 2.579e+02 3.135e+02 3.848e+02 6.941e+02, threshold=6.270e+02, percent-clipped=0.0 +2023-03-09 13:41:18,868 INFO [train.py:898] (0/4) Epoch 21, batch 300, loss[loss=0.1621, simple_loss=0.2565, pruned_loss=0.03384, over 18485.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.2513, pruned_loss=0.03671, over 2801505.41 frames. ], batch size: 53, lr: 5.43e-03, grad_scale: 8.0 +2023-03-09 13:41:20,303 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=72982.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:42:11,291 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73025.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 13:42:16,849 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=73030.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:42:17,822 INFO [train.py:898] (0/4) Epoch 21, batch 350, loss[loss=0.1605, simple_loss=0.2513, pruned_loss=0.0349, over 16982.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2514, pruned_loss=0.03692, over 2979241.45 frames. ], batch size: 78, lr: 5.43e-03, grad_scale: 8.0 +2023-03-09 13:42:38,266 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5412, 3.4571, 4.6531, 4.1297, 3.1205, 2.9795, 4.2082, 4.8553], + device='cuda:0'), covar=tensor([0.0894, 0.1471, 0.0216, 0.0411, 0.1041, 0.1198, 0.0417, 0.0297], + device='cuda:0'), in_proj_covar=tensor([0.0148, 0.0273, 0.0150, 0.0180, 0.0191, 0.0190, 0.0194, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 13:43:11,933 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.597e+02 2.497e+02 3.036e+02 3.677e+02 5.898e+02, threshold=6.073e+02, percent-clipped=0.0 +2023-03-09 13:43:16,484 INFO [train.py:898] (0/4) Epoch 21, batch 400, loss[loss=0.1902, simple_loss=0.2782, pruned_loss=0.0511, over 18298.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2519, pruned_loss=0.03704, over 3116135.05 frames. ], batch size: 57, lr: 5.42e-03, grad_scale: 8.0 +2023-03-09 13:43:54,906 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4225, 3.2380, 2.0519, 4.2814, 2.8230, 4.0486, 2.4496, 3.7510], + device='cuda:0'), covar=tensor([0.0697, 0.0939, 0.1587, 0.0464, 0.0939, 0.0323, 0.1202, 0.0433], + device='cuda:0'), in_proj_covar=tensor([0.0212, 0.0225, 0.0188, 0.0281, 0.0190, 0.0261, 0.0203, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 13:44:06,318 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7965, 3.8849, 5.0649, 4.4393, 3.4225, 3.1238, 4.6101, 5.3460], + device='cuda:0'), covar=tensor([0.0741, 0.1213, 0.0170, 0.0338, 0.0811, 0.1057, 0.0334, 0.0172], + device='cuda:0'), in_proj_covar=tensor([0.0147, 0.0273, 0.0149, 0.0180, 0.0191, 0.0190, 0.0194, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 13:44:14,450 INFO [train.py:898] (0/4) Epoch 21, batch 450, loss[loss=0.1762, simple_loss=0.2681, pruned_loss=0.04219, over 17847.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.2529, pruned_loss=0.0373, over 3222225.44 frames. ], batch size: 70, lr: 5.42e-03, grad_scale: 8.0 +2023-03-09 13:44:14,923 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5960, 2.2354, 2.5361, 2.5829, 3.0906, 4.7831, 4.6276, 3.3550], + device='cuda:0'), covar=tensor([0.1842, 0.2505, 0.3133, 0.1959, 0.2474, 0.0227, 0.0393, 0.0899], + device='cuda:0'), in_proj_covar=tensor([0.0300, 0.0347, 0.0382, 0.0279, 0.0390, 0.0240, 0.0297, 0.0255], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 13:45:06,943 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.937e+02 2.746e+02 3.322e+02 3.984e+02 6.045e+02, threshold=6.644e+02, percent-clipped=0.0 +2023-03-09 13:45:12,779 INFO [train.py:898] (0/4) Epoch 21, batch 500, loss[loss=0.172, simple_loss=0.2661, pruned_loss=0.03896, over 18309.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.2528, pruned_loss=0.03741, over 3302180.70 frames. ], batch size: 57, lr: 5.42e-03, grad_scale: 8.0 +2023-03-09 13:45:38,517 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4862, 2.7149, 2.3524, 2.8390, 3.5711, 3.5381, 3.0184, 2.9047], + device='cuda:0'), covar=tensor([0.0202, 0.0331, 0.0573, 0.0402, 0.0210, 0.0162, 0.0379, 0.0369], + device='cuda:0'), in_proj_covar=tensor([0.0140, 0.0133, 0.0162, 0.0155, 0.0129, 0.0115, 0.0152, 0.0154], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 13:46:10,611 INFO [train.py:898] (0/4) Epoch 21, batch 550, loss[loss=0.1663, simple_loss=0.26, pruned_loss=0.03628, over 18491.00 frames. ], tot_loss[loss=0.164, simple_loss=0.2532, pruned_loss=0.03741, over 3359154.49 frames. ], batch size: 51, lr: 5.42e-03, grad_scale: 8.0 +2023-03-09 13:46:32,269 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73250.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:46:53,140 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.82 vs. limit=5.0 +2023-03-09 13:47:03,604 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.734e+02 2.558e+02 3.003e+02 3.637e+02 1.061e+03, threshold=6.005e+02, percent-clipped=3.0 +2023-03-09 13:47:08,046 INFO [train.py:898] (0/4) Epoch 21, batch 600, loss[loss=0.1587, simple_loss=0.2418, pruned_loss=0.03781, over 17647.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2531, pruned_loss=0.03741, over 3384574.05 frames. ], batch size: 39, lr: 5.42e-03, grad_scale: 8.0 +2023-03-09 13:47:22,554 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=73293.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:47:53,078 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73320.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 13:48:06,375 INFO [train.py:898] (0/4) Epoch 21, batch 650, loss[loss=0.1634, simple_loss=0.2564, pruned_loss=0.03522, over 18333.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2527, pruned_loss=0.03727, over 3420755.43 frames. ], batch size: 57, lr: 5.41e-03, grad_scale: 8.0 +2023-03-09 13:48:33,485 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73354.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:49:00,062 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.798e+02 2.707e+02 3.166e+02 4.000e+02 8.066e+02, threshold=6.331e+02, percent-clipped=2.0 +2023-03-09 13:49:04,586 INFO [train.py:898] (0/4) Epoch 21, batch 700, loss[loss=0.1695, simple_loss=0.2652, pruned_loss=0.03686, over 18564.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.2522, pruned_loss=0.03715, over 3451910.74 frames. ], batch size: 54, lr: 5.41e-03, grad_scale: 8.0 +2023-03-09 13:49:43,811 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6117, 2.1615, 2.4968, 2.5939, 3.0374, 4.9307, 4.8421, 3.6607], + device='cuda:0'), covar=tensor([0.1943, 0.2817, 0.3074, 0.1990, 0.2656, 0.0230, 0.0342, 0.0797], + device='cuda:0'), in_proj_covar=tensor([0.0299, 0.0347, 0.0382, 0.0277, 0.0388, 0.0240, 0.0296, 0.0255], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 13:50:02,188 INFO [train.py:898] (0/4) Epoch 21, batch 750, loss[loss=0.1495, simple_loss=0.2357, pruned_loss=0.03168, over 18302.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2524, pruned_loss=0.03727, over 3484854.11 frames. ], batch size: 49, lr: 5.41e-03, grad_scale: 8.0 +2023-03-09 13:50:09,025 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6729, 3.5128, 3.4132, 3.0227, 3.2805, 2.6942, 2.8344, 3.4964], + device='cuda:0'), covar=tensor([0.0074, 0.0130, 0.0116, 0.0196, 0.0143, 0.0266, 0.0266, 0.0106], + device='cuda:0'), in_proj_covar=tensor([0.0136, 0.0156, 0.0132, 0.0184, 0.0139, 0.0177, 0.0182, 0.0120], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 13:50:15,146 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8894, 4.6312, 4.7853, 3.6938, 3.8849, 3.5649, 2.8601, 2.5286], + device='cuda:0'), covar=tensor([0.0206, 0.0138, 0.0068, 0.0243, 0.0301, 0.0239, 0.0653, 0.0838], + device='cuda:0'), in_proj_covar=tensor([0.0070, 0.0059, 0.0062, 0.0068, 0.0089, 0.0066, 0.0077, 0.0084], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 13:50:54,400 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.921e+02 2.615e+02 3.165e+02 3.785e+02 6.213e+02, threshold=6.329e+02, percent-clipped=0.0 +2023-03-09 13:50:59,647 INFO [train.py:898] (0/4) Epoch 21, batch 800, loss[loss=0.1599, simple_loss=0.2589, pruned_loss=0.03042, over 18480.00 frames. ], tot_loss[loss=0.1642, simple_loss=0.2535, pruned_loss=0.03746, over 3504133.60 frames. ], batch size: 53, lr: 5.41e-03, grad_scale: 8.0 +2023-03-09 13:51:01,383 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-03-09 13:51:56,989 INFO [train.py:898] (0/4) Epoch 21, batch 850, loss[loss=0.1453, simple_loss=0.2327, pruned_loss=0.02899, over 18564.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2523, pruned_loss=0.03682, over 3535271.75 frames. ], batch size: 45, lr: 5.41e-03, grad_scale: 8.0 +2023-03-09 13:52:19,957 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=73550.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:52:33,781 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-09 13:52:50,040 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.633e+02 2.526e+02 3.131e+02 3.569e+02 6.821e+02, threshold=6.262e+02, percent-clipped=1.0 +2023-03-09 13:52:54,607 INFO [train.py:898] (0/4) Epoch 21, batch 900, loss[loss=0.1476, simple_loss=0.2373, pruned_loss=0.02892, over 18289.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.2527, pruned_loss=0.03693, over 3552623.02 frames. ], batch size: 49, lr: 5.41e-03, grad_scale: 8.0 +2023-03-09 13:53:06,078 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=73590.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:53:15,909 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=73598.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:53:40,607 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=73620.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 13:53:53,021 INFO [train.py:898] (0/4) Epoch 21, batch 950, loss[loss=0.148, simple_loss=0.2317, pruned_loss=0.03215, over 18414.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.2532, pruned_loss=0.03718, over 3556540.69 frames. ], batch size: 42, lr: 5.40e-03, grad_scale: 8.0 +2023-03-09 13:54:06,097 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4438, 3.1915, 1.9527, 4.1026, 2.8597, 3.6279, 2.0237, 3.5374], + device='cuda:0'), covar=tensor([0.0644, 0.0966, 0.1677, 0.0610, 0.0896, 0.0399, 0.1637, 0.0539], + device='cuda:0'), in_proj_covar=tensor([0.0215, 0.0228, 0.0190, 0.0284, 0.0193, 0.0264, 0.0204, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 13:54:15,619 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73649.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:54:15,763 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5730, 2.9852, 2.3351, 3.0321, 3.6884, 3.5850, 3.2204, 3.0788], + device='cuda:0'), covar=tensor([0.0195, 0.0238, 0.0688, 0.0298, 0.0180, 0.0143, 0.0292, 0.0307], + device='cuda:0'), in_proj_covar=tensor([0.0140, 0.0133, 0.0162, 0.0155, 0.0130, 0.0115, 0.0151, 0.0155], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 13:54:18,622 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73651.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:54:37,455 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=73668.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 13:54:47,212 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.873e+02 2.598e+02 2.980e+02 3.796e+02 7.725e+02, threshold=5.960e+02, percent-clipped=2.0 +2023-03-09 13:54:48,790 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7219, 3.6722, 4.8652, 4.3341, 3.3021, 2.9880, 4.5138, 5.1409], + device='cuda:0'), covar=tensor([0.0795, 0.1431, 0.0209, 0.0390, 0.0911, 0.1208, 0.0370, 0.0245], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0279, 0.0153, 0.0183, 0.0195, 0.0194, 0.0198, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 13:54:51,742 INFO [train.py:898] (0/4) Epoch 21, batch 1000, loss[loss=0.135, simple_loss=0.2148, pruned_loss=0.02762, over 18406.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.2535, pruned_loss=0.03697, over 3576056.39 frames. ], batch size: 42, lr: 5.40e-03, grad_scale: 8.0 +2023-03-09 13:55:17,727 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3629, 4.2689, 4.4035, 4.1009, 4.1959, 4.2941, 4.5000, 4.3758], + device='cuda:0'), covar=tensor([0.0109, 0.0126, 0.0124, 0.0162, 0.0099, 0.0161, 0.0102, 0.0151], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0069, 0.0074, 0.0092, 0.0075, 0.0104, 0.0086, 0.0086], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 13:55:23,187 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.29 vs. limit=5.0 +2023-03-09 13:55:32,126 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-09 13:55:49,539 INFO [train.py:898] (0/4) Epoch 21, batch 1050, loss[loss=0.1702, simple_loss=0.2654, pruned_loss=0.03753, over 17066.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2536, pruned_loss=0.03675, over 3587947.57 frames. ], batch size: 78, lr: 5.40e-03, grad_scale: 8.0 +2023-03-09 13:55:50,880 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=73732.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:56:06,786 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1712, 5.2824, 5.4066, 5.4319, 5.1225, 6.0217, 5.5852, 5.2284], + device='cuda:0'), covar=tensor([0.1022, 0.0587, 0.0715, 0.0714, 0.1273, 0.0695, 0.0699, 0.1811], + device='cuda:0'), in_proj_covar=tensor([0.0355, 0.0281, 0.0308, 0.0308, 0.0327, 0.0420, 0.0279, 0.0416], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 13:56:14,431 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6656, 3.6369, 2.4213, 4.4857, 3.1413, 4.4546, 2.9142, 4.2344], + device='cuda:0'), covar=tensor([0.0632, 0.0768, 0.1370, 0.0468, 0.0857, 0.0273, 0.0999, 0.0332], + device='cuda:0'), in_proj_covar=tensor([0.0216, 0.0229, 0.0190, 0.0285, 0.0194, 0.0264, 0.0205, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 13:56:43,397 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.657e+02 2.593e+02 3.039e+02 3.732e+02 7.995e+02, threshold=6.078e+02, percent-clipped=2.0 +2023-03-09 13:56:47,972 INFO [train.py:898] (0/4) Epoch 21, batch 1100, loss[loss=0.1599, simple_loss=0.2539, pruned_loss=0.03294, over 15898.00 frames. ], tot_loss[loss=0.1631, simple_loss=0.2524, pruned_loss=0.03687, over 3591540.77 frames. ], batch size: 94, lr: 5.40e-03, grad_scale: 8.0 +2023-03-09 13:57:01,631 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=73793.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 13:57:46,779 INFO [train.py:898] (0/4) Epoch 21, batch 1150, loss[loss=0.1515, simple_loss=0.2388, pruned_loss=0.0321, over 18547.00 frames. ], tot_loss[loss=0.1629, simple_loss=0.2522, pruned_loss=0.03681, over 3579470.91 frames. ], batch size: 49, lr: 5.40e-03, grad_scale: 8.0 +2023-03-09 13:58:40,425 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.636e+02 2.572e+02 3.032e+02 3.666e+02 7.271e+02, threshold=6.063e+02, percent-clipped=3.0 +2023-03-09 13:58:44,859 INFO [train.py:898] (0/4) Epoch 21, batch 1200, loss[loss=0.1597, simple_loss=0.2543, pruned_loss=0.03249, over 18483.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.2525, pruned_loss=0.03705, over 3587239.82 frames. ], batch size: 51, lr: 5.39e-03, grad_scale: 8.0 +2023-03-09 13:58:53,097 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.45 vs. limit=5.0 +2023-03-09 13:59:00,866 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-09 13:59:42,880 INFO [train.py:898] (0/4) Epoch 21, batch 1250, loss[loss=0.1371, simple_loss=0.2218, pruned_loss=0.02621, over 18487.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.2529, pruned_loss=0.03734, over 3585885.45 frames. ], batch size: 44, lr: 5.39e-03, grad_scale: 8.0 +2023-03-09 14:00:00,222 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=73946.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:00:03,835 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=73949.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:00:22,623 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7520, 2.4324, 2.6760, 2.8569, 3.3215, 5.0857, 4.9990, 3.2004], + device='cuda:0'), covar=tensor([0.1790, 0.2290, 0.2903, 0.1703, 0.2345, 0.0173, 0.0293, 0.1018], + device='cuda:0'), in_proj_covar=tensor([0.0299, 0.0347, 0.0382, 0.0278, 0.0388, 0.0240, 0.0296, 0.0255], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 14:00:37,351 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.681e+02 2.538e+02 2.943e+02 3.615e+02 7.778e+02, threshold=5.886e+02, percent-clipped=1.0 +2023-03-09 14:00:41,922 INFO [train.py:898] (0/4) Epoch 21, batch 1300, loss[loss=0.1624, simple_loss=0.2573, pruned_loss=0.0337, over 18290.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.253, pruned_loss=0.03713, over 3584714.29 frames. ], batch size: 54, lr: 5.39e-03, grad_scale: 8.0 +2023-03-09 14:01:00,189 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=73997.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:01:03,804 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-74000.pt +2023-03-09 14:01:10,423 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-09 14:01:45,716 INFO [train.py:898] (0/4) Epoch 21, batch 1350, loss[loss=0.1325, simple_loss=0.2242, pruned_loss=0.02036, over 18244.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2527, pruned_loss=0.03681, over 3591993.21 frames. ], batch size: 47, lr: 5.39e-03, grad_scale: 8.0 +2023-03-09 14:02:19,904 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9993, 3.6447, 5.0224, 2.8402, 4.3022, 2.6203, 3.0973, 1.6853], + device='cuda:0'), covar=tensor([0.1066, 0.0913, 0.0130, 0.0931, 0.0548, 0.2474, 0.2519, 0.2202], + device='cuda:0'), in_proj_covar=tensor([0.0221, 0.0246, 0.0189, 0.0198, 0.0260, 0.0272, 0.0323, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 14:02:22,917 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74064.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:02:39,204 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.953e+02 2.917e+02 3.282e+02 4.135e+02 8.979e+02, threshold=6.564e+02, percent-clipped=10.0 +2023-03-09 14:02:43,779 INFO [train.py:898] (0/4) Epoch 21, batch 1400, loss[loss=0.1696, simple_loss=0.2649, pruned_loss=0.0372, over 18270.00 frames. ], tot_loss[loss=0.1629, simple_loss=0.2524, pruned_loss=0.03674, over 3594227.70 frames. ], batch size: 57, lr: 5.39e-03, grad_scale: 8.0 +2023-03-09 14:02:51,614 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74088.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:03:35,134 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=74125.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:03:41,405 INFO [train.py:898] (0/4) Epoch 21, batch 1450, loss[loss=0.2023, simple_loss=0.2846, pruned_loss=0.05996, over 18355.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2532, pruned_loss=0.03689, over 3603539.47 frames. ], batch size: 56, lr: 5.39e-03, grad_scale: 8.0 +2023-03-09 14:04:35,771 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.991e+02 2.519e+02 2.943e+02 3.724e+02 1.372e+03, threshold=5.886e+02, percent-clipped=4.0 +2023-03-09 14:04:40,826 INFO [train.py:898] (0/4) Epoch 21, batch 1500, loss[loss=0.1509, simple_loss=0.2445, pruned_loss=0.02865, over 18312.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.253, pruned_loss=0.03679, over 3600865.59 frames. ], batch size: 54, lr: 5.38e-03, grad_scale: 8.0 +2023-03-09 14:05:06,285 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8659, 5.2611, 2.8420, 5.0299, 4.9982, 5.2401, 5.0223, 2.7175], + device='cuda:0'), covar=tensor([0.0238, 0.0076, 0.0760, 0.0088, 0.0067, 0.0075, 0.0099, 0.0996], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0081, 0.0095, 0.0095, 0.0084, 0.0075, 0.0084, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 14:05:39,324 INFO [train.py:898] (0/4) Epoch 21, batch 1550, loss[loss=0.1631, simple_loss=0.2493, pruned_loss=0.03842, over 18388.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2522, pruned_loss=0.03652, over 3605021.80 frames. ], batch size: 50, lr: 5.38e-03, grad_scale: 8.0 +2023-03-09 14:05:50,218 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74240.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:05:53,516 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5171, 6.0803, 5.6257, 5.8481, 5.6828, 5.5275, 6.1486, 6.0856], + device='cuda:0'), covar=tensor([0.1099, 0.0643, 0.0451, 0.0709, 0.1279, 0.0659, 0.0562, 0.0656], + device='cuda:0'), in_proj_covar=tensor([0.0609, 0.0524, 0.0379, 0.0547, 0.0737, 0.0542, 0.0739, 0.0566], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 14:05:57,050 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74246.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:05:59,350 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7315, 4.5729, 4.6043, 3.4730, 3.8127, 3.5678, 2.4112, 2.4154], + device='cuda:0'), covar=tensor([0.0229, 0.0138, 0.0067, 0.0279, 0.0324, 0.0212, 0.0804, 0.0889], + device='cuda:0'), in_proj_covar=tensor([0.0069, 0.0058, 0.0061, 0.0067, 0.0087, 0.0065, 0.0076, 0.0083], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 14:06:00,314 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4586, 5.4245, 5.0989, 5.3411, 5.3487, 4.7332, 5.2556, 5.0307], + device='cuda:0'), covar=tensor([0.0431, 0.0425, 0.1365, 0.0789, 0.0548, 0.0469, 0.0467, 0.1071], + device='cuda:0'), in_proj_covar=tensor([0.0502, 0.0565, 0.0706, 0.0436, 0.0451, 0.0511, 0.0546, 0.0688], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 14:06:32,520 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.761e+02 2.609e+02 3.003e+02 3.447e+02 5.872e+02, threshold=6.005e+02, percent-clipped=0.0 +2023-03-09 14:06:37,156 INFO [train.py:898] (0/4) Epoch 21, batch 1600, loss[loss=0.1277, simple_loss=0.2155, pruned_loss=0.01993, over 17647.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2526, pruned_loss=0.03669, over 3608230.22 frames. ], batch size: 39, lr: 5.38e-03, grad_scale: 8.0 +2023-03-09 14:06:46,915 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7854, 5.0320, 4.9671, 5.0605, 4.7776, 5.5843, 5.1675, 4.8774], + device='cuda:0'), covar=tensor([0.1187, 0.0775, 0.0868, 0.0805, 0.1478, 0.0757, 0.0753, 0.1703], + device='cuda:0'), in_proj_covar=tensor([0.0360, 0.0288, 0.0314, 0.0313, 0.0331, 0.0423, 0.0282, 0.0419], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 14:06:49,328 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5499, 3.1569, 4.3177, 3.7569, 2.6764, 4.6592, 4.0145, 2.9804], + device='cuda:0'), covar=tensor([0.0582, 0.1336, 0.0284, 0.0429, 0.1599, 0.0224, 0.0482, 0.0959], + device='cuda:0'), in_proj_covar=tensor([0.0205, 0.0233, 0.0206, 0.0158, 0.0218, 0.0208, 0.0241, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 14:06:53,632 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=74294.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:07:01,864 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=74301.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:07:35,659 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6533, 2.7379, 2.6012, 2.8841, 3.6655, 3.5213, 3.0952, 2.9365], + device='cuda:0'), covar=tensor([0.0197, 0.0300, 0.0551, 0.0384, 0.0184, 0.0150, 0.0399, 0.0387], + device='cuda:0'), in_proj_covar=tensor([0.0140, 0.0133, 0.0162, 0.0156, 0.0128, 0.0116, 0.0153, 0.0154], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 14:07:36,396 INFO [train.py:898] (0/4) Epoch 21, batch 1650, loss[loss=0.1534, simple_loss=0.2536, pruned_loss=0.02661, over 18634.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2528, pruned_loss=0.0366, over 3602905.70 frames. ], batch size: 52, lr: 5.38e-03, grad_scale: 8.0 +2023-03-09 14:08:30,594 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.874e+02 2.715e+02 3.292e+02 3.947e+02 6.741e+02, threshold=6.584e+02, percent-clipped=1.0 +2023-03-09 14:08:35,088 INFO [train.py:898] (0/4) Epoch 21, batch 1700, loss[loss=0.1764, simple_loss=0.263, pruned_loss=0.04494, over 17119.00 frames. ], tot_loss[loss=0.1629, simple_loss=0.2524, pruned_loss=0.03665, over 3595902.84 frames. ], batch size: 78, lr: 5.38e-03, grad_scale: 8.0 +2023-03-09 14:08:44,108 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74388.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:08:59,116 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7447, 5.2346, 5.2081, 5.2341, 4.7961, 5.1560, 4.5870, 5.1275], + device='cuda:0'), covar=tensor([0.0222, 0.0238, 0.0173, 0.0346, 0.0335, 0.0200, 0.0980, 0.0285], + device='cuda:0'), in_proj_covar=tensor([0.0216, 0.0261, 0.0252, 0.0331, 0.0268, 0.0270, 0.0310, 0.0259], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 14:09:20,215 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74420.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:09:33,168 INFO [train.py:898] (0/4) Epoch 21, batch 1750, loss[loss=0.185, simple_loss=0.2747, pruned_loss=0.04771, over 16302.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2533, pruned_loss=0.03681, over 3600438.19 frames. ], batch size: 94, lr: 5.37e-03, grad_scale: 8.0 +2023-03-09 14:09:38,683 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=74436.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:10:18,430 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9617, 5.4313, 2.8261, 5.2226, 5.1406, 5.4058, 5.2004, 2.6524], + device='cuda:0'), covar=tensor([0.0216, 0.0056, 0.0754, 0.0069, 0.0066, 0.0055, 0.0086, 0.1029], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0080, 0.0095, 0.0095, 0.0084, 0.0075, 0.0084, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 14:10:25,800 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.991e+02 2.679e+02 3.149e+02 3.785e+02 6.522e+02, threshold=6.298e+02, percent-clipped=0.0 +2023-03-09 14:10:30,659 INFO [train.py:898] (0/4) Epoch 21, batch 1800, loss[loss=0.157, simple_loss=0.2387, pruned_loss=0.0376, over 18354.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2523, pruned_loss=0.03648, over 3605673.91 frames. ], batch size: 46, lr: 5.37e-03, grad_scale: 8.0 +2023-03-09 14:11:11,665 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.43 vs. limit=2.0 +2023-03-09 14:11:28,025 INFO [train.py:898] (0/4) Epoch 21, batch 1850, loss[loss=0.1678, simple_loss=0.2552, pruned_loss=0.0402, over 18494.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2519, pruned_loss=0.03653, over 3607318.96 frames. ], batch size: 51, lr: 5.37e-03, grad_scale: 8.0 +2023-03-09 14:12:12,560 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.73 vs. limit=2.0 +2023-03-09 14:12:21,974 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.780e+02 2.538e+02 3.055e+02 3.545e+02 5.986e+02, threshold=6.111e+02, percent-clipped=0.0 +2023-03-09 14:12:26,474 INFO [train.py:898] (0/4) Epoch 21, batch 1900, loss[loss=0.1654, simple_loss=0.2549, pruned_loss=0.03792, over 18362.00 frames. ], tot_loss[loss=0.1629, simple_loss=0.252, pruned_loss=0.03686, over 3590011.56 frames. ], batch size: 56, lr: 5.37e-03, grad_scale: 8.0 +2023-03-09 14:12:43,230 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74595.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:12:44,262 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74596.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:13:24,947 INFO [train.py:898] (0/4) Epoch 21, batch 1950, loss[loss=0.2189, simple_loss=0.2971, pruned_loss=0.07033, over 12589.00 frames. ], tot_loss[loss=0.1629, simple_loss=0.2519, pruned_loss=0.0369, over 3579757.67 frames. ], batch size: 131, lr: 5.37e-03, grad_scale: 16.0 +2023-03-09 14:13:40,579 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4848, 3.4522, 3.3929, 3.0544, 3.3523, 2.8418, 2.7562, 3.6022], + device='cuda:0'), covar=tensor([0.0073, 0.0095, 0.0076, 0.0140, 0.0093, 0.0168, 0.0194, 0.0060], + device='cuda:0'), in_proj_covar=tensor([0.0140, 0.0158, 0.0135, 0.0186, 0.0142, 0.0178, 0.0184, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 14:13:56,034 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=74656.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:14:19,405 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.749e+02 2.532e+02 2.978e+02 3.558e+02 7.453e+02, threshold=5.957e+02, percent-clipped=1.0 +2023-03-09 14:14:23,892 INFO [train.py:898] (0/4) Epoch 21, batch 2000, loss[loss=0.1588, simple_loss=0.2534, pruned_loss=0.03208, over 17722.00 frames. ], tot_loss[loss=0.1631, simple_loss=0.2523, pruned_loss=0.03698, over 3582510.16 frames. ], batch size: 70, lr: 5.37e-03, grad_scale: 16.0 +2023-03-09 14:14:38,335 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9676, 3.7343, 4.9636, 4.5182, 3.5438, 3.2152, 4.7998, 5.2923], + device='cuda:0'), covar=tensor([0.0740, 0.1703, 0.0247, 0.0384, 0.0847, 0.1070, 0.0301, 0.0315], + device='cuda:0'), in_proj_covar=tensor([0.0148, 0.0275, 0.0151, 0.0181, 0.0191, 0.0191, 0.0195, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 14:15:09,461 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74720.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:15:21,706 INFO [train.py:898] (0/4) Epoch 21, batch 2050, loss[loss=0.1519, simple_loss=0.2313, pruned_loss=0.03628, over 17729.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2523, pruned_loss=0.03706, over 3582596.30 frames. ], batch size: 39, lr: 5.36e-03, grad_scale: 16.0 +2023-03-09 14:16:05,255 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=74768.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:16:15,321 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.051e+02 2.659e+02 3.081e+02 3.760e+02 7.989e+02, threshold=6.163e+02, percent-clipped=4.0 +2023-03-09 14:16:19,623 INFO [train.py:898] (0/4) Epoch 21, batch 2100, loss[loss=0.1508, simple_loss=0.2309, pruned_loss=0.03529, over 18430.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.2529, pruned_loss=0.03726, over 3573811.65 frames. ], batch size: 43, lr: 5.36e-03, grad_scale: 16.0 +2023-03-09 14:16:55,731 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4825, 6.0523, 5.5444, 5.8032, 5.6377, 5.4442, 6.1039, 6.0306], + device='cuda:0'), covar=tensor([0.1189, 0.0718, 0.0447, 0.0774, 0.1183, 0.0770, 0.0513, 0.0646], + device='cuda:0'), in_proj_covar=tensor([0.0615, 0.0532, 0.0385, 0.0554, 0.0747, 0.0552, 0.0750, 0.0572], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 14:17:17,969 INFO [train.py:898] (0/4) Epoch 21, batch 2150, loss[loss=0.1756, simple_loss=0.2648, pruned_loss=0.04319, over 16374.00 frames. ], tot_loss[loss=0.1639, simple_loss=0.2529, pruned_loss=0.03745, over 3571183.04 frames. ], batch size: 94, lr: 5.36e-03, grad_scale: 16.0 +2023-03-09 14:17:43,757 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-09 14:18:11,058 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.902e+02 2.637e+02 3.175e+02 3.840e+02 7.861e+02, threshold=6.351e+02, percent-clipped=2.0 +2023-03-09 14:18:15,485 INFO [train.py:898] (0/4) Epoch 21, batch 2200, loss[loss=0.1878, simple_loss=0.2776, pruned_loss=0.04897, over 17315.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2521, pruned_loss=0.03711, over 3579894.93 frames. ], batch size: 78, lr: 5.36e-03, grad_scale: 16.0 +2023-03-09 14:18:32,696 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=74896.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:18:47,781 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74909.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:19:13,475 INFO [train.py:898] (0/4) Epoch 21, batch 2250, loss[loss=0.1814, simple_loss=0.2713, pruned_loss=0.04578, over 17926.00 frames. ], tot_loss[loss=0.1622, simple_loss=0.2514, pruned_loss=0.03651, over 3591662.46 frames. ], batch size: 65, lr: 5.36e-03, grad_scale: 16.0 +2023-03-09 14:19:28,438 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=74944.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:19:36,396 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=74951.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:19:36,627 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3341, 3.7853, 2.5026, 3.6763, 4.5964, 2.4442, 3.5004, 3.5469], + device='cuda:0'), covar=tensor([0.0232, 0.1063, 0.1557, 0.0648, 0.0108, 0.1259, 0.0718, 0.0774], + device='cuda:0'), in_proj_covar=tensor([0.0167, 0.0270, 0.0204, 0.0194, 0.0126, 0.0181, 0.0215, 0.0224], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 14:19:51,802 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=74964.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:19:52,289 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.41 vs. limit=5.0 +2023-03-09 14:19:59,239 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=74970.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:20:02,924 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0885, 5.5543, 5.5307, 5.5530, 5.0910, 5.4987, 4.8578, 5.4693], + device='cuda:0'), covar=tensor([0.0216, 0.0256, 0.0189, 0.0330, 0.0347, 0.0207, 0.1051, 0.0270], + device='cuda:0'), in_proj_covar=tensor([0.0218, 0.0262, 0.0255, 0.0331, 0.0271, 0.0270, 0.0311, 0.0261], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 14:20:08,787 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.078e+02 2.763e+02 3.199e+02 3.835e+02 8.109e+02, threshold=6.398e+02, percent-clipped=3.0 +2023-03-09 14:20:12,092 INFO [train.py:898] (0/4) Epoch 21, batch 2300, loss[loss=0.1825, simple_loss=0.2745, pruned_loss=0.04529, over 18254.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2518, pruned_loss=0.03659, over 3588944.45 frames. ], batch size: 60, lr: 5.35e-03, grad_scale: 8.0 +2023-03-09 14:20:22,827 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-09 14:20:45,211 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.70 vs. limit=2.0 +2023-03-09 14:20:45,215 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-09 14:20:48,256 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9585, 5.4048, 2.7014, 5.2278, 5.1651, 5.4336, 5.1780, 2.6900], + device='cuda:0'), covar=tensor([0.0212, 0.0069, 0.0794, 0.0075, 0.0063, 0.0061, 0.0085, 0.0997], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0081, 0.0095, 0.0095, 0.0084, 0.0075, 0.0084, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 14:21:03,792 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=75025.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:21:10,835 INFO [train.py:898] (0/4) Epoch 21, batch 2350, loss[loss=0.1625, simple_loss=0.2658, pruned_loss=0.02966, over 18345.00 frames. ], tot_loss[loss=0.1634, simple_loss=0.2534, pruned_loss=0.03677, over 3596808.74 frames. ], batch size: 55, lr: 5.35e-03, grad_scale: 8.0 +2023-03-09 14:21:37,067 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-09 14:22:04,856 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.756e+02 2.674e+02 3.137e+02 3.722e+02 6.361e+02, threshold=6.274e+02, percent-clipped=0.0 +2023-03-09 14:22:08,960 INFO [train.py:898] (0/4) Epoch 21, batch 2400, loss[loss=0.1562, simple_loss=0.233, pruned_loss=0.03967, over 18426.00 frames. ], tot_loss[loss=0.164, simple_loss=0.254, pruned_loss=0.037, over 3602220.34 frames. ], batch size: 42, lr: 5.35e-03, grad_scale: 8.0 +2023-03-09 14:23:06,749 INFO [train.py:898] (0/4) Epoch 21, batch 2450, loss[loss=0.1711, simple_loss=0.2633, pruned_loss=0.0394, over 18113.00 frames. ], tot_loss[loss=0.1645, simple_loss=0.2543, pruned_loss=0.03734, over 3594726.67 frames. ], batch size: 62, lr: 5.35e-03, grad_scale: 8.0 +2023-03-09 14:24:00,721 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.788e+02 2.611e+02 3.062e+02 3.854e+02 9.257e+02, threshold=6.124e+02, percent-clipped=4.0 +2023-03-09 14:24:03,969 INFO [train.py:898] (0/4) Epoch 21, batch 2500, loss[loss=0.171, simple_loss=0.265, pruned_loss=0.03848, over 17872.00 frames. ], tot_loss[loss=0.1646, simple_loss=0.254, pruned_loss=0.03765, over 3574468.87 frames. ], batch size: 70, lr: 5.35e-03, grad_scale: 8.0 +2023-03-09 14:24:09,779 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-09 14:24:59,555 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=75228.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:25:02,699 INFO [train.py:898] (0/4) Epoch 21, batch 2550, loss[loss=0.1668, simple_loss=0.2538, pruned_loss=0.03987, over 17196.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2543, pruned_loss=0.0378, over 3561397.06 frames. ], batch size: 78, lr: 5.35e-03, grad_scale: 8.0 +2023-03-09 14:25:26,309 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75251.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:25:40,329 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7724, 3.7055, 4.9942, 4.2557, 3.2884, 2.9969, 4.3447, 5.1657], + device='cuda:0'), covar=tensor([0.0735, 0.1331, 0.0165, 0.0401, 0.0898, 0.1104, 0.0380, 0.0205], + device='cuda:0'), in_proj_covar=tensor([0.0146, 0.0270, 0.0149, 0.0179, 0.0188, 0.0187, 0.0192, 0.0194], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 14:25:42,168 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75265.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:25:48,467 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=75270.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:25:57,522 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.566e+02 2.536e+02 3.166e+02 3.977e+02 7.777e+02, threshold=6.331e+02, percent-clipped=2.0 +2023-03-09 14:26:01,084 INFO [train.py:898] (0/4) Epoch 21, batch 2600, loss[loss=0.1546, simple_loss=0.242, pruned_loss=0.03358, over 18526.00 frames. ], tot_loss[loss=0.165, simple_loss=0.2546, pruned_loss=0.03769, over 3563463.83 frames. ], batch size: 49, lr: 5.34e-03, grad_scale: 8.0 +2023-03-09 14:26:11,283 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=75289.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:26:22,470 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=75299.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:26:45,949 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75320.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:26:46,387 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-09 14:26:52,154 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7645, 3.0946, 4.6084, 3.8069, 2.7356, 4.8140, 4.1049, 3.1130], + device='cuda:0'), covar=tensor([0.0488, 0.1372, 0.0211, 0.0475, 0.1641, 0.0197, 0.0444, 0.0976], + device='cuda:0'), in_proj_covar=tensor([0.0209, 0.0238, 0.0210, 0.0162, 0.0222, 0.0211, 0.0244, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 14:26:59,265 INFO [train.py:898] (0/4) Epoch 21, batch 2650, loss[loss=0.1436, simple_loss=0.2244, pruned_loss=0.0314, over 17752.00 frames. ], tot_loss[loss=0.1643, simple_loss=0.2541, pruned_loss=0.03727, over 3568607.42 frames. ], batch size: 39, lr: 5.34e-03, grad_scale: 8.0 +2023-03-09 14:26:59,656 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=75331.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 14:27:53,150 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.804e+02 2.756e+02 3.274e+02 3.890e+02 7.803e+02, threshold=6.547e+02, percent-clipped=3.0 +2023-03-09 14:27:57,176 INFO [train.py:898] (0/4) Epoch 21, batch 2700, loss[loss=0.1487, simple_loss=0.2399, pruned_loss=0.02878, over 18357.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2532, pruned_loss=0.03702, over 3586301.03 frames. ], batch size: 50, lr: 5.34e-03, grad_scale: 8.0 +2023-03-09 14:28:55,176 INFO [train.py:898] (0/4) Epoch 21, batch 2750, loss[loss=0.1699, simple_loss=0.2588, pruned_loss=0.04052, over 17898.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.2531, pruned_loss=0.03679, over 3602652.95 frames. ], batch size: 70, lr: 5.34e-03, grad_scale: 4.0 +2023-03-09 14:29:15,564 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8328, 5.1926, 2.5442, 5.0434, 4.8333, 5.2066, 4.9257, 2.4723], + device='cuda:0'), covar=tensor([0.0246, 0.0065, 0.0858, 0.0085, 0.0095, 0.0075, 0.0111, 0.1133], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0081, 0.0095, 0.0095, 0.0084, 0.0075, 0.0085, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 14:29:50,239 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.664e+02 2.526e+02 3.249e+02 3.895e+02 6.224e+02, threshold=6.498e+02, percent-clipped=0.0 +2023-03-09 14:29:52,449 INFO [train.py:898] (0/4) Epoch 21, batch 2800, loss[loss=0.1594, simple_loss=0.2544, pruned_loss=0.0322, over 18483.00 frames. ], tot_loss[loss=0.1631, simple_loss=0.2528, pruned_loss=0.03668, over 3596919.93 frames. ], batch size: 53, lr: 5.34e-03, grad_scale: 8.0 +2023-03-09 14:30:50,507 INFO [train.py:898] (0/4) Epoch 21, batch 2850, loss[loss=0.1526, simple_loss=0.2331, pruned_loss=0.03602, over 18366.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2532, pruned_loss=0.03691, over 3596879.03 frames. ], batch size: 42, lr: 5.34e-03, grad_scale: 8.0 +2023-03-09 14:30:52,483 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7912, 3.4101, 4.5184, 3.9074, 2.5264, 4.8335, 4.0441, 3.0478], + device='cuda:0'), covar=tensor([0.0511, 0.1165, 0.0243, 0.0444, 0.1738, 0.0210, 0.0530, 0.0933], + device='cuda:0'), in_proj_covar=tensor([0.0211, 0.0239, 0.0212, 0.0162, 0.0223, 0.0212, 0.0246, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 14:30:59,529 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-09 14:31:01,942 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6438, 2.7322, 2.6140, 2.9370, 3.6946, 3.5152, 3.1717, 2.9158], + device='cuda:0'), covar=tensor([0.0193, 0.0329, 0.0613, 0.0356, 0.0235, 0.0199, 0.0358, 0.0376], + device='cuda:0'), in_proj_covar=tensor([0.0139, 0.0134, 0.0162, 0.0158, 0.0129, 0.0117, 0.0153, 0.0154], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 14:31:12,946 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5674, 5.3691, 5.8281, 5.7821, 5.5383, 6.3231, 5.9177, 5.5942], + device='cuda:0'), covar=tensor([0.1045, 0.0586, 0.0573, 0.0628, 0.1247, 0.0657, 0.0566, 0.1581], + device='cuda:0'), in_proj_covar=tensor([0.0355, 0.0286, 0.0310, 0.0310, 0.0330, 0.0420, 0.0281, 0.0417], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0002, 0.0004], + device='cuda:0') +2023-03-09 14:31:31,010 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75565.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:31:46,371 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.918e+02 2.851e+02 3.414e+02 4.346e+02 8.993e+02, threshold=6.828e+02, percent-clipped=4.0 +2023-03-09 14:31:46,960 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.48 vs. limit=2.0 +2023-03-09 14:31:48,602 INFO [train.py:898] (0/4) Epoch 21, batch 2900, loss[loss=0.1794, simple_loss=0.2673, pruned_loss=0.04572, over 18302.00 frames. ], tot_loss[loss=0.1629, simple_loss=0.2524, pruned_loss=0.03664, over 3600659.50 frames. ], batch size: 57, lr: 5.33e-03, grad_scale: 8.0 +2023-03-09 14:31:48,952 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8205, 3.7159, 5.1122, 4.4974, 3.3378, 3.1688, 4.6551, 5.2812], + device='cuda:0'), covar=tensor([0.0830, 0.1544, 0.0184, 0.0378, 0.0983, 0.1207, 0.0372, 0.0277], + device='cuda:0'), in_proj_covar=tensor([0.0148, 0.0275, 0.0152, 0.0181, 0.0191, 0.0190, 0.0195, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 14:31:52,546 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75584.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:32:27,058 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=75613.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:32:34,981 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75620.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:32:41,938 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=75626.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 14:32:47,478 INFO [train.py:898] (0/4) Epoch 21, batch 2950, loss[loss=0.1675, simple_loss=0.2599, pruned_loss=0.03752, over 16192.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2518, pruned_loss=0.03658, over 3598758.98 frames. ], batch size: 94, lr: 5.33e-03, grad_scale: 8.0 +2023-03-09 14:33:31,426 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=75668.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:33:43,525 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.752e+02 2.513e+02 3.131e+02 3.690e+02 6.570e+02, threshold=6.263e+02, percent-clipped=0.0 +2023-03-09 14:33:45,503 INFO [train.py:898] (0/4) Epoch 21, batch 3000, loss[loss=0.1745, simple_loss=0.2649, pruned_loss=0.04203, over 17166.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2518, pruned_loss=0.0366, over 3600623.94 frames. ], batch size: 78, lr: 5.33e-03, grad_scale: 8.0 +2023-03-09 14:33:45,504 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 14:33:57,966 INFO [train.py:932] (0/4) Epoch 21, validation: loss=0.1498, simple_loss=0.2495, pruned_loss=0.02501, over 944034.00 frames. +2023-03-09 14:33:57,967 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 14:34:55,550 INFO [train.py:898] (0/4) Epoch 21, batch 3050, loss[loss=0.1769, simple_loss=0.2641, pruned_loss=0.04486, over 18577.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2517, pruned_loss=0.03672, over 3601432.97 frames. ], batch size: 54, lr: 5.33e-03, grad_scale: 8.0 +2023-03-09 14:35:05,655 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5514, 3.5163, 4.8502, 4.0784, 3.1995, 2.8085, 4.2478, 5.0375], + device='cuda:0'), covar=tensor([0.0823, 0.1432, 0.0170, 0.0472, 0.0961, 0.1264, 0.0399, 0.0203], + device='cuda:0'), in_proj_covar=tensor([0.0147, 0.0273, 0.0152, 0.0181, 0.0190, 0.0190, 0.0194, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 14:35:18,074 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0301, 5.4207, 2.9240, 5.2301, 5.0865, 5.3738, 5.2217, 2.8038], + device='cuda:0'), covar=tensor([0.0203, 0.0059, 0.0727, 0.0073, 0.0074, 0.0074, 0.0085, 0.0997], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0080, 0.0095, 0.0095, 0.0085, 0.0075, 0.0085, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 14:35:19,735 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1146, 5.0830, 4.8073, 5.0343, 5.0167, 4.4623, 4.9430, 4.7455], + device='cuda:0'), covar=tensor([0.0453, 0.0450, 0.1172, 0.0676, 0.0520, 0.0442, 0.0410, 0.1024], + device='cuda:0'), in_proj_covar=tensor([0.0500, 0.0561, 0.0706, 0.0440, 0.0453, 0.0509, 0.0542, 0.0682], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 14:35:51,757 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.562e+02 2.753e+02 3.195e+02 3.749e+02 6.792e+02, threshold=6.390e+02, percent-clipped=2.0 +2023-03-09 14:35:54,447 INFO [train.py:898] (0/4) Epoch 21, batch 3100, loss[loss=0.2155, simple_loss=0.2951, pruned_loss=0.06798, over 12282.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2522, pruned_loss=0.03683, over 3595365.98 frames. ], batch size: 129, lr: 5.33e-03, grad_scale: 8.0 +2023-03-09 14:35:58,991 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4308, 5.4008, 5.0848, 5.3343, 5.3694, 4.7810, 5.2653, 4.9913], + device='cuda:0'), covar=tensor([0.0441, 0.0400, 0.1272, 0.0745, 0.0516, 0.0409, 0.0390, 0.1068], + device='cuda:0'), in_proj_covar=tensor([0.0501, 0.0563, 0.0707, 0.0440, 0.0452, 0.0510, 0.0544, 0.0683], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 14:36:52,846 INFO [train.py:898] (0/4) Epoch 21, batch 3150, loss[loss=0.1759, simple_loss=0.2676, pruned_loss=0.04212, over 18330.00 frames. ], tot_loss[loss=0.1623, simple_loss=0.2518, pruned_loss=0.03642, over 3607050.93 frames. ], batch size: 54, lr: 5.32e-03, grad_scale: 8.0 +2023-03-09 14:37:32,266 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-09 14:37:49,362 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.108e+02 2.640e+02 3.076e+02 3.933e+02 6.282e+02, threshold=6.152e+02, percent-clipped=0.0 +2023-03-09 14:37:51,672 INFO [train.py:898] (0/4) Epoch 21, batch 3200, loss[loss=0.158, simple_loss=0.2518, pruned_loss=0.03207, over 18290.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.2516, pruned_loss=0.03657, over 3596094.54 frames. ], batch size: 60, lr: 5.32e-03, grad_scale: 8.0 +2023-03-09 14:37:55,405 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75884.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:37:58,965 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1091, 5.3934, 2.6760, 5.2638, 5.1216, 5.3335, 5.1802, 2.5862], + device='cuda:0'), covar=tensor([0.0189, 0.0084, 0.0818, 0.0075, 0.0080, 0.0113, 0.0100, 0.1117], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0081, 0.0096, 0.0095, 0.0085, 0.0076, 0.0085, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 14:38:43,101 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6128, 2.3824, 2.6578, 2.6676, 3.1671, 4.8455, 4.7426, 3.4938], + device='cuda:0'), covar=tensor([0.1871, 0.2373, 0.2958, 0.1858, 0.2450, 0.0246, 0.0354, 0.0913], + device='cuda:0'), in_proj_covar=tensor([0.0301, 0.0346, 0.0383, 0.0277, 0.0389, 0.0242, 0.0296, 0.0256], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 14:38:45,210 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=75926.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 14:38:50,635 INFO [train.py:898] (0/4) Epoch 21, batch 3250, loss[loss=0.1735, simple_loss=0.268, pruned_loss=0.03949, over 18502.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2526, pruned_loss=0.03694, over 3591760.32 frames. ], batch size: 51, lr: 5.32e-03, grad_scale: 8.0 +2023-03-09 14:38:51,966 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=75932.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:39:12,201 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2062, 5.6055, 2.6710, 5.4062, 5.3046, 5.5964, 5.4299, 2.8381], + device='cuda:0'), covar=tensor([0.0175, 0.0059, 0.0800, 0.0060, 0.0068, 0.0065, 0.0078, 0.0960], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0081, 0.0096, 0.0096, 0.0086, 0.0076, 0.0086, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 14:39:41,431 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=75974.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:39:46,930 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.093e+02 2.622e+02 3.085e+02 3.781e+02 1.173e+03, threshold=6.171e+02, percent-clipped=7.0 +2023-03-09 14:39:49,198 INFO [train.py:898] (0/4) Epoch 21, batch 3300, loss[loss=0.1359, simple_loss=0.22, pruned_loss=0.02591, over 18502.00 frames. ], tot_loss[loss=0.1629, simple_loss=0.2521, pruned_loss=0.03688, over 3593127.67 frames. ], batch size: 44, lr: 5.32e-03, grad_scale: 8.0 +2023-03-09 14:40:12,047 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-76000.pt +2023-03-09 14:40:52,535 INFO [train.py:898] (0/4) Epoch 21, batch 3350, loss[loss=0.1634, simple_loss=0.2581, pruned_loss=0.03435, over 18482.00 frames. ], tot_loss[loss=0.1638, simple_loss=0.2531, pruned_loss=0.0372, over 3584321.30 frames. ], batch size: 53, lr: 5.32e-03, grad_scale: 8.0 +2023-03-09 14:41:26,077 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.73 vs. limit=5.0 +2023-03-09 14:41:48,801 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.784e+02 2.702e+02 3.276e+02 4.045e+02 1.355e+03, threshold=6.552e+02, percent-clipped=6.0 +2023-03-09 14:41:51,074 INFO [train.py:898] (0/4) Epoch 21, batch 3400, loss[loss=0.1556, simple_loss=0.2332, pruned_loss=0.03901, over 17696.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2526, pruned_loss=0.03719, over 3582961.73 frames. ], batch size: 39, lr: 5.32e-03, grad_scale: 8.0 +2023-03-09 14:42:02,122 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.89 vs. limit=5.0 +2023-03-09 14:42:06,212 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5449, 2.3244, 2.4772, 2.5771, 2.9642, 4.4965, 4.3191, 3.0963], + device='cuda:0'), covar=tensor([0.1819, 0.2426, 0.2815, 0.1941, 0.2540, 0.0249, 0.0424, 0.1008], + device='cuda:0'), in_proj_covar=tensor([0.0303, 0.0349, 0.0386, 0.0280, 0.0392, 0.0244, 0.0297, 0.0257], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 14:42:49,331 INFO [train.py:898] (0/4) Epoch 21, batch 3450, loss[loss=0.1822, simple_loss=0.2718, pruned_loss=0.0463, over 18299.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2526, pruned_loss=0.03721, over 3576855.54 frames. ], batch size: 57, lr: 5.31e-03, grad_scale: 8.0 +2023-03-09 14:43:45,619 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.955e+02 2.600e+02 3.134e+02 3.743e+02 8.321e+02, threshold=6.269e+02, percent-clipped=3.0 +2023-03-09 14:43:47,843 INFO [train.py:898] (0/4) Epoch 21, batch 3500, loss[loss=0.1634, simple_loss=0.2584, pruned_loss=0.03414, over 18304.00 frames. ], tot_loss[loss=0.1637, simple_loss=0.253, pruned_loss=0.03722, over 3559410.99 frames. ], batch size: 54, lr: 5.31e-03, grad_scale: 8.0 +2023-03-09 14:44:39,750 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4326, 3.4135, 3.2885, 3.0199, 3.2300, 2.7053, 2.7786, 3.4681], + device='cuda:0'), covar=tensor([0.0074, 0.0092, 0.0100, 0.0139, 0.0103, 0.0175, 0.0190, 0.0072], + device='cuda:0'), in_proj_covar=tensor([0.0141, 0.0158, 0.0135, 0.0187, 0.0142, 0.0178, 0.0184, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 14:44:43,827 INFO [train.py:898] (0/4) Epoch 21, batch 3550, loss[loss=0.1676, simple_loss=0.2511, pruned_loss=0.04204, over 17734.00 frames. ], tot_loss[loss=0.1635, simple_loss=0.2532, pruned_loss=0.03694, over 3579613.92 frames. ], batch size: 39, lr: 5.31e-03, grad_scale: 4.0 +2023-03-09 14:45:24,484 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76269.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 14:45:36,125 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.098e+02 2.775e+02 3.080e+02 3.770e+02 8.534e+02, threshold=6.159e+02, percent-clipped=2.0 +2023-03-09 14:45:37,246 INFO [train.py:898] (0/4) Epoch 21, batch 3600, loss[loss=0.1667, simple_loss=0.2611, pruned_loss=0.0361, over 16126.00 frames. ], tot_loss[loss=0.1634, simple_loss=0.2533, pruned_loss=0.03676, over 3584760.52 frames. ], batch size: 94, lr: 5.31e-03, grad_scale: 8.0 +2023-03-09 14:45:51,589 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76294.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:46:13,100 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-21.pt +2023-03-09 14:46:42,826 INFO [train.py:898] (0/4) Epoch 22, batch 0, loss[loss=0.1717, simple_loss=0.2605, pruned_loss=0.04141, over 18101.00 frames. ], tot_loss[loss=0.1717, simple_loss=0.2605, pruned_loss=0.04141, over 18101.00 frames. ], batch size: 62, lr: 5.18e-03, grad_scale: 8.0 +2023-03-09 14:46:42,828 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 14:46:54,616 INFO [train.py:932] (0/4) Epoch 22, validation: loss=0.1504, simple_loss=0.25, pruned_loss=0.02541, over 944034.00 frames. +2023-03-09 14:46:54,617 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 14:46:57,459 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-09 14:47:04,758 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76323.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 14:47:13,358 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76330.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 14:47:39,530 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.39 vs. limit=5.0 +2023-03-09 14:47:42,759 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76355.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:47:53,546 INFO [train.py:898] (0/4) Epoch 22, batch 50, loss[loss=0.159, simple_loss=0.2416, pruned_loss=0.03824, over 17732.00 frames. ], tot_loss[loss=0.1644, simple_loss=0.2544, pruned_loss=0.03722, over 802849.32 frames. ], batch size: 39, lr: 5.18e-03, grad_scale: 8.0 +2023-03-09 14:48:11,539 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.962e+02 2.572e+02 3.075e+02 3.753e+02 8.462e+02, threshold=6.150e+02, percent-clipped=5.0 +2023-03-09 14:48:16,371 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76384.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 14:48:52,273 INFO [train.py:898] (0/4) Epoch 22, batch 100, loss[loss=0.1594, simple_loss=0.2496, pruned_loss=0.03461, over 16330.00 frames. ], tot_loss[loss=0.1633, simple_loss=0.253, pruned_loss=0.03682, over 1420717.14 frames. ], batch size: 94, lr: 5.18e-03, grad_scale: 8.0 +2023-03-09 14:49:26,594 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9053, 5.0218, 5.0069, 4.7143, 4.7187, 4.7721, 5.1137, 5.0465], + device='cuda:0'), covar=tensor([0.0073, 0.0072, 0.0063, 0.0123, 0.0071, 0.0144, 0.0091, 0.0124], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0068, 0.0073, 0.0091, 0.0073, 0.0102, 0.0085, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 14:49:49,352 INFO [train.py:898] (0/4) Epoch 22, batch 150, loss[loss=0.1741, simple_loss=0.2639, pruned_loss=0.04215, over 18608.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2524, pruned_loss=0.03679, over 1906315.08 frames. ], batch size: 52, lr: 5.18e-03, grad_scale: 8.0 +2023-03-09 14:50:05,945 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.617e+02 2.647e+02 3.251e+02 3.870e+02 7.271e+02, threshold=6.503e+02, percent-clipped=3.0 +2023-03-09 14:50:46,566 INFO [train.py:898] (0/4) Epoch 22, batch 200, loss[loss=0.1678, simple_loss=0.2672, pruned_loss=0.03422, over 18405.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2523, pruned_loss=0.03706, over 2288005.80 frames. ], batch size: 52, lr: 5.18e-03, grad_scale: 8.0 +2023-03-09 14:51:45,316 INFO [train.py:898] (0/4) Epoch 22, batch 250, loss[loss=0.1644, simple_loss=0.2553, pruned_loss=0.03679, over 18451.00 frames. ], tot_loss[loss=0.1634, simple_loss=0.2531, pruned_loss=0.03685, over 2579950.14 frames. ], batch size: 59, lr: 5.18e-03, grad_scale: 8.0 +2023-03-09 14:52:02,091 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.820e+02 2.535e+02 2.870e+02 3.429e+02 6.205e+02, threshold=5.739e+02, percent-clipped=0.0 +2023-03-09 14:52:44,277 INFO [train.py:898] (0/4) Epoch 22, batch 300, loss[loss=0.1784, simple_loss=0.2754, pruned_loss=0.0407, over 18487.00 frames. ], tot_loss[loss=0.1622, simple_loss=0.2518, pruned_loss=0.03624, over 2806081.92 frames. ], batch size: 59, lr: 5.17e-03, grad_scale: 8.0 +2023-03-09 14:52:55,643 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=76625.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 14:53:25,040 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=76650.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:53:42,870 INFO [train.py:898] (0/4) Epoch 22, batch 350, loss[loss=0.173, simple_loss=0.2667, pruned_loss=0.03967, over 17066.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.252, pruned_loss=0.03663, over 2977800.85 frames. ], batch size: 78, lr: 5.17e-03, grad_scale: 8.0 +2023-03-09 14:53:58,855 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=76679.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 14:53:59,706 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.678e+02 2.562e+02 3.001e+02 3.817e+02 1.277e+03, threshold=6.002e+02, percent-clipped=2.0 +2023-03-09 14:54:41,498 INFO [train.py:898] (0/4) Epoch 22, batch 400, loss[loss=0.1927, simple_loss=0.2831, pruned_loss=0.05118, over 18307.00 frames. ], tot_loss[loss=0.1629, simple_loss=0.2526, pruned_loss=0.03667, over 3113867.54 frames. ], batch size: 57, lr: 5.17e-03, grad_scale: 8.0 +2023-03-09 14:55:28,854 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-09 14:55:30,678 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-09 14:55:40,008 INFO [train.py:898] (0/4) Epoch 22, batch 450, loss[loss=0.1659, simple_loss=0.2661, pruned_loss=0.03283, over 18380.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2533, pruned_loss=0.03693, over 3221695.47 frames. ], batch size: 55, lr: 5.17e-03, grad_scale: 8.0 +2023-03-09 14:55:41,893 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-09 14:55:57,343 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.685e+02 2.597e+02 2.928e+02 3.376e+02 5.957e+02, threshold=5.857e+02, percent-clipped=0.0 +2023-03-09 14:56:28,261 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7096, 5.2889, 5.2255, 5.2959, 4.7368, 5.1651, 4.5680, 5.1347], + device='cuda:0'), covar=tensor([0.0244, 0.0271, 0.0204, 0.0396, 0.0420, 0.0230, 0.1129, 0.0319], + device='cuda:0'), in_proj_covar=tensor([0.0218, 0.0263, 0.0258, 0.0333, 0.0273, 0.0270, 0.0314, 0.0264], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 14:56:38,196 INFO [train.py:898] (0/4) Epoch 22, batch 500, loss[loss=0.1551, simple_loss=0.2431, pruned_loss=0.03359, over 18363.00 frames. ], tot_loss[loss=0.1629, simple_loss=0.2527, pruned_loss=0.03657, over 3309433.33 frames. ], batch size: 46, lr: 5.17e-03, grad_scale: 8.0 +2023-03-09 14:56:42,559 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8589, 3.5199, 4.7755, 2.8359, 4.2568, 2.5045, 2.9387, 1.8435], + device='cuda:0'), covar=tensor([0.1149, 0.0935, 0.0229, 0.0928, 0.0550, 0.2770, 0.2555, 0.2161], + device='cuda:0'), in_proj_covar=tensor([0.0219, 0.0246, 0.0192, 0.0198, 0.0258, 0.0272, 0.0322, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 14:57:06,167 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=76839.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:57:12,996 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9433, 3.7364, 5.0811, 4.6873, 3.4349, 3.1753, 4.6133, 5.3885], + device='cuda:0'), covar=tensor([0.0790, 0.1477, 0.0194, 0.0329, 0.0919, 0.1132, 0.0352, 0.0263], + device='cuda:0'), in_proj_covar=tensor([0.0149, 0.0275, 0.0155, 0.0182, 0.0191, 0.0190, 0.0196, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0002, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 14:57:36,634 INFO [train.py:898] (0/4) Epoch 22, batch 550, loss[loss=0.1675, simple_loss=0.2663, pruned_loss=0.03439, over 18577.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2525, pruned_loss=0.0363, over 3383043.51 frames. ], batch size: 54, lr: 5.17e-03, grad_scale: 8.0 +2023-03-09 14:57:53,021 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3562, 5.9103, 5.4931, 5.6885, 5.4497, 5.3428, 5.9196, 5.8683], + device='cuda:0'), covar=tensor([0.1112, 0.0733, 0.0502, 0.0705, 0.1415, 0.0644, 0.0597, 0.0748], + device='cuda:0'), in_proj_covar=tensor([0.0604, 0.0527, 0.0380, 0.0546, 0.0733, 0.0545, 0.0746, 0.0565], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0005, 0.0004], + device='cuda:0') +2023-03-09 14:57:53,941 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.970e+02 2.654e+02 3.120e+02 3.965e+02 8.304e+02, threshold=6.239e+02, percent-clipped=1.0 +2023-03-09 14:58:16,977 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=76900.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:58:34,771 INFO [train.py:898] (0/4) Epoch 22, batch 600, loss[loss=0.1597, simple_loss=0.2382, pruned_loss=0.04056, over 18265.00 frames. ], tot_loss[loss=0.1616, simple_loss=0.2515, pruned_loss=0.03585, over 3434479.29 frames. ], batch size: 45, lr: 5.16e-03, grad_scale: 8.0 +2023-03-09 14:58:47,279 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=76925.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 14:59:14,887 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=76950.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 14:59:32,245 INFO [train.py:898] (0/4) Epoch 22, batch 650, loss[loss=0.1929, simple_loss=0.2796, pruned_loss=0.05309, over 18359.00 frames. ], tot_loss[loss=0.1619, simple_loss=0.2518, pruned_loss=0.03598, over 3471235.81 frames. ], batch size: 55, lr: 5.16e-03, grad_scale: 8.0 +2023-03-09 14:59:42,054 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=76973.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 14:59:49,476 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=76979.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 14:59:50,128 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.663e+02 2.460e+02 2.912e+02 3.515e+02 5.898e+02, threshold=5.824e+02, percent-clipped=0.0 +2023-03-09 15:00:10,804 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=76998.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:00:30,420 INFO [train.py:898] (0/4) Epoch 22, batch 700, loss[loss=0.1612, simple_loss=0.2455, pruned_loss=0.03847, over 18283.00 frames. ], tot_loss[loss=0.1615, simple_loss=0.2513, pruned_loss=0.03586, over 3510516.45 frames. ], batch size: 47, lr: 5.16e-03, grad_scale: 8.0 +2023-03-09 15:00:31,815 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4554, 5.4029, 5.0412, 5.3930, 5.3765, 4.7129, 5.2302, 5.0308], + device='cuda:0'), covar=tensor([0.0420, 0.0468, 0.1406, 0.0715, 0.0561, 0.0465, 0.0466, 0.1025], + device='cuda:0'), in_proj_covar=tensor([0.0498, 0.0559, 0.0700, 0.0436, 0.0448, 0.0508, 0.0542, 0.0672], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 15:00:42,795 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6581, 2.3611, 2.5504, 2.5951, 3.2459, 4.8922, 4.7357, 3.4086], + device='cuda:0'), covar=tensor([0.1880, 0.2418, 0.3123, 0.1940, 0.2368, 0.0206, 0.0358, 0.0943], + device='cuda:0'), in_proj_covar=tensor([0.0305, 0.0349, 0.0387, 0.0281, 0.0393, 0.0245, 0.0298, 0.0257], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 15:00:43,734 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9908, 5.0520, 5.0516, 4.7426, 4.8063, 4.8634, 5.1400, 5.1509], + device='cuda:0'), covar=tensor([0.0078, 0.0061, 0.0059, 0.0111, 0.0061, 0.0141, 0.0080, 0.0088], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0069, 0.0074, 0.0093, 0.0075, 0.0104, 0.0086, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 15:00:45,314 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=77027.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 15:01:29,146 INFO [train.py:898] (0/4) Epoch 22, batch 750, loss[loss=0.1653, simple_loss=0.2643, pruned_loss=0.03308, over 16961.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.2523, pruned_loss=0.03631, over 3518916.17 frames. ], batch size: 78, lr: 5.16e-03, grad_scale: 8.0 +2023-03-09 15:01:47,411 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.651e+02 2.728e+02 3.361e+02 4.039e+02 1.393e+03, threshold=6.722e+02, percent-clipped=6.0 +2023-03-09 15:02:05,347 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77095.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:02:27,378 INFO [train.py:898] (0/4) Epoch 22, batch 800, loss[loss=0.1902, simple_loss=0.2757, pruned_loss=0.05239, over 16224.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2524, pruned_loss=0.0364, over 3534055.92 frames. ], batch size: 94, lr: 5.16e-03, grad_scale: 8.0 +2023-03-09 15:02:46,074 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4420, 2.7294, 2.4298, 2.7994, 3.5369, 3.3577, 3.0418, 2.8625], + device='cuda:0'), covar=tensor([0.0222, 0.0262, 0.0636, 0.0388, 0.0188, 0.0206, 0.0377, 0.0370], + device='cuda:0'), in_proj_covar=tensor([0.0141, 0.0135, 0.0166, 0.0160, 0.0131, 0.0119, 0.0155, 0.0157], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 15:02:51,176 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3319, 3.6653, 2.3342, 3.5385, 4.5847, 2.5264, 3.5273, 3.6438], + device='cuda:0'), covar=tensor([0.0250, 0.1248, 0.1729, 0.0755, 0.0116, 0.1228, 0.0703, 0.0726], + device='cuda:0'), in_proj_covar=tensor([0.0169, 0.0274, 0.0206, 0.0196, 0.0127, 0.0184, 0.0216, 0.0223], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 15:03:16,287 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77156.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 15:03:25,833 INFO [train.py:898] (0/4) Epoch 22, batch 850, loss[loss=0.1482, simple_loss=0.2394, pruned_loss=0.02853, over 18293.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2521, pruned_loss=0.03644, over 3551818.58 frames. ], batch size: 49, lr: 5.16e-03, grad_scale: 8.0 +2023-03-09 15:03:43,940 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.868e+02 2.628e+02 3.238e+02 3.999e+02 6.972e+02, threshold=6.476e+02, percent-clipped=1.0 +2023-03-09 15:04:02,037 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77195.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:04:20,628 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77211.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:04:24,960 INFO [train.py:898] (0/4) Epoch 22, batch 900, loss[loss=0.1598, simple_loss=0.2576, pruned_loss=0.03103, over 18351.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2524, pruned_loss=0.03677, over 3541038.17 frames. ], batch size: 55, lr: 5.15e-03, grad_scale: 8.0 +2023-03-09 15:05:23,899 INFO [train.py:898] (0/4) Epoch 22, batch 950, loss[loss=0.1574, simple_loss=0.2557, pruned_loss=0.02959, over 18314.00 frames. ], tot_loss[loss=0.1619, simple_loss=0.2515, pruned_loss=0.03612, over 3566249.55 frames. ], batch size: 54, lr: 5.15e-03, grad_scale: 8.0 +2023-03-09 15:05:32,082 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77272.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:05:33,125 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77273.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:05:37,404 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77277.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:05:41,083 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.729e+02 2.727e+02 3.251e+02 3.771e+02 1.514e+03, threshold=6.501e+02, percent-clipped=4.0 +2023-03-09 15:06:22,532 INFO [train.py:898] (0/4) Epoch 22, batch 1000, loss[loss=0.1597, simple_loss=0.2527, pruned_loss=0.0334, over 18381.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2529, pruned_loss=0.03651, over 3571454.40 frames. ], batch size: 46, lr: 5.15e-03, grad_scale: 8.0 +2023-03-09 15:06:30,056 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-09 15:06:45,408 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77334.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:06:50,027 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77338.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:06:51,185 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5324, 5.0351, 5.0002, 5.0194, 4.5803, 4.9525, 4.4136, 4.9270], + device='cuda:0'), covar=tensor([0.0275, 0.0292, 0.0226, 0.0461, 0.0416, 0.0248, 0.1067, 0.0330], + device='cuda:0'), in_proj_covar=tensor([0.0218, 0.0263, 0.0257, 0.0334, 0.0273, 0.0270, 0.0313, 0.0264], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 15:06:55,127 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1965, 3.1333, 2.1071, 3.8089, 2.6437, 3.5528, 2.2954, 3.3644], + device='cuda:0'), covar=tensor([0.0642, 0.0837, 0.1333, 0.0539, 0.0809, 0.0328, 0.1208, 0.0428], + device='cuda:0'), in_proj_covar=tensor([0.0217, 0.0228, 0.0192, 0.0288, 0.0194, 0.0265, 0.0204, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 15:07:01,438 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.82 vs. limit=5.0 +2023-03-09 15:07:09,431 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5736, 5.5454, 5.1639, 5.5060, 5.5053, 4.9574, 5.3574, 5.1251], + device='cuda:0'), covar=tensor([0.0386, 0.0403, 0.1282, 0.0653, 0.0503, 0.0380, 0.0409, 0.1066], + device='cuda:0'), in_proj_covar=tensor([0.0493, 0.0550, 0.0693, 0.0434, 0.0443, 0.0503, 0.0538, 0.0665], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 15:07:21,344 INFO [train.py:898] (0/4) Epoch 22, batch 1050, loss[loss=0.1551, simple_loss=0.2488, pruned_loss=0.03071, over 18345.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2529, pruned_loss=0.03652, over 3578533.25 frames. ], batch size: 55, lr: 5.15e-03, grad_scale: 8.0 +2023-03-09 15:07:38,060 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.901e+02 2.608e+02 3.171e+02 4.041e+02 8.045e+02, threshold=6.343e+02, percent-clipped=4.0 +2023-03-09 15:08:19,733 INFO [train.py:898] (0/4) Epoch 22, batch 1100, loss[loss=0.1688, simple_loss=0.2643, pruned_loss=0.03668, over 18351.00 frames. ], tot_loss[loss=0.1632, simple_loss=0.2531, pruned_loss=0.03664, over 3584185.47 frames. ], batch size: 55, lr: 5.15e-03, grad_scale: 8.0 +2023-03-09 15:08:20,102 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7281, 3.6070, 2.3586, 4.5022, 3.2407, 4.3886, 2.6942, 4.1350], + device='cuda:0'), covar=tensor([0.0629, 0.0809, 0.1442, 0.0512, 0.0794, 0.0307, 0.1140, 0.0383], + device='cuda:0'), in_proj_covar=tensor([0.0215, 0.0227, 0.0191, 0.0286, 0.0192, 0.0263, 0.0202, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 15:09:02,477 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77451.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 15:09:18,512 INFO [train.py:898] (0/4) Epoch 22, batch 1150, loss[loss=0.1663, simple_loss=0.2597, pruned_loss=0.03648, over 17781.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2523, pruned_loss=0.03647, over 3595606.95 frames. ], batch size: 70, lr: 5.15e-03, grad_scale: 8.0 +2023-03-09 15:09:35,428 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.520e+02 2.405e+02 2.730e+02 3.168e+02 5.091e+02, threshold=5.460e+02, percent-clipped=0.0 +2023-03-09 15:09:52,922 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77495.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:10:16,605 INFO [train.py:898] (0/4) Epoch 22, batch 1200, loss[loss=0.1843, simple_loss=0.2724, pruned_loss=0.04805, over 18305.00 frames. ], tot_loss[loss=0.1629, simple_loss=0.2528, pruned_loss=0.03649, over 3609353.89 frames. ], batch size: 57, lr: 5.14e-03, grad_scale: 8.0 +2023-03-09 15:10:42,506 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-09 15:10:48,602 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=77543.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:11:15,448 INFO [train.py:898] (0/4) Epoch 22, batch 1250, loss[loss=0.1722, simple_loss=0.2531, pruned_loss=0.04567, over 18356.00 frames. ], tot_loss[loss=0.1623, simple_loss=0.2523, pruned_loss=0.03617, over 3603039.04 frames. ], batch size: 46, lr: 5.14e-03, grad_scale: 8.0 +2023-03-09 15:11:18,513 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77567.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:11:32,785 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.900e+02 2.671e+02 3.249e+02 3.869e+02 7.848e+02, threshold=6.498e+02, percent-clipped=7.0 +2023-03-09 15:11:39,232 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-09 15:12:04,819 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9845, 4.9635, 4.6439, 4.8750, 4.9134, 4.3826, 4.8434, 4.6321], + device='cuda:0'), covar=tensor([0.0433, 0.0504, 0.1238, 0.0773, 0.0528, 0.0447, 0.0419, 0.0996], + device='cuda:0'), in_proj_covar=tensor([0.0500, 0.0560, 0.0700, 0.0442, 0.0449, 0.0510, 0.0545, 0.0672], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 15:12:13,380 INFO [train.py:898] (0/4) Epoch 22, batch 1300, loss[loss=0.1717, simple_loss=0.2647, pruned_loss=0.03931, over 18269.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2526, pruned_loss=0.03619, over 3600474.05 frames. ], batch size: 60, lr: 5.14e-03, grad_scale: 8.0 +2023-03-09 15:12:14,882 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77616.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:12:30,233 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77629.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:12:34,905 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77633.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:12:41,786 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0074, 5.4661, 5.4516, 5.4474, 4.9544, 5.4057, 4.8232, 5.3715], + device='cuda:0'), covar=tensor([0.0243, 0.0292, 0.0181, 0.0400, 0.0441, 0.0222, 0.1093, 0.0333], + device='cuda:0'), in_proj_covar=tensor([0.0218, 0.0263, 0.0257, 0.0335, 0.0274, 0.0270, 0.0313, 0.0264], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 15:12:52,257 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-09 15:13:12,280 INFO [train.py:898] (0/4) Epoch 22, batch 1350, loss[loss=0.1395, simple_loss=0.2245, pruned_loss=0.0273, over 18368.00 frames. ], tot_loss[loss=0.1619, simple_loss=0.2518, pruned_loss=0.03601, over 3591155.95 frames. ], batch size: 42, lr: 5.14e-03, grad_scale: 8.0 +2023-03-09 15:13:16,126 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4359, 3.3381, 2.0195, 4.2879, 2.8644, 4.0535, 2.3273, 3.6308], + device='cuda:0'), covar=tensor([0.0674, 0.0907, 0.1644, 0.0433, 0.0890, 0.0292, 0.1367, 0.0481], + device='cuda:0'), in_proj_covar=tensor([0.0216, 0.0228, 0.0193, 0.0288, 0.0193, 0.0265, 0.0205, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 15:13:22,277 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-09 15:13:26,528 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=77677.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:13:26,551 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6636, 3.6695, 3.5132, 3.1866, 3.4088, 2.7914, 2.7342, 3.7617], + device='cuda:0'), covar=tensor([0.0064, 0.0088, 0.0097, 0.0146, 0.0108, 0.0204, 0.0221, 0.0058], + device='cuda:0'), in_proj_covar=tensor([0.0141, 0.0161, 0.0135, 0.0189, 0.0143, 0.0180, 0.0183, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 15:13:26,555 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77677.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:13:29,476 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.757e+02 2.485e+02 2.994e+02 3.704e+02 6.281e+02, threshold=5.988e+02, percent-clipped=0.0 +2023-03-09 15:14:10,447 INFO [train.py:898] (0/4) Epoch 22, batch 1400, loss[loss=0.1549, simple_loss=0.2458, pruned_loss=0.03207, over 18398.00 frames. ], tot_loss[loss=0.1621, simple_loss=0.2519, pruned_loss=0.03618, over 3574523.43 frames. ], batch size: 52, lr: 5.14e-03, grad_scale: 4.0 +2023-03-09 15:14:14,107 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6404, 3.5651, 2.2319, 4.4617, 3.0794, 4.3634, 2.5391, 3.9224], + device='cuda:0'), covar=tensor([0.0662, 0.0841, 0.1553, 0.0504, 0.0876, 0.0366, 0.1202, 0.0444], + device='cuda:0'), in_proj_covar=tensor([0.0216, 0.0229, 0.0193, 0.0289, 0.0193, 0.0266, 0.0205, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 15:14:36,562 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=77738.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:14:50,905 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77751.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 15:15:07,977 INFO [train.py:898] (0/4) Epoch 22, batch 1450, loss[loss=0.1489, simple_loss=0.2307, pruned_loss=0.03356, over 18156.00 frames. ], tot_loss[loss=0.1627, simple_loss=0.2524, pruned_loss=0.03645, over 3572052.87 frames. ], batch size: 44, lr: 5.14e-03, grad_scale: 4.0 +2023-03-09 15:15:26,459 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.943e+02 2.691e+02 3.335e+02 4.479e+02 1.440e+03, threshold=6.670e+02, percent-clipped=5.0 +2023-03-09 15:15:46,838 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=77799.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:16:06,819 INFO [train.py:898] (0/4) Epoch 22, batch 1500, loss[loss=0.1767, simple_loss=0.2717, pruned_loss=0.04089, over 18394.00 frames. ], tot_loss[loss=0.1621, simple_loss=0.2521, pruned_loss=0.03602, over 3581615.02 frames. ], batch size: 52, lr: 5.13e-03, grad_scale: 4.0 +2023-03-09 15:16:42,432 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7365, 2.3629, 2.6178, 2.5899, 3.1863, 4.6551, 4.5599, 3.3517], + device='cuda:0'), covar=tensor([0.1736, 0.2380, 0.3048, 0.1861, 0.2430, 0.0253, 0.0397, 0.0927], + device='cuda:0'), in_proj_covar=tensor([0.0303, 0.0347, 0.0385, 0.0280, 0.0392, 0.0244, 0.0297, 0.0257], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 15:17:04,737 INFO [train.py:898] (0/4) Epoch 22, batch 1550, loss[loss=0.1508, simple_loss=0.2409, pruned_loss=0.03033, over 18358.00 frames. ], tot_loss[loss=0.1619, simple_loss=0.2517, pruned_loss=0.03605, over 3592464.44 frames. ], batch size: 46, lr: 5.13e-03, grad_scale: 4.0 +2023-03-09 15:17:07,434 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77867.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:17:23,123 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.906e+02 2.725e+02 3.263e+02 3.842e+02 6.752e+02, threshold=6.526e+02, percent-clipped=2.0 +2023-03-09 15:18:03,237 INFO [train.py:898] (0/4) Epoch 22, batch 1600, loss[loss=0.1452, simple_loss=0.2279, pruned_loss=0.03124, over 18516.00 frames. ], tot_loss[loss=0.162, simple_loss=0.2516, pruned_loss=0.03622, over 3580815.71 frames. ], batch size: 44, lr: 5.13e-03, grad_scale: 8.0 +2023-03-09 15:18:03,403 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=77915.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:18:19,171 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77929.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:18:24,161 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=77933.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:18:33,126 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.4390, 3.8415, 2.5238, 3.7524, 4.7739, 2.5507, 3.5607, 3.6656], + device='cuda:0'), covar=tensor([0.0239, 0.1097, 0.1593, 0.0688, 0.0115, 0.1256, 0.0739, 0.0829], + device='cuda:0'), in_proj_covar=tensor([0.0171, 0.0275, 0.0207, 0.0197, 0.0130, 0.0186, 0.0216, 0.0228], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 15:18:52,550 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-09 15:19:00,947 INFO [train.py:898] (0/4) Epoch 22, batch 1650, loss[loss=0.1789, simple_loss=0.2738, pruned_loss=0.042, over 18115.00 frames. ], tot_loss[loss=0.1619, simple_loss=0.2514, pruned_loss=0.03623, over 3589395.20 frames. ], batch size: 62, lr: 5.13e-03, grad_scale: 8.0 +2023-03-09 15:19:09,461 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=77972.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:19:15,204 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=77977.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:19:19,785 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=77981.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:19:20,794 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.050e+02 2.525e+02 3.024e+02 3.859e+02 1.001e+03, threshold=6.048e+02, percent-clipped=3.0 +2023-03-09 15:19:26,253 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5117, 2.8442, 3.9669, 3.5494, 2.7928, 4.2196, 3.6369, 2.6772], + device='cuda:0'), covar=tensor([0.0472, 0.1296, 0.0298, 0.0388, 0.1348, 0.0231, 0.0626, 0.1017], + device='cuda:0'), in_proj_covar=tensor([0.0212, 0.0238, 0.0211, 0.0163, 0.0222, 0.0211, 0.0247, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 15:19:42,207 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-78000.pt +2023-03-09 15:20:03,106 INFO [train.py:898] (0/4) Epoch 22, batch 1700, loss[loss=0.1708, simple_loss=0.2614, pruned_loss=0.04006, over 18494.00 frames. ], tot_loss[loss=0.162, simple_loss=0.2516, pruned_loss=0.03617, over 3591280.12 frames. ], batch size: 53, lr: 5.13e-03, grad_scale: 4.0 +2023-03-09 15:20:25,568 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78033.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:21:01,846 INFO [train.py:898] (0/4) Epoch 22, batch 1750, loss[loss=0.1418, simple_loss=0.2262, pruned_loss=0.02871, over 17803.00 frames. ], tot_loss[loss=0.1616, simple_loss=0.2512, pruned_loss=0.036, over 3599456.30 frames. ], batch size: 39, lr: 5.13e-03, grad_scale: 4.0 +2023-03-09 15:21:22,782 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.851e+02 2.720e+02 3.295e+02 3.984e+02 2.464e+03, threshold=6.591e+02, percent-clipped=4.0 +2023-03-09 15:22:00,369 INFO [train.py:898] (0/4) Epoch 22, batch 1800, loss[loss=0.1356, simple_loss=0.2241, pruned_loss=0.02359, over 18257.00 frames. ], tot_loss[loss=0.1618, simple_loss=0.2516, pruned_loss=0.03599, over 3603163.08 frames. ], batch size: 45, lr: 5.12e-03, grad_scale: 4.0 +2023-03-09 15:22:58,103 INFO [train.py:898] (0/4) Epoch 22, batch 1850, loss[loss=0.1813, simple_loss=0.2771, pruned_loss=0.04278, over 18250.00 frames. ], tot_loss[loss=0.1621, simple_loss=0.2522, pruned_loss=0.03606, over 3604077.79 frames. ], batch size: 57, lr: 5.12e-03, grad_scale: 4.0 +2023-03-09 15:23:00,818 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0382, 4.4449, 2.7236, 4.2250, 5.3896, 2.8539, 4.1270, 4.0962], + device='cuda:0'), covar=tensor([0.0165, 0.1015, 0.1481, 0.0588, 0.0100, 0.1060, 0.0549, 0.0710], + device='cuda:0'), in_proj_covar=tensor([0.0168, 0.0271, 0.0203, 0.0195, 0.0128, 0.0182, 0.0214, 0.0224], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 15:23:04,202 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8761, 3.1365, 4.5510, 3.8969, 3.1630, 4.8378, 4.0219, 3.1462], + device='cuda:0'), covar=tensor([0.0434, 0.1307, 0.0210, 0.0405, 0.1320, 0.0169, 0.0571, 0.0908], + device='cuda:0'), in_proj_covar=tensor([0.0213, 0.0237, 0.0212, 0.0163, 0.0222, 0.0211, 0.0246, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 15:23:19,115 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.849e+02 2.526e+02 2.955e+02 3.559e+02 8.124e+02, threshold=5.910e+02, percent-clipped=2.0 +2023-03-09 15:23:57,002 INFO [train.py:898] (0/4) Epoch 22, batch 1900, loss[loss=0.1866, simple_loss=0.2744, pruned_loss=0.04944, over 12650.00 frames. ], tot_loss[loss=0.1622, simple_loss=0.2521, pruned_loss=0.0361, over 3602735.91 frames. ], batch size: 130, lr: 5.12e-03, grad_scale: 4.0 +2023-03-09 15:24:09,953 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8265, 4.5171, 4.5036, 3.3737, 3.6752, 3.5183, 2.5219, 2.3769], + device='cuda:0'), covar=tensor([0.0235, 0.0178, 0.0088, 0.0326, 0.0334, 0.0237, 0.0778, 0.0884], + device='cuda:0'), in_proj_covar=tensor([0.0071, 0.0061, 0.0063, 0.0069, 0.0089, 0.0067, 0.0077, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 15:24:18,221 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78232.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:24:21,502 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8103, 4.5102, 4.4928, 3.3412, 3.6501, 3.4388, 2.3825, 2.4186], + device='cuda:0'), covar=tensor([0.0228, 0.0157, 0.0088, 0.0319, 0.0352, 0.0242, 0.0819, 0.0837], + device='cuda:0'), in_proj_covar=tensor([0.0071, 0.0060, 0.0062, 0.0069, 0.0089, 0.0067, 0.0077, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 15:24:43,776 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7962, 3.2118, 3.9154, 2.9097, 3.6625, 2.6864, 2.7494, 2.3313], + device='cuda:0'), covar=tensor([0.1047, 0.0967, 0.0315, 0.0717, 0.0654, 0.2192, 0.2275, 0.1730], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0247, 0.0197, 0.0201, 0.0260, 0.0276, 0.0329, 0.0237], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 15:24:48,042 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5653, 3.3929, 2.3065, 4.4395, 3.0635, 4.1284, 2.3395, 3.8677], + device='cuda:0'), covar=tensor([0.0652, 0.0919, 0.1471, 0.0505, 0.0819, 0.0401, 0.1300, 0.0461], + device='cuda:0'), in_proj_covar=tensor([0.0215, 0.0227, 0.0192, 0.0286, 0.0193, 0.0266, 0.0204, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 15:24:55,256 INFO [train.py:898] (0/4) Epoch 22, batch 1950, loss[loss=0.1568, simple_loss=0.2485, pruned_loss=0.03252, over 18562.00 frames. ], tot_loss[loss=0.1621, simple_loss=0.252, pruned_loss=0.03608, over 3608018.34 frames. ], batch size: 49, lr: 5.12e-03, grad_scale: 4.0 +2023-03-09 15:25:03,384 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=78272.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:25:09,291 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6022, 2.6125, 2.6210, 2.5007, 2.5964, 2.2560, 2.2903, 2.6975], + device='cuda:0'), covar=tensor([0.0091, 0.0115, 0.0080, 0.0116, 0.0104, 0.0162, 0.0197, 0.0074], + device='cuda:0'), in_proj_covar=tensor([0.0141, 0.0162, 0.0135, 0.0189, 0.0144, 0.0180, 0.0182, 0.0122], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 15:25:14,974 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.944e+02 2.648e+02 3.049e+02 3.777e+02 6.458e+02, threshold=6.098e+02, percent-clipped=2.0 +2023-03-09 15:25:28,629 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=78293.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 15:25:35,359 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5406, 2.7822, 2.5620, 2.9378, 3.6298, 3.5324, 3.0673, 2.9133], + device='cuda:0'), covar=tensor([0.0167, 0.0296, 0.0546, 0.0350, 0.0170, 0.0151, 0.0372, 0.0412], + device='cuda:0'), in_proj_covar=tensor([0.0141, 0.0137, 0.0163, 0.0159, 0.0132, 0.0120, 0.0155, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 15:25:53,379 INFO [train.py:898] (0/4) Epoch 22, batch 2000, loss[loss=0.1682, simple_loss=0.2669, pruned_loss=0.03478, over 17971.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2525, pruned_loss=0.03637, over 3597730.25 frames. ], batch size: 65, lr: 5.12e-03, grad_scale: 8.0 +2023-03-09 15:25:59,291 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=78320.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:26:14,112 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=78333.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:26:47,356 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0737, 3.6692, 5.1816, 2.9095, 4.5911, 2.7569, 3.0941, 1.9354], + device='cuda:0'), covar=tensor([0.1113, 0.0962, 0.0149, 0.1018, 0.0494, 0.2664, 0.2801, 0.2136], + device='cuda:0'), in_proj_covar=tensor([0.0221, 0.0244, 0.0196, 0.0199, 0.0257, 0.0273, 0.0325, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 15:26:52,352 INFO [train.py:898] (0/4) Epoch 22, batch 2050, loss[loss=0.1855, simple_loss=0.2681, pruned_loss=0.0514, over 18492.00 frames. ], tot_loss[loss=0.1628, simple_loss=0.2524, pruned_loss=0.03659, over 3581234.25 frames. ], batch size: 59, lr: 5.12e-03, grad_scale: 8.0 +2023-03-09 15:27:10,383 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=78381.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:27:11,260 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.641e+02 2.890e+02 3.341e+02 3.968e+02 7.791e+02, threshold=6.681e+02, percent-clipped=3.0 +2023-03-09 15:27:36,001 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0598, 5.0563, 5.1263, 4.8627, 4.8230, 4.9122, 5.2259, 5.2129], + device='cuda:0'), covar=tensor([0.0074, 0.0074, 0.0060, 0.0115, 0.0068, 0.0186, 0.0073, 0.0100], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0070, 0.0076, 0.0094, 0.0076, 0.0104, 0.0087, 0.0087], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 15:27:50,918 INFO [train.py:898] (0/4) Epoch 22, batch 2100, loss[loss=0.1636, simple_loss=0.2536, pruned_loss=0.03685, over 18249.00 frames. ], tot_loss[loss=0.163, simple_loss=0.2529, pruned_loss=0.03654, over 3584813.85 frames. ], batch size: 47, lr: 5.11e-03, grad_scale: 8.0 +2023-03-09 15:28:07,860 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6059, 5.1207, 5.1030, 5.0820, 4.6001, 5.0187, 4.4634, 5.0210], + device='cuda:0'), covar=tensor([0.0312, 0.0320, 0.0213, 0.0501, 0.0436, 0.0252, 0.1111, 0.0320], + device='cuda:0'), in_proj_covar=tensor([0.0219, 0.0263, 0.0256, 0.0334, 0.0273, 0.0271, 0.0311, 0.0262], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 15:28:28,195 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-09 15:28:31,351 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6444, 2.4472, 2.7123, 2.6486, 3.2532, 4.7356, 4.6445, 3.3235], + device='cuda:0'), covar=tensor([0.1927, 0.2393, 0.2949, 0.1895, 0.2263, 0.0253, 0.0394, 0.0990], + device='cuda:0'), in_proj_covar=tensor([0.0307, 0.0351, 0.0389, 0.0282, 0.0395, 0.0249, 0.0299, 0.0260], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 15:28:37,826 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3150, 5.8948, 5.4433, 5.6059, 5.4597, 5.3002, 5.9082, 5.8622], + device='cuda:0'), covar=tensor([0.1260, 0.0749, 0.0488, 0.0720, 0.1383, 0.0792, 0.0613, 0.0689], + device='cuda:0'), in_proj_covar=tensor([0.0606, 0.0529, 0.0375, 0.0552, 0.0741, 0.0546, 0.0760, 0.0571], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 15:28:43,484 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78460.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:28:49,225 INFO [train.py:898] (0/4) Epoch 22, batch 2150, loss[loss=0.1547, simple_loss=0.2431, pruned_loss=0.03317, over 18243.00 frames. ], tot_loss[loss=0.1626, simple_loss=0.2526, pruned_loss=0.03631, over 3593329.43 frames. ], batch size: 47, lr: 5.11e-03, grad_scale: 8.0 +2023-03-09 15:28:54,378 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-09 15:29:01,824 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2456, 5.8064, 5.3812, 5.5535, 5.4005, 5.2158, 5.8451, 5.7742], + device='cuda:0'), covar=tensor([0.1298, 0.0796, 0.0555, 0.0726, 0.1451, 0.0819, 0.0654, 0.0761], + device='cuda:0'), in_proj_covar=tensor([0.0610, 0.0532, 0.0377, 0.0554, 0.0743, 0.0549, 0.0761, 0.0573], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 15:29:08,263 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.757e+02 2.625e+02 3.036e+02 3.633e+02 1.479e+03, threshold=6.073e+02, percent-clipped=4.0 +2023-03-09 15:29:31,382 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.76 vs. limit=2.0 +2023-03-09 15:29:47,486 INFO [train.py:898] (0/4) Epoch 22, batch 2200, loss[loss=0.1613, simple_loss=0.2523, pruned_loss=0.03512, over 18486.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2522, pruned_loss=0.03646, over 3586529.22 frames. ], batch size: 51, lr: 5.11e-03, grad_scale: 8.0 +2023-03-09 15:29:55,185 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=78521.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:30:20,727 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-09 15:30:30,368 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0148, 5.5188, 5.4899, 5.4609, 4.9796, 5.4142, 4.8113, 5.3761], + device='cuda:0'), covar=tensor([0.0229, 0.0254, 0.0170, 0.0410, 0.0361, 0.0221, 0.1050, 0.0299], + device='cuda:0'), in_proj_covar=tensor([0.0220, 0.0263, 0.0259, 0.0336, 0.0275, 0.0273, 0.0314, 0.0264], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 15:30:35,214 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78556.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:30:44,767 INFO [train.py:898] (0/4) Epoch 22, batch 2250, loss[loss=0.165, simple_loss=0.2562, pruned_loss=0.03691, over 18384.00 frames. ], tot_loss[loss=0.1623, simple_loss=0.2519, pruned_loss=0.03638, over 3599568.86 frames. ], batch size: 52, lr: 5.11e-03, grad_scale: 8.0 +2023-03-09 15:31:04,898 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.048e+02 2.745e+02 3.101e+02 3.633e+02 7.718e+02, threshold=6.202e+02, percent-clipped=2.0 +2023-03-09 15:31:11,530 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78588.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 15:31:25,003 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-09 15:31:27,141 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-09 15:31:36,141 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8329, 3.4537, 4.7005, 2.9000, 4.1173, 2.6623, 3.0245, 1.9708], + device='cuda:0'), covar=tensor([0.1133, 0.0952, 0.0170, 0.0911, 0.0602, 0.2303, 0.2386, 0.1969], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0247, 0.0197, 0.0200, 0.0260, 0.0275, 0.0327, 0.0236], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 15:31:43,706 INFO [train.py:898] (0/4) Epoch 22, batch 2300, loss[loss=0.1438, simple_loss=0.2315, pruned_loss=0.02806, over 18471.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2524, pruned_loss=0.03636, over 3592088.43 frames. ], batch size: 44, lr: 5.11e-03, grad_scale: 8.0 +2023-03-09 15:31:46,297 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=78617.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:32:05,308 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-03-09 15:32:42,002 INFO [train.py:898] (0/4) Epoch 22, batch 2350, loss[loss=0.1344, simple_loss=0.2225, pruned_loss=0.02322, over 18550.00 frames. ], tot_loss[loss=0.1612, simple_loss=0.2509, pruned_loss=0.03575, over 3601744.95 frames. ], batch size: 49, lr: 5.11e-03, grad_scale: 8.0 +2023-03-09 15:33:01,735 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.907e+02 2.584e+02 2.934e+02 3.285e+02 5.590e+02, threshold=5.868e+02, percent-clipped=0.0 +2023-03-09 15:33:35,138 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7931, 3.1902, 4.5204, 3.7249, 2.8448, 4.7848, 4.0960, 3.0104], + device='cuda:0'), covar=tensor([0.0544, 0.1377, 0.0262, 0.0486, 0.1508, 0.0192, 0.0480, 0.0992], + device='cuda:0'), in_proj_covar=tensor([0.0215, 0.0240, 0.0215, 0.0166, 0.0224, 0.0212, 0.0249, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 15:33:40,407 INFO [train.py:898] (0/4) Epoch 22, batch 2400, loss[loss=0.161, simple_loss=0.2587, pruned_loss=0.03166, over 18360.00 frames. ], tot_loss[loss=0.1614, simple_loss=0.251, pruned_loss=0.03587, over 3605330.04 frames. ], batch size: 55, lr: 5.10e-03, grad_scale: 8.0 +2023-03-09 15:34:15,200 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7731, 4.3888, 4.2955, 3.3422, 3.6423, 3.4309, 2.6046, 2.2828], + device='cuda:0'), covar=tensor([0.0204, 0.0136, 0.0098, 0.0309, 0.0329, 0.0229, 0.0679, 0.0897], + device='cuda:0'), in_proj_covar=tensor([0.0071, 0.0060, 0.0063, 0.0069, 0.0089, 0.0067, 0.0077, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 15:34:39,527 INFO [train.py:898] (0/4) Epoch 22, batch 2450, loss[loss=0.1679, simple_loss=0.2638, pruned_loss=0.036, over 18312.00 frames. ], tot_loss[loss=0.1616, simple_loss=0.2515, pruned_loss=0.03584, over 3604814.60 frames. ], batch size: 54, lr: 5.10e-03, grad_scale: 4.0 +2023-03-09 15:34:51,776 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-09 15:35:00,356 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.881e+02 2.648e+02 3.233e+02 3.743e+02 6.587e+02, threshold=6.467e+02, percent-clipped=2.0 +2023-03-09 15:35:38,713 INFO [train.py:898] (0/4) Epoch 22, batch 2500, loss[loss=0.1861, simple_loss=0.2813, pruned_loss=0.04542, over 18244.00 frames. ], tot_loss[loss=0.1618, simple_loss=0.2517, pruned_loss=0.03588, over 3585309.74 frames. ], batch size: 60, lr: 5.10e-03, grad_scale: 4.0 +2023-03-09 15:35:39,937 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78816.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:36:37,295 INFO [train.py:898] (0/4) Epoch 22, batch 2550, loss[loss=0.1488, simple_loss=0.2441, pruned_loss=0.02673, over 18304.00 frames. ], tot_loss[loss=0.1611, simple_loss=0.2507, pruned_loss=0.03571, over 3584179.79 frames. ], batch size: 54, lr: 5.10e-03, grad_scale: 4.0 +2023-03-09 15:36:47,864 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.87 vs. limit=2.0 +2023-03-09 15:36:57,000 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.832e+02 2.718e+02 3.165e+02 3.980e+02 2.438e+03, threshold=6.330e+02, percent-clipped=4.0 +2023-03-09 15:36:58,283 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3203, 5.2756, 5.6066, 5.7308, 5.2529, 6.1912, 5.8964, 5.5197], + device='cuda:0'), covar=tensor([0.1176, 0.0615, 0.0725, 0.0752, 0.1414, 0.0654, 0.0522, 0.1474], + device='cuda:0'), in_proj_covar=tensor([0.0362, 0.0288, 0.0318, 0.0319, 0.0333, 0.0430, 0.0287, 0.0423], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 15:37:03,648 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=78888.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:37:30,287 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=78912.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:37:30,688 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-09 15:37:34,108 INFO [train.py:898] (0/4) Epoch 22, batch 2600, loss[loss=0.1757, simple_loss=0.2685, pruned_loss=0.04145, over 18372.00 frames. ], tot_loss[loss=0.1614, simple_loss=0.2512, pruned_loss=0.0358, over 3588443.12 frames. ], batch size: 56, lr: 5.10e-03, grad_scale: 4.0 +2023-03-09 15:37:38,515 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-03-09 15:37:45,100 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7339, 3.8023, 3.6475, 3.3479, 3.6182, 3.0373, 2.9698, 3.8253], + device='cuda:0'), covar=tensor([0.0059, 0.0078, 0.0075, 0.0114, 0.0088, 0.0163, 0.0183, 0.0060], + device='cuda:0'), in_proj_covar=tensor([0.0143, 0.0162, 0.0136, 0.0189, 0.0144, 0.0180, 0.0182, 0.0123], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 15:37:58,480 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=78936.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:38:17,344 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5533, 2.7558, 2.5251, 2.8536, 3.6109, 3.4990, 3.0073, 2.8086], + device='cuda:0'), covar=tensor([0.0239, 0.0317, 0.0644, 0.0413, 0.0199, 0.0196, 0.0443, 0.0506], + device='cuda:0'), in_proj_covar=tensor([0.0143, 0.0139, 0.0166, 0.0162, 0.0134, 0.0121, 0.0158, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 15:38:31,478 INFO [train.py:898] (0/4) Epoch 22, batch 2650, loss[loss=0.1693, simple_loss=0.2605, pruned_loss=0.03903, over 18495.00 frames. ], tot_loss[loss=0.1613, simple_loss=0.2509, pruned_loss=0.03582, over 3588304.69 frames. ], batch size: 53, lr: 5.10e-03, grad_scale: 4.0 +2023-03-09 15:38:31,713 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=78965.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:38:52,451 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.743e+02 2.628e+02 3.171e+02 3.718e+02 9.310e+02, threshold=6.343e+02, percent-clipped=1.0 +2023-03-09 15:39:22,908 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.40 vs. limit=5.0 +2023-03-09 15:39:28,931 INFO [train.py:898] (0/4) Epoch 22, batch 2700, loss[loss=0.1706, simple_loss=0.2661, pruned_loss=0.0375, over 18110.00 frames. ], tot_loss[loss=0.1627, simple_loss=0.2525, pruned_loss=0.03648, over 3574705.40 frames. ], batch size: 62, lr: 5.09e-03, grad_scale: 4.0 +2023-03-09 15:39:42,880 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79026.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:40:27,076 INFO [train.py:898] (0/4) Epoch 22, batch 2750, loss[loss=0.1688, simple_loss=0.2599, pruned_loss=0.03888, over 18355.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2522, pruned_loss=0.0364, over 3563325.12 frames. ], batch size: 55, lr: 5.09e-03, grad_scale: 4.0 +2023-03-09 15:40:48,101 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.755e+02 2.673e+02 3.245e+02 3.869e+02 1.413e+03, threshold=6.491e+02, percent-clipped=1.0 +2023-03-09 15:41:25,103 INFO [train.py:898] (0/4) Epoch 22, batch 2800, loss[loss=0.1662, simple_loss=0.2584, pruned_loss=0.03702, over 17749.00 frames. ], tot_loss[loss=0.1622, simple_loss=0.2523, pruned_loss=0.03608, over 3579334.53 frames. ], batch size: 70, lr: 5.09e-03, grad_scale: 8.0 +2023-03-09 15:41:26,572 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79116.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:42:22,265 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=79164.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:42:23,247 INFO [train.py:898] (0/4) Epoch 22, batch 2850, loss[loss=0.143, simple_loss=0.2264, pruned_loss=0.02983, over 18397.00 frames. ], tot_loss[loss=0.1616, simple_loss=0.2516, pruned_loss=0.03586, over 3586822.34 frames. ], batch size: 42, lr: 5.09e-03, grad_scale: 8.0 +2023-03-09 15:42:45,534 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.599e+02 2.640e+02 3.106e+02 3.691e+02 7.833e+02, threshold=6.212e+02, percent-clipped=1.0 +2023-03-09 15:43:01,438 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8326, 5.3438, 5.2956, 5.3203, 4.7872, 5.2084, 4.6968, 5.2347], + device='cuda:0'), covar=tensor([0.0205, 0.0246, 0.0186, 0.0368, 0.0408, 0.0212, 0.0980, 0.0293], + device='cuda:0'), in_proj_covar=tensor([0.0221, 0.0266, 0.0261, 0.0339, 0.0277, 0.0277, 0.0314, 0.0266], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 15:43:15,116 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7836, 3.5834, 4.9526, 4.3059, 3.3318, 3.0139, 4.4502, 5.2091], + device='cuda:0'), covar=tensor([0.0827, 0.1690, 0.0202, 0.0461, 0.0926, 0.1201, 0.0371, 0.0219], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0279, 0.0159, 0.0184, 0.0194, 0.0193, 0.0198, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 15:43:19,575 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79212.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:43:20,779 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4061, 2.4647, 3.7888, 3.4161, 2.4128, 3.8882, 3.5487, 2.5186], + device='cuda:0'), covar=tensor([0.0503, 0.1615, 0.0392, 0.0401, 0.1647, 0.0259, 0.0648, 0.1091], + device='cuda:0'), in_proj_covar=tensor([0.0215, 0.0239, 0.0213, 0.0166, 0.0224, 0.0211, 0.0249, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 15:43:22,656 INFO [train.py:898] (0/4) Epoch 22, batch 2900, loss[loss=0.1419, simple_loss=0.2268, pruned_loss=0.02849, over 18256.00 frames. ], tot_loss[loss=0.1615, simple_loss=0.2516, pruned_loss=0.03565, over 3582947.28 frames. ], batch size: 47, lr: 5.09e-03, grad_scale: 4.0 +2023-03-09 15:43:52,315 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9681, 2.9716, 4.5828, 3.9212, 2.9128, 4.7765, 4.0985, 3.1633], + device='cuda:0'), covar=tensor([0.0405, 0.1436, 0.0250, 0.0394, 0.1320, 0.0203, 0.0509, 0.0809], + device='cuda:0'), in_proj_covar=tensor([0.0215, 0.0240, 0.0214, 0.0166, 0.0224, 0.0212, 0.0249, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 15:44:15,495 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=79260.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:44:21,242 INFO [train.py:898] (0/4) Epoch 22, batch 2950, loss[loss=0.1872, simple_loss=0.2703, pruned_loss=0.05207, over 12433.00 frames. ], tot_loss[loss=0.1617, simple_loss=0.2518, pruned_loss=0.03579, over 3571081.72 frames. ], batch size: 130, lr: 5.09e-03, grad_scale: 4.0 +2023-03-09 15:44:32,704 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.12 vs. limit=5.0 +2023-03-09 15:44:42,794 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.956e+02 2.705e+02 3.045e+02 3.689e+02 6.352e+02, threshold=6.090e+02, percent-clipped=1.0 +2023-03-09 15:44:44,780 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79285.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 15:45:10,035 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2419, 4.1871, 4.0191, 4.1326, 4.1738, 3.7163, 4.1536, 4.0102], + device='cuda:0'), covar=tensor([0.0453, 0.0699, 0.1159, 0.0749, 0.0641, 0.0492, 0.0484, 0.0987], + device='cuda:0'), in_proj_covar=tensor([0.0500, 0.0561, 0.0703, 0.0442, 0.0454, 0.0514, 0.0548, 0.0680], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 15:45:19,366 INFO [train.py:898] (0/4) Epoch 22, batch 3000, loss[loss=0.1436, simple_loss=0.2306, pruned_loss=0.02826, over 18266.00 frames. ], tot_loss[loss=0.1612, simple_loss=0.2512, pruned_loss=0.03558, over 3579293.29 frames. ], batch size: 47, lr: 5.09e-03, grad_scale: 4.0 +2023-03-09 15:45:19,367 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 15:45:31,243 INFO [train.py:932] (0/4) Epoch 22, validation: loss=0.1498, simple_loss=0.249, pruned_loss=0.02526, over 944034.00 frames. +2023-03-09 15:45:31,244 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 15:45:38,815 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79321.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:45:46,633 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2216, 5.2539, 5.5798, 5.5804, 5.2026, 6.0600, 5.7307, 5.3586], + device='cuda:0'), covar=tensor([0.1042, 0.0589, 0.0689, 0.0656, 0.1235, 0.0697, 0.0664, 0.1385], + device='cuda:0'), in_proj_covar=tensor([0.0361, 0.0285, 0.0316, 0.0316, 0.0330, 0.0428, 0.0285, 0.0420], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 15:45:53,359 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79333.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:46:08,189 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79346.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 15:46:13,796 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.74 vs. limit=2.0 +2023-03-09 15:46:28,548 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79364.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:46:29,399 INFO [train.py:898] (0/4) Epoch 22, batch 3050, loss[loss=0.1624, simple_loss=0.26, pruned_loss=0.03237, over 17830.00 frames. ], tot_loss[loss=0.1618, simple_loss=0.2519, pruned_loss=0.03587, over 3580037.98 frames. ], batch size: 70, lr: 5.08e-03, grad_scale: 4.0 +2023-03-09 15:46:41,857 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79375.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 15:46:48,390 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9676, 3.2393, 4.4934, 3.8527, 2.9214, 4.7499, 4.0820, 3.1379], + device='cuda:0'), covar=tensor([0.0444, 0.1346, 0.0288, 0.0442, 0.1468, 0.0203, 0.0568, 0.0945], + device='cuda:0'), in_proj_covar=tensor([0.0213, 0.0237, 0.0213, 0.0165, 0.0223, 0.0210, 0.0247, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 15:46:52,689 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.974e+02 2.697e+02 3.145e+02 3.968e+02 7.194e+02, threshold=6.290e+02, percent-clipped=6.0 +2023-03-09 15:47:04,517 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79394.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:47:28,504 INFO [train.py:898] (0/4) Epoch 22, batch 3100, loss[loss=0.1714, simple_loss=0.2662, pruned_loss=0.03832, over 18097.00 frames. ], tot_loss[loss=0.1622, simple_loss=0.2525, pruned_loss=0.03597, over 3580511.64 frames. ], batch size: 62, lr: 5.08e-03, grad_scale: 4.0 +2023-03-09 15:47:38,380 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9825, 4.1841, 2.5525, 3.9567, 5.2766, 2.6772, 3.7938, 4.0300], + device='cuda:0'), covar=tensor([0.0177, 0.1102, 0.1604, 0.0736, 0.0088, 0.1242, 0.0702, 0.0724], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0270, 0.0203, 0.0196, 0.0128, 0.0183, 0.0216, 0.0224], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 15:47:40,573 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79425.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:47:54,524 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79436.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 15:48:18,291 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-09 15:48:27,404 INFO [train.py:898] (0/4) Epoch 22, batch 3150, loss[loss=0.1634, simple_loss=0.2381, pruned_loss=0.04435, over 17647.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.2526, pruned_loss=0.03604, over 3568603.28 frames. ], batch size: 39, lr: 5.08e-03, grad_scale: 4.0 +2023-03-09 15:48:50,537 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.854e+02 2.702e+02 3.215e+02 3.838e+02 6.492e+02, threshold=6.430e+02, percent-clipped=3.0 +2023-03-09 15:49:09,945 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5324, 2.8272, 2.6069, 2.9365, 3.6214, 3.6060, 3.1123, 2.8365], + device='cuda:0'), covar=tensor([0.0181, 0.0361, 0.0538, 0.0364, 0.0185, 0.0146, 0.0340, 0.0387], + device='cuda:0'), in_proj_covar=tensor([0.0144, 0.0141, 0.0166, 0.0164, 0.0136, 0.0122, 0.0158, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 15:49:25,662 INFO [train.py:898] (0/4) Epoch 22, batch 3200, loss[loss=0.1461, simple_loss=0.2318, pruned_loss=0.03018, over 18400.00 frames. ], tot_loss[loss=0.1618, simple_loss=0.2519, pruned_loss=0.03586, over 3577296.84 frames. ], batch size: 50, lr: 5.08e-03, grad_scale: 8.0 +2023-03-09 15:49:55,877 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8746, 3.1760, 4.5707, 3.8101, 2.8608, 4.7640, 4.0050, 3.1734], + device='cuda:0'), covar=tensor([0.0491, 0.1373, 0.0313, 0.0451, 0.1553, 0.0257, 0.0615, 0.0898], + device='cuda:0'), in_proj_covar=tensor([0.0213, 0.0237, 0.0211, 0.0165, 0.0223, 0.0211, 0.0248, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 15:50:24,049 INFO [train.py:898] (0/4) Epoch 22, batch 3250, loss[loss=0.1505, simple_loss=0.2372, pruned_loss=0.03189, over 17750.00 frames. ], tot_loss[loss=0.162, simple_loss=0.2521, pruned_loss=0.036, over 3573867.89 frames. ], batch size: 39, lr: 5.08e-03, grad_scale: 8.0 +2023-03-09 15:50:46,084 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.828e+02 2.603e+02 3.051e+02 3.520e+02 6.535e+02, threshold=6.103e+02, percent-clipped=1.0 +2023-03-09 15:50:47,552 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79585.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:51:03,228 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-09 15:51:21,967 INFO [train.py:898] (0/4) Epoch 22, batch 3300, loss[loss=0.1438, simple_loss=0.2274, pruned_loss=0.03009, over 18244.00 frames. ], tot_loss[loss=0.1621, simple_loss=0.252, pruned_loss=0.03604, over 3580262.42 frames. ], batch size: 45, lr: 5.08e-03, grad_scale: 8.0 +2023-03-09 15:51:28,951 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79621.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:51:52,908 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79641.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 15:51:58,614 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79646.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:52:10,785 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79657.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:52:19,966 INFO [train.py:898] (0/4) Epoch 22, batch 3350, loss[loss=0.1418, simple_loss=0.221, pruned_loss=0.03135, over 18491.00 frames. ], tot_loss[loss=0.1628, simple_loss=0.2526, pruned_loss=0.03655, over 3576568.45 frames. ], batch size: 44, lr: 5.07e-03, grad_scale: 4.0 +2023-03-09 15:52:24,605 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=79669.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:52:42,678 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.916e+02 2.736e+02 3.253e+02 3.881e+02 1.172e+03, threshold=6.507e+02, percent-clipped=7.0 +2023-03-09 15:52:47,535 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79689.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:53:18,220 INFO [train.py:898] (0/4) Epoch 22, batch 3400, loss[loss=0.1385, simple_loss=0.2159, pruned_loss=0.03053, over 18431.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2521, pruned_loss=0.03641, over 3582908.55 frames. ], batch size: 43, lr: 5.07e-03, grad_scale: 4.0 +2023-03-09 15:53:22,081 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79718.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:53:24,081 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79720.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:53:36,262 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79731.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 15:54:12,517 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=79762.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:54:15,568 INFO [train.py:898] (0/4) Epoch 22, batch 3450, loss[loss=0.1555, simple_loss=0.2439, pruned_loss=0.03352, over 18304.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.252, pruned_loss=0.03649, over 3586983.92 frames. ], batch size: 60, lr: 5.07e-03, grad_scale: 4.0 +2023-03-09 15:54:38,366 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.822e+02 2.532e+02 2.897e+02 3.710e+02 6.365e+02, threshold=5.795e+02, percent-clipped=0.0 +2023-03-09 15:55:14,087 INFO [train.py:898] (0/4) Epoch 22, batch 3500, loss[loss=0.1424, simple_loss=0.2256, pruned_loss=0.0296, over 18478.00 frames. ], tot_loss[loss=0.1621, simple_loss=0.2517, pruned_loss=0.03623, over 3585379.82 frames. ], batch size: 44, lr: 5.07e-03, grad_scale: 4.0 +2023-03-09 15:55:23,978 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=79823.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:56:09,525 INFO [train.py:898] (0/4) Epoch 22, batch 3550, loss[loss=0.1525, simple_loss=0.2455, pruned_loss=0.02972, over 18506.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.2522, pruned_loss=0.03634, over 3572831.97 frames. ], batch size: 51, lr: 5.07e-03, grad_scale: 4.0 +2023-03-09 15:56:32,290 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.670e+02 2.485e+02 2.929e+02 3.619e+02 1.143e+03, threshold=5.859e+02, percent-clipped=2.0 +2023-03-09 15:56:38,819 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-09 15:57:05,048 INFO [train.py:898] (0/4) Epoch 22, batch 3600, loss[loss=0.1414, simple_loss=0.2217, pruned_loss=0.03055, over 18439.00 frames. ], tot_loss[loss=0.1625, simple_loss=0.2523, pruned_loss=0.03632, over 3576896.77 frames. ], batch size: 43, lr: 5.07e-03, grad_scale: 8.0 +2023-03-09 15:57:32,482 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=79941.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:57:32,617 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79941.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 15:57:40,252 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-22.pt +2023-03-09 15:58:07,278 INFO [train.py:898] (0/4) Epoch 23, batch 0, loss[loss=0.1437, simple_loss=0.2232, pruned_loss=0.03212, over 17626.00 frames. ], tot_loss[loss=0.1437, simple_loss=0.2232, pruned_loss=0.03212, over 17626.00 frames. ], batch size: 39, lr: 4.95e-03, grad_scale: 8.0 +2023-03-09 15:58:07,281 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 15:58:18,937 INFO [train.py:932] (0/4) Epoch 23, validation: loss=0.1494, simple_loss=0.2493, pruned_loss=0.02473, over 944034.00 frames. +2023-03-09 15:58:18,937 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 15:59:01,335 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.492e+02 2.757e+02 3.307e+02 4.142e+02 8.059e+02, threshold=6.615e+02, percent-clipped=1.0 +2023-03-09 15:59:06,092 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=79989.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 15:59:06,131 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=79989.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:59:17,065 INFO [train.py:898] (0/4) Epoch 23, batch 50, loss[loss=0.1535, simple_loss=0.2453, pruned_loss=0.03079, over 18306.00 frames. ], tot_loss[loss=0.1592, simple_loss=0.2495, pruned_loss=0.03449, over 825353.99 frames. ], batch size: 54, lr: 4.95e-03, grad_scale: 8.0 +2023-03-09 15:59:18,674 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-80000.pt +2023-03-09 15:59:37,722 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80013.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:59:46,516 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80020.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 15:59:50,438 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3150, 5.3072, 5.5711, 5.5542, 5.2294, 6.1091, 5.7125, 5.3851], + device='cuda:0'), covar=tensor([0.1146, 0.0637, 0.0771, 0.0869, 0.1479, 0.0710, 0.0696, 0.1626], + device='cuda:0'), in_proj_covar=tensor([0.0365, 0.0287, 0.0317, 0.0316, 0.0329, 0.0434, 0.0287, 0.0422], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 16:00:00,050 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80031.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 16:00:06,681 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=80037.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:00:20,039 INFO [train.py:898] (0/4) Epoch 23, batch 100, loss[loss=0.1591, simple_loss=0.2561, pruned_loss=0.03099, over 18345.00 frames. ], tot_loss[loss=0.1601, simple_loss=0.2499, pruned_loss=0.0352, over 1443109.29 frames. ], batch size: 55, lr: 4.95e-03, grad_scale: 8.0 +2023-03-09 16:00:41,817 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=80068.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:00:56,573 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=80079.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 16:01:02,964 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.638e+02 2.619e+02 2.981e+02 3.602e+02 9.245e+02, threshold=5.963e+02, percent-clipped=1.0 +2023-03-09 16:01:18,865 INFO [train.py:898] (0/4) Epoch 23, batch 150, loss[loss=0.1722, simple_loss=0.259, pruned_loss=0.04275, over 18253.00 frames. ], tot_loss[loss=0.1611, simple_loss=0.2508, pruned_loss=0.03576, over 1932510.00 frames. ], batch size: 47, lr: 4.95e-03, grad_scale: 8.0 +2023-03-09 16:01:19,259 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8616, 4.5241, 4.6121, 3.4310, 3.6844, 3.5400, 2.7250, 2.5598], + device='cuda:0'), covar=tensor([0.0199, 0.0151, 0.0066, 0.0300, 0.0302, 0.0204, 0.0694, 0.0826], + device='cuda:0'), in_proj_covar=tensor([0.0070, 0.0060, 0.0064, 0.0070, 0.0089, 0.0067, 0.0077, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 16:01:40,092 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80118.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:01:58,592 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-09 16:02:04,964 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6590, 3.7231, 3.5693, 3.1679, 3.3771, 2.8344, 2.7646, 3.7018], + device='cuda:0'), covar=tensor([0.0063, 0.0089, 0.0091, 0.0138, 0.0110, 0.0195, 0.0225, 0.0076], + device='cuda:0'), in_proj_covar=tensor([0.0144, 0.0164, 0.0137, 0.0191, 0.0145, 0.0181, 0.0184, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 16:02:07,361 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7989, 3.4404, 4.8700, 2.6640, 4.2337, 2.5441, 3.0079, 1.7584], + device='cuda:0'), covar=tensor([0.1212, 0.1007, 0.0172, 0.1046, 0.0594, 0.2700, 0.2548, 0.2268], + device='cuda:0'), in_proj_covar=tensor([0.0225, 0.0249, 0.0203, 0.0203, 0.0262, 0.0275, 0.0328, 0.0241], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 16:02:10,705 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6785, 2.8979, 2.4962, 2.9384, 3.7425, 3.6717, 3.1878, 2.9359], + device='cuda:0'), covar=tensor([0.0192, 0.0322, 0.0607, 0.0398, 0.0191, 0.0157, 0.0409, 0.0433], + device='cuda:0'), in_proj_covar=tensor([0.0144, 0.0142, 0.0168, 0.0164, 0.0136, 0.0122, 0.0159, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:02:17,083 INFO [train.py:898] (0/4) Epoch 23, batch 200, loss[loss=0.1705, simple_loss=0.2758, pruned_loss=0.0326, over 18387.00 frames. ], tot_loss[loss=0.1611, simple_loss=0.2508, pruned_loss=0.03573, over 2291792.47 frames. ], batch size: 50, lr: 4.95e-03, grad_scale: 8.0 +2023-03-09 16:02:59,527 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.861e+02 2.844e+02 3.404e+02 4.018e+02 9.589e+02, threshold=6.808e+02, percent-clipped=5.0 +2023-03-09 16:03:15,929 INFO [train.py:898] (0/4) Epoch 23, batch 250, loss[loss=0.1704, simple_loss=0.2657, pruned_loss=0.03757, over 17045.00 frames. ], tot_loss[loss=0.1615, simple_loss=0.2512, pruned_loss=0.03589, over 2588592.87 frames. ], batch size: 78, lr: 4.94e-03, grad_scale: 8.0 +2023-03-09 16:03:33,257 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3209, 5.3005, 4.9448, 5.2163, 5.2258, 4.6339, 5.1461, 4.9145], + device='cuda:0'), covar=tensor([0.0441, 0.0458, 0.1239, 0.0825, 0.0661, 0.0426, 0.0450, 0.1169], + device='cuda:0'), in_proj_covar=tensor([0.0500, 0.0567, 0.0708, 0.0444, 0.0459, 0.0518, 0.0550, 0.0687], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 16:04:03,409 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0109, 5.0928, 5.1396, 4.8410, 4.9006, 4.9226, 5.2102, 5.1912], + device='cuda:0'), covar=tensor([0.0069, 0.0063, 0.0053, 0.0104, 0.0058, 0.0156, 0.0086, 0.0086], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0071, 0.0075, 0.0094, 0.0076, 0.0105, 0.0088, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 16:04:05,491 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80241.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:04:14,921 INFO [train.py:898] (0/4) Epoch 23, batch 300, loss[loss=0.1597, simple_loss=0.2474, pruned_loss=0.036, over 18556.00 frames. ], tot_loss[loss=0.1615, simple_loss=0.2515, pruned_loss=0.0358, over 2827731.46 frames. ], batch size: 49, lr: 4.94e-03, grad_scale: 8.0 +2023-03-09 16:04:52,886 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80283.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:04:54,783 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.957e+02 2.544e+02 3.151e+02 3.664e+02 8.600e+02, threshold=6.302e+02, percent-clipped=1.0 +2023-03-09 16:05:00,692 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=80289.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:05:12,210 INFO [train.py:898] (0/4) Epoch 23, batch 350, loss[loss=0.1451, simple_loss=0.2327, pruned_loss=0.02873, over 18509.00 frames. ], tot_loss[loss=0.1613, simple_loss=0.2513, pruned_loss=0.03561, over 3002539.07 frames. ], batch size: 47, lr: 4.94e-03, grad_scale: 8.0 +2023-03-09 16:05:21,621 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3326, 2.7064, 2.3376, 2.7054, 3.5133, 3.3624, 2.9679, 2.7658], + device='cuda:0'), covar=tensor([0.0219, 0.0293, 0.0581, 0.0389, 0.0195, 0.0191, 0.0419, 0.0398], + device='cuda:0'), in_proj_covar=tensor([0.0143, 0.0140, 0.0166, 0.0163, 0.0135, 0.0121, 0.0158, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:05:27,945 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80313.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:05:59,689 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3892, 5.3648, 5.7493, 5.7686, 5.2886, 6.2575, 5.8958, 5.4808], + device='cuda:0'), covar=tensor([0.1111, 0.0562, 0.0688, 0.0777, 0.1454, 0.0630, 0.0569, 0.1438], + device='cuda:0'), in_proj_covar=tensor([0.0364, 0.0287, 0.0317, 0.0315, 0.0330, 0.0432, 0.0285, 0.0422], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 16:06:04,343 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80344.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:06:09,756 INFO [train.py:898] (0/4) Epoch 23, batch 400, loss[loss=0.1482, simple_loss=0.2406, pruned_loss=0.0279, over 18568.00 frames. ], tot_loss[loss=0.1612, simple_loss=0.2512, pruned_loss=0.03564, over 3130417.50 frames. ], batch size: 54, lr: 4.94e-03, grad_scale: 8.0 +2023-03-09 16:06:23,658 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=80361.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:06:50,966 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.659e+02 2.745e+02 3.129e+02 3.783e+02 6.813e+02, threshold=6.257e+02, percent-clipped=3.0 +2023-03-09 16:07:02,235 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80393.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:07:08,711 INFO [train.py:898] (0/4) Epoch 23, batch 450, loss[loss=0.1689, simple_loss=0.2663, pruned_loss=0.03571, over 18505.00 frames. ], tot_loss[loss=0.162, simple_loss=0.2522, pruned_loss=0.03589, over 3222550.63 frames. ], batch size: 53, lr: 4.94e-03, grad_scale: 8.0 +2023-03-09 16:07:31,433 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80418.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:08:07,200 INFO [train.py:898] (0/4) Epoch 23, batch 500, loss[loss=0.1421, simple_loss=0.2256, pruned_loss=0.0293, over 18504.00 frames. ], tot_loss[loss=0.1608, simple_loss=0.2507, pruned_loss=0.03542, over 3310731.85 frames. ], batch size: 44, lr: 4.94e-03, grad_scale: 8.0 +2023-03-09 16:08:13,182 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80454.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:08:23,632 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80463.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:08:26,718 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=80466.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:08:33,838 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-09 16:08:47,894 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.857e+02 2.536e+02 3.027e+02 3.548e+02 8.560e+02, threshold=6.054e+02, percent-clipped=2.0 +2023-03-09 16:08:58,984 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80494.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:09:05,399 INFO [train.py:898] (0/4) Epoch 23, batch 550, loss[loss=0.1674, simple_loss=0.2546, pruned_loss=0.04015, over 18148.00 frames. ], tot_loss[loss=0.161, simple_loss=0.2506, pruned_loss=0.03571, over 3379103.58 frames. ], batch size: 62, lr: 4.93e-03, grad_scale: 8.0 +2023-03-09 16:09:33,966 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80524.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:10:02,835 INFO [train.py:898] (0/4) Epoch 23, batch 600, loss[loss=0.1757, simple_loss=0.2691, pruned_loss=0.04115, over 18242.00 frames. ], tot_loss[loss=0.1604, simple_loss=0.2498, pruned_loss=0.03555, over 3424492.41 frames. ], batch size: 60, lr: 4.93e-03, grad_scale: 8.0 +2023-03-09 16:10:05,887 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0708, 4.5548, 4.2914, 4.3360, 4.1304, 4.7482, 4.4961, 4.1633], + device='cuda:0'), covar=tensor([0.1419, 0.1071, 0.1054, 0.0923, 0.1597, 0.1196, 0.0797, 0.1880], + device='cuda:0'), in_proj_covar=tensor([0.0362, 0.0285, 0.0316, 0.0314, 0.0329, 0.0430, 0.0284, 0.0422], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 16:10:10,578 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80555.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:10:44,260 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.750e+02 2.691e+02 3.309e+02 4.060e+02 6.105e+02, threshold=6.618e+02, percent-clipped=2.0 +2023-03-09 16:10:59,880 INFO [train.py:898] (0/4) Epoch 23, batch 650, loss[loss=0.1754, simple_loss=0.2637, pruned_loss=0.04358, over 18488.00 frames. ], tot_loss[loss=0.1605, simple_loss=0.2499, pruned_loss=0.03551, over 3466682.50 frames. ], batch size: 53, lr: 4.93e-03, grad_scale: 8.0 +2023-03-09 16:11:14,915 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-09 16:11:24,912 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6469, 2.8781, 2.7090, 2.9288, 3.7387, 3.6793, 3.1742, 3.0482], + device='cuda:0'), covar=tensor([0.0158, 0.0294, 0.0508, 0.0358, 0.0171, 0.0140, 0.0387, 0.0353], + device='cuda:0'), in_proj_covar=tensor([0.0142, 0.0140, 0.0164, 0.0161, 0.0135, 0.0120, 0.0157, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:11:35,153 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9041, 4.2134, 2.2772, 4.0739, 5.1720, 2.6583, 3.4697, 3.7517], + device='cuda:0'), covar=tensor([0.0219, 0.1190, 0.1778, 0.0680, 0.0107, 0.1227, 0.0917, 0.0963], + device='cuda:0'), in_proj_covar=tensor([0.0170, 0.0270, 0.0204, 0.0196, 0.0130, 0.0182, 0.0215, 0.0225], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:11:47,298 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80639.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:11:58,519 INFO [train.py:898] (0/4) Epoch 23, batch 700, loss[loss=0.1669, simple_loss=0.2674, pruned_loss=0.03314, over 17731.00 frames. ], tot_loss[loss=0.1605, simple_loss=0.25, pruned_loss=0.03546, over 3500157.40 frames. ], batch size: 70, lr: 4.93e-03, grad_scale: 8.0 +2023-03-09 16:12:41,893 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.765e+02 2.683e+02 3.069e+02 3.814e+02 7.668e+02, threshold=6.138e+02, percent-clipped=2.0 +2023-03-09 16:12:46,765 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3994, 5.9050, 5.5114, 5.6692, 5.5272, 5.3496, 6.0133, 5.9098], + device='cuda:0'), covar=tensor([0.1290, 0.0840, 0.0464, 0.0753, 0.1464, 0.0721, 0.0605, 0.0791], + device='cuda:0'), in_proj_covar=tensor([0.0623, 0.0541, 0.0387, 0.0565, 0.0765, 0.0560, 0.0774, 0.0589], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 16:12:57,646 INFO [train.py:898] (0/4) Epoch 23, batch 750, loss[loss=0.1621, simple_loss=0.2569, pruned_loss=0.03368, over 18362.00 frames. ], tot_loss[loss=0.1611, simple_loss=0.2512, pruned_loss=0.03553, over 3521328.92 frames. ], batch size: 55, lr: 4.93e-03, grad_scale: 8.0 +2023-03-09 16:13:41,892 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80736.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:13:56,501 INFO [train.py:898] (0/4) Epoch 23, batch 800, loss[loss=0.1646, simple_loss=0.2609, pruned_loss=0.03414, over 18254.00 frames. ], tot_loss[loss=0.1598, simple_loss=0.2499, pruned_loss=0.03481, over 3549228.71 frames. ], batch size: 47, lr: 4.93e-03, grad_scale: 8.0 +2023-03-09 16:13:56,734 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80749.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:14:22,830 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9664, 5.4929, 5.4787, 5.4761, 4.9626, 5.3846, 4.8614, 5.3487], + device='cuda:0'), covar=tensor([0.0206, 0.0243, 0.0152, 0.0328, 0.0368, 0.0183, 0.0898, 0.0282], + device='cuda:0'), in_proj_covar=tensor([0.0218, 0.0260, 0.0256, 0.0332, 0.0272, 0.0268, 0.0305, 0.0261], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 16:14:25,099 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=80773.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:14:27,561 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5725, 3.4366, 4.5333, 4.0747, 3.1526, 2.8406, 4.0496, 4.7234], + device='cuda:0'), covar=tensor([0.0869, 0.1459, 0.0256, 0.0443, 0.0985, 0.1260, 0.0466, 0.0259], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0279, 0.0160, 0.0184, 0.0193, 0.0193, 0.0199, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:14:38,914 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.739e+02 2.483e+02 2.864e+02 3.599e+02 5.476e+02, threshold=5.727e+02, percent-clipped=0.0 +2023-03-09 16:14:53,004 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80797.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:14:54,773 INFO [train.py:898] (0/4) Epoch 23, batch 850, loss[loss=0.1618, simple_loss=0.255, pruned_loss=0.03432, over 17095.00 frames. ], tot_loss[loss=0.1604, simple_loss=0.2508, pruned_loss=0.03505, over 3557734.05 frames. ], batch size: 78, lr: 4.93e-03, grad_scale: 8.0 +2023-03-09 16:15:18,300 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80819.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:15:36,592 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=80834.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:15:51,333 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-09 16:15:52,795 INFO [train.py:898] (0/4) Epoch 23, batch 900, loss[loss=0.1492, simple_loss=0.2354, pruned_loss=0.03156, over 18356.00 frames. ], tot_loss[loss=0.1602, simple_loss=0.2504, pruned_loss=0.03501, over 3571041.69 frames. ], batch size: 46, lr: 4.92e-03, grad_scale: 8.0 +2023-03-09 16:15:54,056 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=80850.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:16:34,842 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.923e+02 2.539e+02 2.932e+02 3.679e+02 1.116e+03, threshold=5.864e+02, percent-clipped=6.0 +2023-03-09 16:16:51,059 INFO [train.py:898] (0/4) Epoch 23, batch 950, loss[loss=0.1407, simple_loss=0.2244, pruned_loss=0.02849, over 18491.00 frames. ], tot_loss[loss=0.1602, simple_loss=0.2503, pruned_loss=0.03505, over 3589880.87 frames. ], batch size: 44, lr: 4.92e-03, grad_scale: 8.0 +2023-03-09 16:17:37,160 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=80939.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:17:48,619 INFO [train.py:898] (0/4) Epoch 23, batch 1000, loss[loss=0.1697, simple_loss=0.2645, pruned_loss=0.03748, over 18304.00 frames. ], tot_loss[loss=0.1596, simple_loss=0.2496, pruned_loss=0.03483, over 3601336.67 frames. ], batch size: 57, lr: 4.92e-03, grad_scale: 8.0 +2023-03-09 16:18:20,617 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7211, 3.6094, 4.9895, 4.5029, 3.2749, 3.0209, 4.4408, 5.1644], + device='cuda:0'), covar=tensor([0.0796, 0.1491, 0.0177, 0.0347, 0.0996, 0.1183, 0.0392, 0.0270], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0278, 0.0159, 0.0184, 0.0192, 0.0192, 0.0198, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:18:30,653 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.676e+02 2.676e+02 3.165e+02 3.583e+02 7.202e+02, threshold=6.331e+02, percent-clipped=5.0 +2023-03-09 16:18:33,108 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=80987.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:18:41,726 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-09 16:18:46,992 INFO [train.py:898] (0/4) Epoch 23, batch 1050, loss[loss=0.1505, simple_loss=0.2438, pruned_loss=0.02862, over 18502.00 frames. ], tot_loss[loss=0.1598, simple_loss=0.2499, pruned_loss=0.03483, over 3604073.89 frames. ], batch size: 53, lr: 4.92e-03, grad_scale: 8.0 +2023-03-09 16:19:11,169 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8310, 3.7527, 5.0779, 4.5908, 3.2946, 3.0559, 4.5710, 5.3053], + device='cuda:0'), covar=tensor([0.0808, 0.1376, 0.0182, 0.0359, 0.1013, 0.1203, 0.0345, 0.0208], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0281, 0.0160, 0.0185, 0.0194, 0.0194, 0.0199, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:19:12,371 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6848, 2.4323, 2.7802, 2.8258, 3.3904, 4.9263, 4.7905, 3.3959], + device='cuda:0'), covar=tensor([0.1874, 0.2517, 0.2834, 0.1853, 0.2299, 0.0251, 0.0380, 0.0988], + device='cuda:0'), in_proj_covar=tensor([0.0309, 0.0352, 0.0389, 0.0283, 0.0393, 0.0253, 0.0299, 0.0260], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 16:19:45,395 INFO [train.py:898] (0/4) Epoch 23, batch 1100, loss[loss=0.1367, simple_loss=0.2242, pruned_loss=0.02458, over 18514.00 frames. ], tot_loss[loss=0.1596, simple_loss=0.2497, pruned_loss=0.03477, over 3611775.40 frames. ], batch size: 44, lr: 4.92e-03, grad_scale: 8.0 +2023-03-09 16:19:45,615 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81049.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:20:27,976 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.813e+02 2.500e+02 2.967e+02 3.515e+02 7.145e+02, threshold=5.934e+02, percent-clipped=1.0 +2023-03-09 16:20:35,907 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81092.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:20:41,602 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=81097.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:20:43,687 INFO [train.py:898] (0/4) Epoch 23, batch 1150, loss[loss=0.166, simple_loss=0.2676, pruned_loss=0.03216, over 18552.00 frames. ], tot_loss[loss=0.1604, simple_loss=0.2507, pruned_loss=0.03501, over 3602862.60 frames. ], batch size: 54, lr: 4.92e-03, grad_scale: 8.0 +2023-03-09 16:20:51,662 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.75 vs. limit=2.0 +2023-03-09 16:21:03,407 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81116.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:21:06,588 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81119.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:21:18,130 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81129.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:21:41,473 INFO [train.py:898] (0/4) Epoch 23, batch 1200, loss[loss=0.1736, simple_loss=0.2685, pruned_loss=0.0393, over 18226.00 frames. ], tot_loss[loss=0.1602, simple_loss=0.2501, pruned_loss=0.0351, over 3602122.53 frames. ], batch size: 60, lr: 4.91e-03, grad_scale: 8.0 +2023-03-09 16:21:42,827 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81150.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:22:02,578 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=81167.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:22:02,916 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6863, 2.3204, 2.6835, 2.6865, 3.1077, 4.7251, 4.6559, 3.3718], + device='cuda:0'), covar=tensor([0.1868, 0.2606, 0.3097, 0.1858, 0.2552, 0.0284, 0.0408, 0.1006], + device='cuda:0'), in_proj_covar=tensor([0.0309, 0.0353, 0.0390, 0.0284, 0.0393, 0.0253, 0.0299, 0.0261], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 16:22:13,711 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81177.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:22:14,875 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6290, 3.2079, 4.3177, 3.9738, 3.0903, 2.9293, 3.8611, 4.4431], + device='cuda:0'), covar=tensor([0.0878, 0.1565, 0.0266, 0.0447, 0.1010, 0.1180, 0.0519, 0.0426], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0279, 0.0160, 0.0185, 0.0193, 0.0192, 0.0199, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:22:22,853 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.778e+02 2.664e+02 3.163e+02 3.705e+02 6.920e+02, threshold=6.326e+02, percent-clipped=3.0 +2023-03-09 16:22:38,931 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=81198.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:22:39,916 INFO [train.py:898] (0/4) Epoch 23, batch 1250, loss[loss=0.1497, simple_loss=0.2463, pruned_loss=0.02658, over 17721.00 frames. ], tot_loss[loss=0.1611, simple_loss=0.2507, pruned_loss=0.03576, over 3583326.90 frames. ], batch size: 70, lr: 4.91e-03, grad_scale: 8.0 +2023-03-09 16:23:39,292 INFO [train.py:898] (0/4) Epoch 23, batch 1300, loss[loss=0.1729, simple_loss=0.2669, pruned_loss=0.0395, over 18303.00 frames. ], tot_loss[loss=0.1605, simple_loss=0.2503, pruned_loss=0.03535, over 3581897.03 frames. ], batch size: 57, lr: 4.91e-03, grad_scale: 8.0 +2023-03-09 16:24:22,196 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.771e+02 2.557e+02 2.964e+02 3.868e+02 7.836e+02, threshold=5.928e+02, percent-clipped=1.0 +2023-03-09 16:24:37,825 INFO [train.py:898] (0/4) Epoch 23, batch 1350, loss[loss=0.1902, simple_loss=0.275, pruned_loss=0.05274, over 18408.00 frames. ], tot_loss[loss=0.161, simple_loss=0.251, pruned_loss=0.03547, over 3589477.58 frames. ], batch size: 48, lr: 4.91e-03, grad_scale: 4.0 +2023-03-09 16:24:38,212 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81299.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:24:40,411 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81301.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:25:13,982 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9748, 3.2274, 4.6129, 3.8833, 3.2243, 4.9268, 4.1190, 3.2346], + device='cuda:0'), covar=tensor([0.0415, 0.1351, 0.0304, 0.0460, 0.1323, 0.0183, 0.0517, 0.0880], + device='cuda:0'), in_proj_covar=tensor([0.0211, 0.0237, 0.0213, 0.0164, 0.0223, 0.0209, 0.0248, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 16:25:36,790 INFO [train.py:898] (0/4) Epoch 23, batch 1400, loss[loss=0.1581, simple_loss=0.2537, pruned_loss=0.03125, over 15982.00 frames. ], tot_loss[loss=0.1606, simple_loss=0.2505, pruned_loss=0.03534, over 3585840.28 frames. ], batch size: 94, lr: 4.91e-03, grad_scale: 4.0 +2023-03-09 16:25:49,534 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81360.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:25:51,881 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81362.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:26:19,227 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.872e+02 2.620e+02 3.077e+02 3.710e+02 7.565e+02, threshold=6.154e+02, percent-clipped=6.0 +2023-03-09 16:26:26,898 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81392.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:26:35,533 INFO [train.py:898] (0/4) Epoch 23, batch 1450, loss[loss=0.1617, simple_loss=0.2579, pruned_loss=0.0328, over 18375.00 frames. ], tot_loss[loss=0.1598, simple_loss=0.2497, pruned_loss=0.035, over 3584836.22 frames. ], batch size: 55, lr: 4.91e-03, grad_scale: 4.0 +2023-03-09 16:26:36,004 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7928, 3.2344, 3.9703, 2.9033, 3.6401, 2.7141, 2.8150, 2.3277], + device='cuda:0'), covar=tensor([0.1072, 0.1040, 0.0330, 0.0757, 0.0694, 0.2192, 0.2198, 0.1753], + device='cuda:0'), in_proj_covar=tensor([0.0225, 0.0250, 0.0205, 0.0202, 0.0261, 0.0276, 0.0329, 0.0240], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 16:27:10,659 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81429.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:27:23,521 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=81440.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:27:34,458 INFO [train.py:898] (0/4) Epoch 23, batch 1500, loss[loss=0.1684, simple_loss=0.256, pruned_loss=0.04037, over 17795.00 frames. ], tot_loss[loss=0.1598, simple_loss=0.2498, pruned_loss=0.03494, over 3584880.16 frames. ], batch size: 70, lr: 4.91e-03, grad_scale: 4.0 +2023-03-09 16:28:01,899 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81472.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:28:07,466 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=81477.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:28:17,384 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.729e+02 2.620e+02 2.957e+02 3.344e+02 7.866e+02, threshold=5.914e+02, percent-clipped=3.0 +2023-03-09 16:28:32,684 INFO [train.py:898] (0/4) Epoch 23, batch 1550, loss[loss=0.1474, simple_loss=0.2378, pruned_loss=0.02845, over 18285.00 frames. ], tot_loss[loss=0.1605, simple_loss=0.2505, pruned_loss=0.03524, over 3569322.89 frames. ], batch size: 49, lr: 4.90e-03, grad_scale: 4.0 +2023-03-09 16:29:31,016 INFO [train.py:898] (0/4) Epoch 23, batch 1600, loss[loss=0.1708, simple_loss=0.2605, pruned_loss=0.04052, over 18622.00 frames. ], tot_loss[loss=0.1612, simple_loss=0.2514, pruned_loss=0.03548, over 3566670.58 frames. ], batch size: 52, lr: 4.90e-03, grad_scale: 8.0 +2023-03-09 16:29:54,485 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1575, 3.4265, 3.3485, 2.9388, 3.0067, 2.9507, 2.4348, 2.3528], + device='cuda:0'), covar=tensor([0.0258, 0.0174, 0.0146, 0.0301, 0.0354, 0.0232, 0.0634, 0.0717], + device='cuda:0'), in_proj_covar=tensor([0.0073, 0.0061, 0.0065, 0.0070, 0.0091, 0.0067, 0.0078, 0.0086], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 16:30:06,141 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-09 16:30:15,102 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.871e+02 2.795e+02 3.306e+02 4.129e+02 9.714e+02, threshold=6.611e+02, percent-clipped=7.0 +2023-03-09 16:30:20,089 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9266, 3.6591, 4.9969, 2.9402, 4.4042, 2.6570, 3.0222, 1.7066], + device='cuda:0'), covar=tensor([0.1188, 0.0941, 0.0180, 0.0940, 0.0533, 0.2739, 0.2919, 0.2395], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0248, 0.0204, 0.0200, 0.0259, 0.0275, 0.0326, 0.0239], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 16:30:29,165 INFO [train.py:898] (0/4) Epoch 23, batch 1650, loss[loss=0.1735, simple_loss=0.2627, pruned_loss=0.04221, over 16924.00 frames. ], tot_loss[loss=0.1611, simple_loss=0.251, pruned_loss=0.03565, over 3563150.15 frames. ], batch size: 78, lr: 4.90e-03, grad_scale: 4.0 +2023-03-09 16:31:23,558 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8313, 3.6166, 5.0037, 4.4452, 3.3885, 3.0551, 4.3817, 5.1705], + device='cuda:0'), covar=tensor([0.0824, 0.1636, 0.0211, 0.0402, 0.0966, 0.1224, 0.0434, 0.0293], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0279, 0.0161, 0.0185, 0.0194, 0.0193, 0.0199, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:31:28,181 INFO [train.py:898] (0/4) Epoch 23, batch 1700, loss[loss=0.1819, simple_loss=0.2795, pruned_loss=0.04214, over 18070.00 frames. ], tot_loss[loss=0.1606, simple_loss=0.2505, pruned_loss=0.03538, over 3569844.98 frames. ], batch size: 62, lr: 4.90e-03, grad_scale: 4.0 +2023-03-09 16:31:35,085 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81655.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:31:38,686 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=81657.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:31:53,574 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81670.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:32:12,319 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.4587, 2.1081, 2.0741, 2.1302, 2.4226, 2.5200, 2.3619, 2.1741], + device='cuda:0'), covar=tensor([0.0218, 0.0227, 0.0384, 0.0331, 0.0205, 0.0165, 0.0336, 0.0309], + device='cuda:0'), in_proj_covar=tensor([0.0141, 0.0139, 0.0164, 0.0162, 0.0134, 0.0120, 0.0156, 0.0160], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:32:13,034 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.642e+02 2.453e+02 2.824e+02 3.386e+02 8.042e+02, threshold=5.649e+02, percent-clipped=1.0 +2023-03-09 16:32:26,758 INFO [train.py:898] (0/4) Epoch 23, batch 1750, loss[loss=0.1443, simple_loss=0.2329, pruned_loss=0.02784, over 18329.00 frames. ], tot_loss[loss=0.161, simple_loss=0.251, pruned_loss=0.03549, over 3559228.50 frames. ], batch size: 46, lr: 4.90e-03, grad_scale: 4.0 +2023-03-09 16:32:35,495 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5629, 3.2755, 2.2279, 4.2774, 3.0161, 4.0514, 2.4069, 3.8433], + device='cuda:0'), covar=tensor([0.0674, 0.0936, 0.1507, 0.0573, 0.0881, 0.0368, 0.1317, 0.0422], + device='cuda:0'), in_proj_covar=tensor([0.0214, 0.0226, 0.0189, 0.0285, 0.0192, 0.0265, 0.0203, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:32:39,905 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7902, 3.7400, 3.6554, 3.3184, 3.5095, 3.0145, 2.9653, 3.8627], + device='cuda:0'), covar=tensor([0.0068, 0.0101, 0.0084, 0.0134, 0.0107, 0.0185, 0.0205, 0.0063], + device='cuda:0'), in_proj_covar=tensor([0.0144, 0.0165, 0.0137, 0.0192, 0.0146, 0.0182, 0.0186, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 16:33:04,546 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=81731.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 16:33:25,097 INFO [train.py:898] (0/4) Epoch 23, batch 1800, loss[loss=0.1524, simple_loss=0.2475, pruned_loss=0.02865, over 18272.00 frames. ], tot_loss[loss=0.1604, simple_loss=0.2502, pruned_loss=0.03528, over 3562972.26 frames. ], batch size: 49, lr: 4.90e-03, grad_scale: 2.0 +2023-03-09 16:33:52,152 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81772.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:33:52,619 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.92 vs. limit=5.0 +2023-03-09 16:33:58,033 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9741, 5.4589, 2.9800, 5.2795, 5.1762, 5.5068, 5.3329, 2.9556], + device='cuda:0'), covar=tensor([0.0219, 0.0063, 0.0699, 0.0077, 0.0068, 0.0067, 0.0082, 0.0883], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0082, 0.0096, 0.0097, 0.0087, 0.0077, 0.0086, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 16:34:10,541 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.038e+02 2.659e+02 3.003e+02 3.499e+02 8.184e+02, threshold=6.005e+02, percent-clipped=4.0 +2023-03-09 16:34:23,258 INFO [train.py:898] (0/4) Epoch 23, batch 1850, loss[loss=0.1565, simple_loss=0.2502, pruned_loss=0.03141, over 18378.00 frames. ], tot_loss[loss=0.1599, simple_loss=0.2497, pruned_loss=0.03505, over 3566746.10 frames. ], batch size: 50, lr: 4.90e-03, grad_scale: 2.0 +2023-03-09 16:34:27,263 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5860, 2.7889, 2.6522, 2.8460, 3.6348, 3.4677, 3.0981, 2.9169], + device='cuda:0'), covar=tensor([0.0196, 0.0349, 0.0582, 0.0444, 0.0207, 0.0199, 0.0418, 0.0472], + device='cuda:0'), in_proj_covar=tensor([0.0141, 0.0139, 0.0164, 0.0162, 0.0134, 0.0121, 0.0156, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:34:48,396 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=81820.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:35:21,108 INFO [train.py:898] (0/4) Epoch 23, batch 1900, loss[loss=0.1603, simple_loss=0.252, pruned_loss=0.03427, over 18255.00 frames. ], tot_loss[loss=0.1606, simple_loss=0.2506, pruned_loss=0.0353, over 3567919.95 frames. ], batch size: 45, lr: 4.89e-03, grad_scale: 2.0 +2023-03-09 16:35:27,054 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0256, 5.5190, 3.2409, 5.3646, 5.2748, 5.5992, 5.4815, 3.1618], + device='cuda:0'), covar=tensor([0.0204, 0.0084, 0.0618, 0.0068, 0.0072, 0.0072, 0.0072, 0.0787], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0081, 0.0095, 0.0096, 0.0086, 0.0077, 0.0085, 0.0096], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 16:36:07,196 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.831e+02 2.605e+02 3.105e+02 3.631e+02 5.941e+02, threshold=6.209e+02, percent-clipped=0.0 +2023-03-09 16:36:20,013 INFO [train.py:898] (0/4) Epoch 23, batch 1950, loss[loss=0.1454, simple_loss=0.2328, pruned_loss=0.02901, over 18238.00 frames. ], tot_loss[loss=0.1612, simple_loss=0.2511, pruned_loss=0.03561, over 3577713.73 frames. ], batch size: 45, lr: 4.89e-03, grad_scale: 2.0 +2023-03-09 16:37:17,997 INFO [train.py:898] (0/4) Epoch 23, batch 2000, loss[loss=0.19, simple_loss=0.2765, pruned_loss=0.0517, over 18221.00 frames. ], tot_loss[loss=0.1613, simple_loss=0.2512, pruned_loss=0.03574, over 3582433.66 frames. ], batch size: 60, lr: 4.89e-03, grad_scale: 4.0 +2023-03-09 16:37:25,082 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81955.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:37:27,438 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=81957.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:37:30,987 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6426, 3.5063, 5.1730, 3.0290, 4.4904, 2.6365, 3.0453, 1.7949], + device='cuda:0'), covar=tensor([0.1272, 0.0954, 0.0133, 0.0842, 0.0498, 0.2507, 0.2613, 0.2211], + device='cuda:0'), in_proj_covar=tensor([0.0224, 0.0247, 0.0204, 0.0201, 0.0259, 0.0276, 0.0327, 0.0240], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 16:37:41,806 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5411, 3.9100, 2.4992, 3.8221, 4.8035, 2.8112, 3.5025, 3.4875], + device='cuda:0'), covar=tensor([0.0204, 0.1286, 0.1494, 0.0677, 0.0111, 0.1038, 0.0725, 0.0862], + device='cuda:0'), in_proj_covar=tensor([0.0174, 0.0277, 0.0208, 0.0202, 0.0134, 0.0187, 0.0220, 0.0229], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:38:03,261 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=81987.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:38:04,114 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.810e+02 2.510e+02 2.896e+02 3.401e+02 7.542e+02, threshold=5.791e+02, percent-clipped=1.0 +2023-03-09 16:38:08,109 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-09 16:38:09,090 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8857, 3.6735, 5.0703, 3.0164, 4.4137, 2.6818, 3.1526, 1.8348], + device='cuda:0'), covar=tensor([0.1130, 0.0935, 0.0157, 0.0895, 0.0523, 0.2506, 0.2671, 0.2136], + device='cuda:0'), in_proj_covar=tensor([0.0224, 0.0247, 0.0204, 0.0201, 0.0259, 0.0276, 0.0327, 0.0240], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 16:38:16,682 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.63 vs. limit=5.0 +2023-03-09 16:38:17,072 INFO [train.py:898] (0/4) Epoch 23, batch 2050, loss[loss=0.1481, simple_loss=0.2322, pruned_loss=0.03194, over 18395.00 frames. ], tot_loss[loss=0.1607, simple_loss=0.2507, pruned_loss=0.0354, over 3597362.65 frames. ], batch size: 42, lr: 4.89e-03, grad_scale: 4.0 +2023-03-09 16:38:18,580 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-82000.pt +2023-03-09 16:38:26,567 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=82003.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:38:28,779 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=82005.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:38:54,186 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82026.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 16:38:57,091 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.44 vs. limit=2.0 +2023-03-09 16:39:18,929 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82048.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:39:19,762 INFO [train.py:898] (0/4) Epoch 23, batch 2100, loss[loss=0.1921, simple_loss=0.2713, pruned_loss=0.05642, over 12757.00 frames. ], tot_loss[loss=0.1613, simple_loss=0.2512, pruned_loss=0.03569, over 3604384.68 frames. ], batch size: 129, lr: 4.89e-03, grad_scale: 4.0 +2023-03-09 16:39:27,593 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82055.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:40:05,445 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.125e+02 2.645e+02 3.113e+02 3.776e+02 5.395e+02, threshold=6.226e+02, percent-clipped=0.0 +2023-03-09 16:40:18,009 INFO [train.py:898] (0/4) Epoch 23, batch 2150, loss[loss=0.1544, simple_loss=0.2382, pruned_loss=0.0353, over 18069.00 frames. ], tot_loss[loss=0.1609, simple_loss=0.2507, pruned_loss=0.03558, over 3605966.57 frames. ], batch size: 40, lr: 4.89e-03, grad_scale: 4.0 +2023-03-09 16:40:38,283 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82116.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:40:43,151 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2464, 5.2242, 4.8716, 5.1697, 5.1687, 4.5975, 5.0464, 4.8536], + device='cuda:0'), covar=tensor([0.0416, 0.0484, 0.1229, 0.0660, 0.0655, 0.0391, 0.0443, 0.1002], + device='cuda:0'), in_proj_covar=tensor([0.0503, 0.0569, 0.0710, 0.0439, 0.0460, 0.0518, 0.0549, 0.0685], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 16:41:16,326 INFO [train.py:898] (0/4) Epoch 23, batch 2200, loss[loss=0.1704, simple_loss=0.268, pruned_loss=0.03638, over 18028.00 frames. ], tot_loss[loss=0.1617, simple_loss=0.2517, pruned_loss=0.03584, over 3595105.39 frames. ], batch size: 65, lr: 4.88e-03, grad_scale: 4.0 +2023-03-09 16:41:21,843 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0379, 5.4772, 5.4852, 5.4519, 4.9980, 5.4097, 4.8894, 5.3989], + device='cuda:0'), covar=tensor([0.0190, 0.0250, 0.0161, 0.0342, 0.0372, 0.0196, 0.0857, 0.0252], + device='cuda:0'), in_proj_covar=tensor([0.0218, 0.0263, 0.0259, 0.0333, 0.0273, 0.0269, 0.0307, 0.0261], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 16:41:38,478 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.80 vs. limit=2.0 +2023-03-09 16:42:02,195 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.746e+02 2.684e+02 3.246e+02 4.259e+02 8.856e+02, threshold=6.492e+02, percent-clipped=7.0 +2023-03-09 16:42:14,734 INFO [train.py:898] (0/4) Epoch 23, batch 2250, loss[loss=0.1562, simple_loss=0.2461, pruned_loss=0.03314, over 17139.00 frames. ], tot_loss[loss=0.1618, simple_loss=0.2519, pruned_loss=0.03592, over 3574985.54 frames. ], batch size: 78, lr: 4.88e-03, grad_scale: 4.0 +2023-03-09 16:43:06,997 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.61 vs. limit=2.0 +2023-03-09 16:43:13,052 INFO [train.py:898] (0/4) Epoch 23, batch 2300, loss[loss=0.1347, simple_loss=0.2188, pruned_loss=0.02526, over 18485.00 frames. ], tot_loss[loss=0.1615, simple_loss=0.2514, pruned_loss=0.03584, over 3575739.08 frames. ], batch size: 44, lr: 4.88e-03, grad_scale: 4.0 +2023-03-09 16:43:59,036 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.010e+02 2.750e+02 3.270e+02 3.982e+02 6.063e+02, threshold=6.540e+02, percent-clipped=0.0 +2023-03-09 16:44:11,843 INFO [train.py:898] (0/4) Epoch 23, batch 2350, loss[loss=0.1678, simple_loss=0.2495, pruned_loss=0.04307, over 18526.00 frames. ], tot_loss[loss=0.1613, simple_loss=0.2512, pruned_loss=0.0357, over 3591905.33 frames. ], batch size: 49, lr: 4.88e-03, grad_scale: 4.0 +2023-03-09 16:44:43,135 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=82326.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:45:04,035 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82343.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:45:07,636 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82346.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:45:10,706 INFO [train.py:898] (0/4) Epoch 23, batch 2400, loss[loss=0.1513, simple_loss=0.2427, pruned_loss=0.02993, over 18548.00 frames. ], tot_loss[loss=0.161, simple_loss=0.251, pruned_loss=0.03546, over 3593150.99 frames. ], batch size: 49, lr: 4.88e-03, grad_scale: 8.0 +2023-03-09 16:45:39,428 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=82374.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:45:55,568 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.666e+02 2.631e+02 3.119e+02 3.737e+02 9.140e+02, threshold=6.238e+02, percent-clipped=2.0 +2023-03-09 16:46:09,287 INFO [train.py:898] (0/4) Epoch 23, batch 2450, loss[loss=0.138, simple_loss=0.2222, pruned_loss=0.02686, over 18243.00 frames. ], tot_loss[loss=0.1605, simple_loss=0.2503, pruned_loss=0.03536, over 3592493.93 frames. ], batch size: 45, lr: 4.88e-03, grad_scale: 8.0 +2023-03-09 16:46:18,868 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82407.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:46:23,214 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82411.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:46:31,426 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0611, 4.2110, 2.5766, 4.1633, 5.2401, 2.7831, 3.8627, 4.2426], + device='cuda:0'), covar=tensor([0.0199, 0.1432, 0.1568, 0.0624, 0.0093, 0.1180, 0.0702, 0.0626], + device='cuda:0'), in_proj_covar=tensor([0.0173, 0.0275, 0.0206, 0.0199, 0.0133, 0.0185, 0.0218, 0.0227], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:47:08,006 INFO [train.py:898] (0/4) Epoch 23, batch 2500, loss[loss=0.1842, simple_loss=0.2792, pruned_loss=0.04455, over 18391.00 frames. ], tot_loss[loss=0.161, simple_loss=0.2511, pruned_loss=0.03544, over 3592926.68 frames. ], batch size: 56, lr: 4.88e-03, grad_scale: 8.0 +2023-03-09 16:47:16,155 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-09 16:47:52,902 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.741e+02 2.534e+02 3.083e+02 3.529e+02 5.828e+02, threshold=6.166e+02, percent-clipped=0.0 +2023-03-09 16:47:54,417 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4977, 2.8284, 2.4764, 2.7859, 3.6077, 3.5709, 3.1623, 2.8721], + device='cuda:0'), covar=tensor([0.0249, 0.0330, 0.0555, 0.0368, 0.0199, 0.0164, 0.0343, 0.0404], + device='cuda:0'), in_proj_covar=tensor([0.0141, 0.0140, 0.0164, 0.0161, 0.0136, 0.0121, 0.0157, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:47:55,469 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4811, 3.2372, 2.1009, 4.3363, 2.9955, 4.0311, 2.4403, 3.7520], + device='cuda:0'), covar=tensor([0.0719, 0.0992, 0.1684, 0.0513, 0.0970, 0.0389, 0.1284, 0.0500], + device='cuda:0'), in_proj_covar=tensor([0.0219, 0.0230, 0.0192, 0.0291, 0.0195, 0.0269, 0.0206, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:47:58,829 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82493.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 16:48:06,026 INFO [train.py:898] (0/4) Epoch 23, batch 2550, loss[loss=0.147, simple_loss=0.2444, pruned_loss=0.02474, over 18269.00 frames. ], tot_loss[loss=0.1611, simple_loss=0.2511, pruned_loss=0.03554, over 3590930.64 frames. ], batch size: 49, lr: 4.87e-03, grad_scale: 8.0 +2023-03-09 16:48:06,675 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.62 vs. limit=5.0 +2023-03-09 16:49:03,666 INFO [train.py:898] (0/4) Epoch 23, batch 2600, loss[loss=0.156, simple_loss=0.2493, pruned_loss=0.03141, over 18493.00 frames. ], tot_loss[loss=0.1606, simple_loss=0.2506, pruned_loss=0.03536, over 3593018.58 frames. ], batch size: 53, lr: 4.87e-03, grad_scale: 8.0 +2023-03-09 16:49:09,571 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-09 16:49:10,215 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82554.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 16:49:21,814 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-09 16:49:27,014 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82569.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:49:49,705 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.492e+02 2.620e+02 3.051e+02 3.714e+02 9.693e+02, threshold=6.103e+02, percent-clipped=7.0 +2023-03-09 16:50:01,837 INFO [train.py:898] (0/4) Epoch 23, batch 2650, loss[loss=0.159, simple_loss=0.2425, pruned_loss=0.03769, over 18445.00 frames. ], tot_loss[loss=0.1605, simple_loss=0.2505, pruned_loss=0.03521, over 3598346.44 frames. ], batch size: 43, lr: 4.87e-03, grad_scale: 4.0 +2023-03-09 16:50:20,495 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9150, 3.5049, 2.5500, 3.3545, 4.0263, 2.5609, 3.3666, 3.4295], + device='cuda:0'), covar=tensor([0.0266, 0.0963, 0.1386, 0.0631, 0.0152, 0.1082, 0.0658, 0.0731], + device='cuda:0'), in_proj_covar=tensor([0.0174, 0.0275, 0.0207, 0.0199, 0.0133, 0.0186, 0.0219, 0.0228], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:50:38,522 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=82630.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:50:54,200 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=82643.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:51:00,472 INFO [train.py:898] (0/4) Epoch 23, batch 2700, loss[loss=0.1781, simple_loss=0.2688, pruned_loss=0.04373, over 18323.00 frames. ], tot_loss[loss=0.1607, simple_loss=0.2507, pruned_loss=0.03536, over 3573587.58 frames. ], batch size: 54, lr: 4.87e-03, grad_scale: 4.0 +2023-03-09 16:51:46,588 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.924e+02 2.721e+02 3.267e+02 3.832e+02 9.086e+02, threshold=6.534e+02, percent-clipped=2.0 +2023-03-09 16:51:49,041 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=82691.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:51:58,553 INFO [train.py:898] (0/4) Epoch 23, batch 2750, loss[loss=0.1533, simple_loss=0.2449, pruned_loss=0.03084, over 18626.00 frames. ], tot_loss[loss=0.1602, simple_loss=0.2501, pruned_loss=0.03516, over 3580657.97 frames. ], batch size: 52, lr: 4.87e-03, grad_scale: 4.0 +2023-03-09 16:52:02,233 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82702.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:52:12,997 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=82711.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:52:55,993 INFO [train.py:898] (0/4) Epoch 23, batch 2800, loss[loss=0.1512, simple_loss=0.2432, pruned_loss=0.02962, over 18408.00 frames. ], tot_loss[loss=0.1609, simple_loss=0.2507, pruned_loss=0.03558, over 3575015.37 frames. ], batch size: 48, lr: 4.87e-03, grad_scale: 8.0 +2023-03-09 16:53:07,956 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=82759.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:53:42,292 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.820e+02 2.582e+02 3.010e+02 3.495e+02 9.650e+02, threshold=6.020e+02, percent-clipped=2.0 +2023-03-09 16:53:47,365 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3341, 5.8834, 5.3950, 5.6264, 5.4403, 5.2537, 5.9111, 5.9064], + device='cuda:0'), covar=tensor([0.1161, 0.0689, 0.0521, 0.0640, 0.1356, 0.0711, 0.0575, 0.0636], + device='cuda:0'), in_proj_covar=tensor([0.0617, 0.0538, 0.0387, 0.0561, 0.0762, 0.0556, 0.0769, 0.0582], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 16:53:54,217 INFO [train.py:898] (0/4) Epoch 23, batch 2850, loss[loss=0.1537, simple_loss=0.2312, pruned_loss=0.03809, over 18165.00 frames. ], tot_loss[loss=0.1613, simple_loss=0.2512, pruned_loss=0.03571, over 3581548.55 frames. ], batch size: 44, lr: 4.87e-03, grad_scale: 8.0 +2023-03-09 16:54:27,458 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7834, 3.7556, 5.1528, 4.6895, 3.3769, 3.1968, 4.6066, 5.3108], + device='cuda:0'), covar=tensor([0.0810, 0.1536, 0.0162, 0.0316, 0.0920, 0.1096, 0.0348, 0.0277], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0281, 0.0162, 0.0186, 0.0195, 0.0195, 0.0200, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:54:43,436 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2499, 4.4195, 2.8187, 4.3482, 5.3353, 2.7414, 4.0640, 4.2192], + device='cuda:0'), covar=tensor([0.0140, 0.1085, 0.1360, 0.0565, 0.0089, 0.1114, 0.0607, 0.0680], + device='cuda:0'), in_proj_covar=tensor([0.0175, 0.0277, 0.0209, 0.0201, 0.0135, 0.0186, 0.0221, 0.0230], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 16:54:52,995 INFO [train.py:898] (0/4) Epoch 23, batch 2900, loss[loss=0.1345, simple_loss=0.2153, pruned_loss=0.02685, over 18414.00 frames. ], tot_loss[loss=0.1615, simple_loss=0.2514, pruned_loss=0.03581, over 3588316.13 frames. ], batch size: 42, lr: 4.86e-03, grad_scale: 8.0 +2023-03-09 16:54:53,145 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82849.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 16:55:39,402 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.765e+02 2.643e+02 2.969e+02 3.593e+02 6.828e+02, threshold=5.939e+02, percent-clipped=2.0 +2023-03-09 16:55:51,332 INFO [train.py:898] (0/4) Epoch 23, batch 2950, loss[loss=0.1585, simple_loss=0.258, pruned_loss=0.02947, over 18302.00 frames. ], tot_loss[loss=0.1617, simple_loss=0.2517, pruned_loss=0.03588, over 3593496.22 frames. ], batch size: 54, lr: 4.86e-03, grad_scale: 8.0 +2023-03-09 16:56:02,465 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5809, 2.4063, 2.5431, 2.6601, 2.9694, 4.1996, 4.1210, 3.1938], + device='cuda:0'), covar=tensor([0.1948, 0.2493, 0.2854, 0.1952, 0.2503, 0.0359, 0.0479, 0.0982], + device='cuda:0'), in_proj_covar=tensor([0.0310, 0.0352, 0.0391, 0.0284, 0.0393, 0.0251, 0.0298, 0.0262], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 16:56:22,404 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=82925.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:56:24,853 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1573, 3.1042, 2.9867, 2.8184, 3.0223, 2.4410, 2.5184, 3.1225], + device='cuda:0'), covar=tensor([0.0069, 0.0102, 0.0094, 0.0123, 0.0094, 0.0192, 0.0203, 0.0071], + device='cuda:0'), in_proj_covar=tensor([0.0145, 0.0165, 0.0137, 0.0190, 0.0145, 0.0181, 0.0184, 0.0124], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 16:56:48,951 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7330, 3.7614, 3.5737, 3.2134, 3.5398, 2.8423, 2.9033, 3.8098], + device='cuda:0'), covar=tensor([0.0065, 0.0098, 0.0086, 0.0128, 0.0095, 0.0199, 0.0201, 0.0065], + device='cuda:0'), in_proj_covar=tensor([0.0145, 0.0165, 0.0137, 0.0190, 0.0145, 0.0181, 0.0185, 0.0125], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 16:56:49,686 INFO [train.py:898] (0/4) Epoch 23, batch 3000, loss[loss=0.152, simple_loss=0.2454, pruned_loss=0.02926, over 18276.00 frames. ], tot_loss[loss=0.1619, simple_loss=0.2522, pruned_loss=0.03578, over 3585925.10 frames. ], batch size: 49, lr: 4.86e-03, grad_scale: 8.0 +2023-03-09 16:56:49,688 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 16:56:56,358 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1681, 4.8118, 4.3997, 4.5395, 4.3027, 4.8022, 4.6760, 4.2450], + device='cuda:0'), covar=tensor([0.1591, 0.0526, 0.1182, 0.0732, 0.1245, 0.1170, 0.0726, 0.2360], + device='cuda:0'), in_proj_covar=tensor([0.0364, 0.0293, 0.0313, 0.0319, 0.0334, 0.0430, 0.0287, 0.0420], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 16:57:01,578 INFO [train.py:932] (0/4) Epoch 23, validation: loss=0.1503, simple_loss=0.2492, pruned_loss=0.02572, over 944034.00 frames. +2023-03-09 16:57:01,579 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 16:57:19,134 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=82963.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:57:22,046 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-09 16:57:28,546 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.47 vs. limit=2.0 +2023-03-09 16:57:41,558 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5852, 5.4374, 5.7758, 5.8338, 5.5006, 6.2927, 5.9899, 5.5058], + device='cuda:0'), covar=tensor([0.0964, 0.0507, 0.0668, 0.0604, 0.1288, 0.0615, 0.0567, 0.1653], + device='cuda:0'), in_proj_covar=tensor([0.0362, 0.0291, 0.0311, 0.0318, 0.0333, 0.0428, 0.0286, 0.0419], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 16:57:46,015 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.46 vs. limit=2.0 +2023-03-09 16:57:48,596 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.819e+02 2.491e+02 2.993e+02 3.582e+02 9.758e+02, threshold=5.985e+02, percent-clipped=2.0 +2023-03-09 16:57:53,140 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.66 vs. limit=2.0 +2023-03-09 16:58:00,492 INFO [train.py:898] (0/4) Epoch 23, batch 3050, loss[loss=0.1666, simple_loss=0.261, pruned_loss=0.03607, over 18280.00 frames. ], tot_loss[loss=0.1619, simple_loss=0.2521, pruned_loss=0.03581, over 3582196.03 frames. ], batch size: 54, lr: 4.86e-03, grad_scale: 8.0 +2023-03-09 16:58:04,654 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83002.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:58:14,423 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.6318, 6.1544, 5.6849, 5.9761, 5.7592, 5.6199, 6.2571, 6.1650], + device='cuda:0'), covar=tensor([0.1269, 0.0678, 0.0391, 0.0617, 0.1353, 0.0725, 0.0568, 0.0688], + device='cuda:0'), in_proj_covar=tensor([0.0619, 0.0541, 0.0388, 0.0566, 0.0763, 0.0556, 0.0771, 0.0587], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 16:58:20,206 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6188, 5.0790, 5.0543, 5.0477, 4.5649, 4.9656, 4.4524, 4.9439], + device='cuda:0'), covar=tensor([0.0252, 0.0261, 0.0194, 0.0435, 0.0407, 0.0231, 0.1016, 0.0315], + device='cuda:0'), in_proj_covar=tensor([0.0220, 0.0265, 0.0259, 0.0337, 0.0277, 0.0274, 0.0309, 0.0264], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 16:58:30,339 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83024.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:58:45,476 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.80 vs. limit=2.0 +2023-03-09 16:58:59,524 INFO [train.py:898] (0/4) Epoch 23, batch 3100, loss[loss=0.1503, simple_loss=0.235, pruned_loss=0.03274, over 18274.00 frames. ], tot_loss[loss=0.1609, simple_loss=0.2509, pruned_loss=0.03544, over 3575508.69 frames. ], batch size: 45, lr: 4.86e-03, grad_scale: 4.0 +2023-03-09 16:59:00,887 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=83050.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 16:59:14,401 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5234, 5.4954, 5.1392, 5.4092, 5.4515, 4.8267, 5.3548, 5.0827], + device='cuda:0'), covar=tensor([0.0409, 0.0372, 0.1217, 0.0849, 0.0508, 0.0473, 0.0403, 0.1043], + device='cuda:0'), in_proj_covar=tensor([0.0499, 0.0569, 0.0710, 0.0439, 0.0458, 0.0518, 0.0546, 0.0685], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0004, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 16:59:46,641 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.874e+02 2.570e+02 3.037e+02 3.721e+02 1.351e+03, threshold=6.073e+02, percent-clipped=1.0 +2023-03-09 16:59:57,763 INFO [train.py:898] (0/4) Epoch 23, batch 3150, loss[loss=0.1677, simple_loss=0.2662, pruned_loss=0.0346, over 18338.00 frames. ], tot_loss[loss=0.1611, simple_loss=0.251, pruned_loss=0.03563, over 3565410.79 frames. ], batch size: 55, lr: 4.86e-03, grad_scale: 4.0 +2023-03-09 17:00:55,290 INFO [train.py:898] (0/4) Epoch 23, batch 3200, loss[loss=0.182, simple_loss=0.273, pruned_loss=0.04548, over 16125.00 frames. ], tot_loss[loss=0.162, simple_loss=0.2522, pruned_loss=0.03585, over 3571471.49 frames. ], batch size: 95, lr: 4.86e-03, grad_scale: 8.0 +2023-03-09 17:00:55,633 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83149.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 17:01:25,269 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6395, 3.3977, 2.4702, 4.4419, 3.1911, 4.2761, 2.5857, 4.0003], + device='cuda:0'), covar=tensor([0.0622, 0.0856, 0.1304, 0.0486, 0.0772, 0.0336, 0.1194, 0.0405], + device='cuda:0'), in_proj_covar=tensor([0.0216, 0.0227, 0.0190, 0.0288, 0.0192, 0.0267, 0.0204, 0.0203], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 17:01:43,024 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.642e+02 2.631e+02 3.256e+02 4.142e+02 1.137e+03, threshold=6.512e+02, percent-clipped=5.0 +2023-03-09 17:01:49,052 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0706, 5.1780, 5.3305, 5.3949, 4.9654, 5.8956, 5.4751, 5.0027], + device='cuda:0'), covar=tensor([0.1182, 0.0681, 0.0780, 0.0681, 0.1400, 0.0672, 0.0690, 0.1795], + device='cuda:0'), in_proj_covar=tensor([0.0367, 0.0293, 0.0316, 0.0321, 0.0335, 0.0432, 0.0290, 0.0424], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 17:01:50,282 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83196.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:01:51,142 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=83197.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 17:01:53,670 INFO [train.py:898] (0/4) Epoch 23, batch 3250, loss[loss=0.172, simple_loss=0.2631, pruned_loss=0.04045, over 18114.00 frames. ], tot_loss[loss=0.1614, simple_loss=0.2517, pruned_loss=0.03561, over 3580470.63 frames. ], batch size: 62, lr: 4.85e-03, grad_scale: 8.0 +2023-03-09 17:02:25,229 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83225.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:02:51,888 INFO [train.py:898] (0/4) Epoch 23, batch 3300, loss[loss=0.1821, simple_loss=0.2768, pruned_loss=0.04368, over 18117.00 frames. ], tot_loss[loss=0.1617, simple_loss=0.2521, pruned_loss=0.03567, over 3574393.38 frames. ], batch size: 62, lr: 4.85e-03, grad_scale: 8.0 +2023-03-09 17:03:02,445 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83257.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:03:21,829 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=83273.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:03:34,347 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83284.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:03:40,706 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.708e+02 2.590e+02 3.037e+02 3.666e+02 6.079e+02, threshold=6.073e+02, percent-clipped=0.0 +2023-03-09 17:03:50,690 INFO [train.py:898] (0/4) Epoch 23, batch 3350, loss[loss=0.188, simple_loss=0.2776, pruned_loss=0.04923, over 18369.00 frames. ], tot_loss[loss=0.1617, simple_loss=0.252, pruned_loss=0.03568, over 3573168.85 frames. ], batch size: 56, lr: 4.85e-03, grad_scale: 8.0 +2023-03-09 17:04:00,824 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-09 17:04:14,280 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=83319.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:04:45,120 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83345.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:04:49,292 INFO [train.py:898] (0/4) Epoch 23, batch 3400, loss[loss=0.1654, simple_loss=0.2514, pruned_loss=0.03968, over 18380.00 frames. ], tot_loss[loss=0.1619, simple_loss=0.252, pruned_loss=0.03583, over 3572125.49 frames. ], batch size: 50, lr: 4.85e-03, grad_scale: 8.0 +2023-03-09 17:05:37,372 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.783e+02 2.666e+02 3.258e+02 4.159e+02 1.161e+03, threshold=6.516e+02, percent-clipped=10.0 +2023-03-09 17:05:47,416 INFO [train.py:898] (0/4) Epoch 23, batch 3450, loss[loss=0.1369, simple_loss=0.2219, pruned_loss=0.02595, over 18507.00 frames. ], tot_loss[loss=0.1624, simple_loss=0.2524, pruned_loss=0.03616, over 3579056.85 frames. ], batch size: 47, lr: 4.85e-03, grad_scale: 8.0 +2023-03-09 17:06:03,360 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.45 vs. limit=5.0 +2023-03-09 17:06:45,592 INFO [train.py:898] (0/4) Epoch 23, batch 3500, loss[loss=0.1357, simple_loss=0.2111, pruned_loss=0.03015, over 18410.00 frames. ], tot_loss[loss=0.1613, simple_loss=0.2515, pruned_loss=0.03559, over 3590578.17 frames. ], batch size: 42, lr: 4.85e-03, grad_scale: 8.0 +2023-03-09 17:07:31,221 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.962e+02 2.487e+02 3.071e+02 3.696e+02 9.873e+02, threshold=6.142e+02, percent-clipped=4.0 +2023-03-09 17:07:41,350 INFO [train.py:898] (0/4) Epoch 23, batch 3550, loss[loss=0.1744, simple_loss=0.263, pruned_loss=0.04293, over 18229.00 frames. ], tot_loss[loss=0.1615, simple_loss=0.2514, pruned_loss=0.0358, over 3585015.92 frames. ], batch size: 60, lr: 4.85e-03, grad_scale: 8.0 +2023-03-09 17:08:21,755 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8247, 3.8428, 3.6369, 3.3221, 3.5198, 2.8848, 2.8764, 3.7989], + device='cuda:0'), covar=tensor([0.0062, 0.0102, 0.0087, 0.0125, 0.0102, 0.0177, 0.0209, 0.0070], + device='cuda:0'), in_proj_covar=tensor([0.0145, 0.0166, 0.0138, 0.0191, 0.0147, 0.0182, 0.0186, 0.0126], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 17:08:35,263 INFO [train.py:898] (0/4) Epoch 23, batch 3600, loss[loss=0.1473, simple_loss=0.2331, pruned_loss=0.0308, over 18547.00 frames. ], tot_loss[loss=0.1614, simple_loss=0.2513, pruned_loss=0.03578, over 3580373.77 frames. ], batch size: 49, lr: 4.84e-03, grad_scale: 8.0 +2023-03-09 17:08:38,649 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=83552.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:09:10,311 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-23.pt +2023-03-09 17:09:39,854 INFO [train.py:898] (0/4) Epoch 24, batch 0, loss[loss=0.1322, simple_loss=0.2144, pruned_loss=0.02504, over 17638.00 frames. ], tot_loss[loss=0.1322, simple_loss=0.2144, pruned_loss=0.02504, over 17638.00 frames. ], batch size: 39, lr: 4.74e-03, grad_scale: 8.0 +2023-03-09 17:09:39,856 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 17:09:49,002 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9126, 2.3295, 3.0281, 2.7012, 2.2885, 3.1349, 3.0507, 2.2212], + device='cuda:0'), covar=tensor([0.0620, 0.1428, 0.0596, 0.0609, 0.1605, 0.0416, 0.0806, 0.1050], + device='cuda:0'), in_proj_covar=tensor([0.0212, 0.0238, 0.0219, 0.0166, 0.0226, 0.0212, 0.0251, 0.0195], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 17:09:51,564 INFO [train.py:932] (0/4) Epoch 24, validation: loss=0.1502, simple_loss=0.2499, pruned_loss=0.02529, over 944034.00 frames. +2023-03-09 17:09:51,564 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 17:10:00,560 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.679e+02 2.611e+02 3.172e+02 4.204e+02 1.377e+03, threshold=6.343e+02, percent-clipped=6.0 +2023-03-09 17:10:34,156 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83619.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:10:50,365 INFO [train.py:898] (0/4) Epoch 24, batch 50, loss[loss=0.1924, simple_loss=0.285, pruned_loss=0.0499, over 18340.00 frames. ], tot_loss[loss=0.1582, simple_loss=0.2481, pruned_loss=0.03418, over 813714.63 frames. ], batch size: 56, lr: 4.74e-03, grad_scale: 8.0 +2023-03-09 17:10:58,585 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=83640.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:11:18,199 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.96 vs. limit=2.0 +2023-03-09 17:11:30,049 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=83667.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:11:48,510 INFO [train.py:898] (0/4) Epoch 24, batch 100, loss[loss=0.1307, simple_loss=0.2141, pruned_loss=0.02365, over 18553.00 frames. ], tot_loss[loss=0.1604, simple_loss=0.2508, pruned_loss=0.03501, over 1427881.22 frames. ], batch size: 45, lr: 4.74e-03, grad_scale: 8.0 +2023-03-09 17:11:58,110 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.623e+02 2.468e+02 2.945e+02 3.630e+02 7.467e+02, threshold=5.891e+02, percent-clipped=1.0 +2023-03-09 17:12:46,020 INFO [train.py:898] (0/4) Epoch 24, batch 150, loss[loss=0.1415, simple_loss=0.2276, pruned_loss=0.02767, over 18499.00 frames. ], tot_loss[loss=0.1591, simple_loss=0.2496, pruned_loss=0.03432, over 1915707.64 frames. ], batch size: 44, lr: 4.73e-03, grad_scale: 8.0 +2023-03-09 17:13:18,082 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83760.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:13:30,049 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.34 vs. limit=5.0 +2023-03-09 17:13:39,270 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.38 vs. limit=2.0 +2023-03-09 17:13:43,891 INFO [train.py:898] (0/4) Epoch 24, batch 200, loss[loss=0.156, simple_loss=0.256, pruned_loss=0.02798, over 18364.00 frames. ], tot_loss[loss=0.1588, simple_loss=0.249, pruned_loss=0.03426, over 2299764.75 frames. ], batch size: 55, lr: 4.73e-03, grad_scale: 4.0 +2023-03-09 17:13:53,810 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.895e+02 2.575e+02 2.989e+02 3.608e+02 5.254e+02, threshold=5.979e+02, percent-clipped=0.0 +2023-03-09 17:14:28,999 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=83821.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:14:31,441 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3902, 2.8170, 2.3617, 2.6909, 3.5132, 3.4017, 2.9194, 2.8177], + device='cuda:0'), covar=tensor([0.0203, 0.0280, 0.0642, 0.0411, 0.0190, 0.0179, 0.0422, 0.0390], + device='cuda:0'), in_proj_covar=tensor([0.0142, 0.0141, 0.0165, 0.0162, 0.0135, 0.0121, 0.0158, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 17:14:42,255 INFO [train.py:898] (0/4) Epoch 24, batch 250, loss[loss=0.2109, simple_loss=0.2915, pruned_loss=0.0652, over 12707.00 frames. ], tot_loss[loss=0.1597, simple_loss=0.25, pruned_loss=0.03469, over 2586860.80 frames. ], batch size: 129, lr: 4.73e-03, grad_scale: 4.0 +2023-03-09 17:14:56,245 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0683, 5.5314, 5.5238, 5.5428, 5.0279, 5.4231, 4.8755, 5.3848], + device='cuda:0'), covar=tensor([0.0210, 0.0244, 0.0166, 0.0384, 0.0364, 0.0211, 0.0949, 0.0298], + device='cuda:0'), in_proj_covar=tensor([0.0221, 0.0266, 0.0259, 0.0338, 0.0278, 0.0274, 0.0310, 0.0266], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 17:15:04,749 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83852.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:15:40,425 INFO [train.py:898] (0/4) Epoch 24, batch 300, loss[loss=0.1538, simple_loss=0.2497, pruned_loss=0.02897, over 18397.00 frames. ], tot_loss[loss=0.1596, simple_loss=0.2502, pruned_loss=0.03454, over 2818668.12 frames. ], batch size: 52, lr: 4.73e-03, grad_scale: 4.0 +2023-03-09 17:15:50,699 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.883e+02 2.677e+02 3.133e+02 3.610e+02 5.796e+02, threshold=6.266e+02, percent-clipped=0.0 +2023-03-09 17:16:00,987 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=83900.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:16:25,832 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8619, 3.1197, 2.7996, 3.0812, 3.9108, 3.7955, 3.3771, 3.1563], + device='cuda:0'), covar=tensor([0.0227, 0.0301, 0.0574, 0.0384, 0.0150, 0.0171, 0.0354, 0.0389], + device='cuda:0'), in_proj_covar=tensor([0.0142, 0.0141, 0.0165, 0.0162, 0.0135, 0.0121, 0.0157, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 17:16:38,566 INFO [train.py:898] (0/4) Epoch 24, batch 350, loss[loss=0.1699, simple_loss=0.2635, pruned_loss=0.03819, over 18500.00 frames. ], tot_loss[loss=0.1601, simple_loss=0.2503, pruned_loss=0.03495, over 2992324.81 frames. ], batch size: 51, lr: 4.73e-03, grad_scale: 2.0 +2023-03-09 17:16:46,852 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=83940.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:16:47,184 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.02 vs. limit=5.0 +2023-03-09 17:16:56,924 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83949.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:17:34,052 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8055, 3.5270, 5.3176, 3.2643, 4.5986, 2.5430, 3.0080, 1.9629], + device='cuda:0'), covar=tensor([0.1234, 0.1023, 0.0126, 0.0768, 0.0461, 0.2828, 0.2784, 0.2158], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0246, 0.0206, 0.0202, 0.0259, 0.0274, 0.0328, 0.0240], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 17:17:36,994 INFO [train.py:898] (0/4) Epoch 24, batch 400, loss[loss=0.1373, simple_loss=0.2188, pruned_loss=0.0279, over 18473.00 frames. ], tot_loss[loss=0.1607, simple_loss=0.2514, pruned_loss=0.03501, over 3134652.20 frames. ], batch size: 43, lr: 4.73e-03, grad_scale: 4.0 +2023-03-09 17:17:39,555 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=83985.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:17:42,759 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=83988.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:17:48,087 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.909e+02 2.479e+02 2.899e+02 3.539e+02 5.465e+02, threshold=5.798e+02, percent-clipped=0.0 +2023-03-09 17:17:56,309 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-84000.pt +2023-03-09 17:18:13,503 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84010.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 17:18:39,725 INFO [train.py:898] (0/4) Epoch 24, batch 450, loss[loss=0.1656, simple_loss=0.2582, pruned_loss=0.03646, over 18368.00 frames. ], tot_loss[loss=0.1609, simple_loss=0.2514, pruned_loss=0.03521, over 3241727.53 frames. ], batch size: 56, lr: 4.73e-03, grad_scale: 4.0 +2023-03-09 17:18:54,962 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84046.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:18:57,503 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-09 17:19:01,609 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8252, 3.8303, 3.7131, 3.3338, 3.5299, 3.0022, 3.0509, 3.8040], + device='cuda:0'), covar=tensor([0.0062, 0.0090, 0.0070, 0.0131, 0.0102, 0.0183, 0.0199, 0.0080], + device='cuda:0'), in_proj_covar=tensor([0.0146, 0.0167, 0.0139, 0.0192, 0.0148, 0.0182, 0.0188, 0.0126], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 17:19:38,436 INFO [train.py:898] (0/4) Epoch 24, batch 500, loss[loss=0.1539, simple_loss=0.2465, pruned_loss=0.03061, over 18362.00 frames. ], tot_loss[loss=0.16, simple_loss=0.2503, pruned_loss=0.03487, over 3315547.93 frames. ], batch size: 46, lr: 4.73e-03, grad_scale: 4.0 +2023-03-09 17:19:40,938 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3568, 5.3926, 5.5620, 5.5536, 5.2905, 6.1025, 5.7479, 5.4574], + device='cuda:0'), covar=tensor([0.1018, 0.0583, 0.0799, 0.0768, 0.1288, 0.0716, 0.0648, 0.1581], + device='cuda:0'), in_proj_covar=tensor([0.0364, 0.0293, 0.0315, 0.0321, 0.0331, 0.0430, 0.0290, 0.0423], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 17:19:49,795 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.687e+02 2.790e+02 3.387e+02 4.220e+02 9.031e+02, threshold=6.774e+02, percent-clipped=2.0 +2023-03-09 17:19:54,556 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84097.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:20:16,873 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84116.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:20:36,453 INFO [train.py:898] (0/4) Epoch 24, batch 550, loss[loss=0.1543, simple_loss=0.2414, pruned_loss=0.03357, over 18489.00 frames. ], tot_loss[loss=0.1603, simple_loss=0.2506, pruned_loss=0.035, over 3371113.61 frames. ], batch size: 47, lr: 4.72e-03, grad_scale: 4.0 +2023-03-09 17:21:04,954 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84158.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:21:12,478 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1324, 2.5598, 3.2711, 3.0616, 2.5102, 3.3991, 3.2203, 2.4995], + device='cuda:0'), covar=tensor([0.0521, 0.1328, 0.0498, 0.0450, 0.1413, 0.0396, 0.0760, 0.0971], + device='cuda:0'), in_proj_covar=tensor([0.0213, 0.0239, 0.0218, 0.0166, 0.0225, 0.0213, 0.0251, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 17:21:15,975 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8749, 5.3241, 2.8437, 5.1826, 5.0424, 5.3046, 5.1905, 2.6361], + device='cuda:0'), covar=tensor([0.0221, 0.0056, 0.0762, 0.0066, 0.0074, 0.0066, 0.0079, 0.0989], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0082, 0.0097, 0.0096, 0.0088, 0.0077, 0.0087, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 17:21:33,937 INFO [train.py:898] (0/4) Epoch 24, batch 600, loss[loss=0.1823, simple_loss=0.2748, pruned_loss=0.04495, over 18351.00 frames. ], tot_loss[loss=0.1606, simple_loss=0.2508, pruned_loss=0.03516, over 3398157.19 frames. ], batch size: 56, lr: 4.72e-03, grad_scale: 4.0 +2023-03-09 17:21:45,739 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.833e+02 2.553e+02 3.031e+02 3.741e+02 6.860e+02, threshold=6.062e+02, percent-clipped=1.0 +2023-03-09 17:21:53,088 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84199.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:22:32,600 INFO [train.py:898] (0/4) Epoch 24, batch 650, loss[loss=0.1781, simple_loss=0.2718, pruned_loss=0.04217, over 18279.00 frames. ], tot_loss[loss=0.161, simple_loss=0.2515, pruned_loss=0.03522, over 3434772.89 frames. ], batch size: 57, lr: 4.72e-03, grad_scale: 4.0 +2023-03-09 17:22:36,436 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84236.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:23:04,080 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84260.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:23:12,743 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4817, 5.4951, 5.7514, 5.7621, 5.4093, 6.2972, 5.9488, 5.5635], + device='cuda:0'), covar=tensor([0.1142, 0.0635, 0.0631, 0.0866, 0.1535, 0.0700, 0.0732, 0.1678], + device='cuda:0'), in_proj_covar=tensor([0.0370, 0.0297, 0.0319, 0.0324, 0.0335, 0.0434, 0.0291, 0.0425], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 17:23:30,930 INFO [train.py:898] (0/4) Epoch 24, batch 700, loss[loss=0.1485, simple_loss=0.2298, pruned_loss=0.03362, over 18421.00 frames. ], tot_loss[loss=0.1599, simple_loss=0.2502, pruned_loss=0.03478, over 3474128.57 frames. ], batch size: 43, lr: 4.72e-03, grad_scale: 4.0 +2023-03-09 17:23:42,560 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.851e+02 2.577e+02 2.948e+02 3.805e+02 8.453e+02, threshold=5.897e+02, percent-clipped=4.0 +2023-03-09 17:23:47,428 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84297.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:23:54,297 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9952, 3.3339, 4.6330, 4.0453, 3.2205, 4.9015, 4.1884, 3.4018], + device='cuda:0'), covar=tensor([0.0420, 0.1241, 0.0288, 0.0407, 0.1184, 0.0210, 0.0503, 0.0775], + device='cuda:0'), in_proj_covar=tensor([0.0215, 0.0242, 0.0220, 0.0168, 0.0227, 0.0216, 0.0254, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 17:23:56,335 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84305.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 17:24:06,599 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6757, 2.9194, 4.4446, 3.7246, 2.8328, 4.6475, 3.9523, 2.8414], + device='cuda:0'), covar=tensor([0.0523, 0.1514, 0.0247, 0.0487, 0.1559, 0.0229, 0.0578, 0.1057], + device='cuda:0'), in_proj_covar=tensor([0.0215, 0.0243, 0.0221, 0.0168, 0.0227, 0.0216, 0.0255, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 17:24:29,563 INFO [train.py:898] (0/4) Epoch 24, batch 750, loss[loss=0.1639, simple_loss=0.2559, pruned_loss=0.03589, over 18221.00 frames. ], tot_loss[loss=0.1595, simple_loss=0.2498, pruned_loss=0.03459, over 3503025.71 frames. ], batch size: 60, lr: 4.72e-03, grad_scale: 4.0 +2023-03-09 17:24:33,174 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84336.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:24:38,594 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84341.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:25:27,052 INFO [train.py:898] (0/4) Epoch 24, batch 800, loss[loss=0.1428, simple_loss=0.2337, pruned_loss=0.02597, over 18291.00 frames. ], tot_loss[loss=0.1598, simple_loss=0.2501, pruned_loss=0.03479, over 3524464.51 frames. ], batch size: 49, lr: 4.72e-03, grad_scale: 8.0 +2023-03-09 17:25:39,054 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.586e+02 2.623e+02 3.073e+02 3.761e+02 8.070e+02, threshold=6.147e+02, percent-clipped=7.0 +2023-03-09 17:25:43,749 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84397.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:26:05,549 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84416.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:26:16,419 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-09 17:26:21,953 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.15 vs. limit=2.0 +2023-03-09 17:26:25,762 INFO [train.py:898] (0/4) Epoch 24, batch 850, loss[loss=0.1787, simple_loss=0.2669, pruned_loss=0.04526, over 18344.00 frames. ], tot_loss[loss=0.1596, simple_loss=0.2496, pruned_loss=0.03479, over 3535206.48 frames. ], batch size: 56, lr: 4.72e-03, grad_scale: 8.0 +2023-03-09 17:26:49,151 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84453.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:27:01,315 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=84464.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:27:23,392 INFO [train.py:898] (0/4) Epoch 24, batch 900, loss[loss=0.1608, simple_loss=0.2501, pruned_loss=0.03579, over 18474.00 frames. ], tot_loss[loss=0.1603, simple_loss=0.2502, pruned_loss=0.03518, over 3541975.70 frames. ], batch size: 53, lr: 4.71e-03, grad_scale: 8.0 +2023-03-09 17:27:28,200 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6963, 2.6449, 2.6404, 2.4583, 2.5607, 2.2331, 2.3107, 2.6907], + device='cuda:0'), covar=tensor([0.0091, 0.0106, 0.0095, 0.0126, 0.0131, 0.0165, 0.0197, 0.0090], + device='cuda:0'), in_proj_covar=tensor([0.0147, 0.0168, 0.0139, 0.0192, 0.0150, 0.0183, 0.0188, 0.0126], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 17:27:35,099 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.772e+02 2.570e+02 3.063e+02 4.034e+02 9.733e+02, threshold=6.126e+02, percent-clipped=4.0 +2023-03-09 17:27:43,981 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.45 vs. limit=2.0 +2023-03-09 17:28:11,906 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84525.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:28:21,161 INFO [train.py:898] (0/4) Epoch 24, batch 950, loss[loss=0.1671, simple_loss=0.2565, pruned_loss=0.03881, over 17810.00 frames. ], tot_loss[loss=0.1608, simple_loss=0.251, pruned_loss=0.03529, over 3537869.41 frames. ], batch size: 70, lr: 4.71e-03, grad_scale: 8.0 +2023-03-09 17:28:46,194 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5178, 4.9221, 4.8883, 4.9293, 4.4051, 4.8040, 4.3145, 4.8150], + device='cuda:0'), covar=tensor([0.0253, 0.0302, 0.0220, 0.0387, 0.0435, 0.0232, 0.1023, 0.0309], + device='cuda:0'), in_proj_covar=tensor([0.0222, 0.0269, 0.0263, 0.0341, 0.0281, 0.0277, 0.0314, 0.0268], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 17:28:47,310 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84555.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:29:19,839 INFO [train.py:898] (0/4) Epoch 24, batch 1000, loss[loss=0.1538, simple_loss=0.2417, pruned_loss=0.03297, over 18536.00 frames. ], tot_loss[loss=0.161, simple_loss=0.2515, pruned_loss=0.03529, over 3551187.00 frames. ], batch size: 49, lr: 4.71e-03, grad_scale: 8.0 +2023-03-09 17:29:23,605 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84586.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:29:30,738 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84592.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:29:31,652 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.810e+02 2.603e+02 3.017e+02 3.476e+02 5.470e+02, threshold=6.035e+02, percent-clipped=0.0 +2023-03-09 17:29:38,160 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5498, 5.5148, 5.1154, 5.4767, 5.4140, 4.8619, 5.3631, 5.0433], + device='cuda:0'), covar=tensor([0.0398, 0.0407, 0.1340, 0.0698, 0.0580, 0.0423, 0.0448, 0.1073], + device='cuda:0'), in_proj_covar=tensor([0.0506, 0.0570, 0.0720, 0.0444, 0.0468, 0.0522, 0.0554, 0.0693], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 17:29:46,272 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84605.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:29:48,590 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6465, 3.3429, 4.5470, 4.0177, 3.0830, 2.8732, 3.9996, 4.7114], + device='cuda:0'), covar=tensor([0.0902, 0.1602, 0.0244, 0.0482, 0.1087, 0.1299, 0.0473, 0.0292], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0281, 0.0164, 0.0185, 0.0196, 0.0195, 0.0200, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 17:30:17,803 INFO [train.py:898] (0/4) Epoch 24, batch 1050, loss[loss=0.1344, simple_loss=0.2182, pruned_loss=0.02535, over 18155.00 frames. ], tot_loss[loss=0.1605, simple_loss=0.2512, pruned_loss=0.03494, over 3565180.38 frames. ], batch size: 44, lr: 4.71e-03, grad_scale: 8.0 +2023-03-09 17:30:18,235 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84633.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:30:27,864 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84641.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:30:41,759 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=84653.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:30:42,005 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2708, 3.7091, 2.4167, 3.5243, 4.4839, 2.5298, 3.4410, 3.6206], + device='cuda:0'), covar=tensor([0.0259, 0.1135, 0.1655, 0.0737, 0.0130, 0.1171, 0.0714, 0.0743], + device='cuda:0'), in_proj_covar=tensor([0.0174, 0.0272, 0.0206, 0.0200, 0.0133, 0.0184, 0.0218, 0.0226], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 17:31:16,086 INFO [train.py:898] (0/4) Epoch 24, batch 1100, loss[loss=0.2067, simple_loss=0.2799, pruned_loss=0.06678, over 12546.00 frames. ], tot_loss[loss=0.1605, simple_loss=0.2508, pruned_loss=0.0351, over 3558823.89 frames. ], batch size: 129, lr: 4.71e-03, grad_scale: 8.0 +2023-03-09 17:31:23,088 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=84689.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:31:27,254 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84692.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:31:28,854 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.674e+02 2.619e+02 3.034e+02 3.649e+02 7.124e+02, threshold=6.067e+02, percent-clipped=2.0 +2023-03-09 17:31:30,450 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84694.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:32:14,535 INFO [train.py:898] (0/4) Epoch 24, batch 1150, loss[loss=0.1693, simple_loss=0.2607, pruned_loss=0.03896, over 18260.00 frames. ], tot_loss[loss=0.1611, simple_loss=0.2515, pruned_loss=0.03533, over 3567118.75 frames. ], batch size: 60, lr: 4.71e-03, grad_scale: 8.0 +2023-03-09 17:32:32,883 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8748, 3.9289, 3.7472, 3.3598, 3.6279, 2.9867, 3.0208, 3.9203], + device='cuda:0'), covar=tensor([0.0066, 0.0078, 0.0075, 0.0138, 0.0098, 0.0189, 0.0198, 0.0064], + device='cuda:0'), in_proj_covar=tensor([0.0146, 0.0166, 0.0140, 0.0192, 0.0148, 0.0181, 0.0187, 0.0126], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 17:32:36,145 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0499, 5.4423, 2.6355, 5.2528, 5.2127, 5.4529, 5.3021, 2.9462], + device='cuda:0'), covar=tensor([0.0204, 0.0062, 0.0801, 0.0075, 0.0063, 0.0067, 0.0077, 0.0842], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0081, 0.0097, 0.0096, 0.0087, 0.0077, 0.0086, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 17:32:38,368 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84753.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:32:54,998 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.45 vs. limit=5.0 +2023-03-09 17:33:11,351 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0683, 5.1621, 5.2921, 5.4442, 5.0209, 5.9370, 5.4980, 5.1057], + device='cuda:0'), covar=tensor([0.1118, 0.0680, 0.0897, 0.0801, 0.1463, 0.0757, 0.0692, 0.1802], + device='cuda:0'), in_proj_covar=tensor([0.0368, 0.0296, 0.0318, 0.0321, 0.0333, 0.0434, 0.0290, 0.0425], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 17:33:12,317 INFO [train.py:898] (0/4) Epoch 24, batch 1200, loss[loss=0.1652, simple_loss=0.2616, pruned_loss=0.03434, over 18455.00 frames. ], tot_loss[loss=0.1615, simple_loss=0.2519, pruned_loss=0.03556, over 3575199.31 frames. ], batch size: 59, lr: 4.71e-03, grad_scale: 8.0 +2023-03-09 17:33:24,632 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.879e+02 2.822e+02 3.199e+02 4.147e+02 1.139e+03, threshold=6.397e+02, percent-clipped=6.0 +2023-03-09 17:33:32,589 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84800.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:33:33,883 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=84801.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:34:10,446 INFO [train.py:898] (0/4) Epoch 24, batch 1250, loss[loss=0.174, simple_loss=0.27, pruned_loss=0.03903, over 18367.00 frames. ], tot_loss[loss=0.1609, simple_loss=0.2513, pruned_loss=0.03527, over 3577559.18 frames. ], batch size: 50, lr: 4.70e-03, grad_scale: 4.0 +2023-03-09 17:34:36,663 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84855.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:34:39,396 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.57 vs. limit=5.0 +2023-03-09 17:34:42,527 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0125, 5.4043, 2.5432, 5.2292, 5.1649, 5.4266, 5.2722, 2.7567], + device='cuda:0'), covar=tensor([0.0202, 0.0073, 0.0827, 0.0073, 0.0070, 0.0068, 0.0080, 0.0964], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0081, 0.0096, 0.0096, 0.0087, 0.0077, 0.0085, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 17:34:43,641 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=84861.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:35:06,942 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84881.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:35:08,994 INFO [train.py:898] (0/4) Epoch 24, batch 1300, loss[loss=0.1526, simple_loss=0.2446, pruned_loss=0.03033, over 18387.00 frames. ], tot_loss[loss=0.1606, simple_loss=0.251, pruned_loss=0.0351, over 3591164.09 frames. ], batch size: 50, lr: 4.70e-03, grad_scale: 4.0 +2023-03-09 17:35:19,189 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84892.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:35:21,286 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.784e+02 2.717e+02 3.045e+02 3.603e+02 7.814e+02, threshold=6.090e+02, percent-clipped=3.0 +2023-03-09 17:35:31,907 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=84903.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:36:07,825 INFO [train.py:898] (0/4) Epoch 24, batch 1350, loss[loss=0.1799, simple_loss=0.2744, pruned_loss=0.04267, over 18569.00 frames. ], tot_loss[loss=0.1603, simple_loss=0.2509, pruned_loss=0.03482, over 3600876.25 frames. ], batch size: 54, lr: 4.70e-03, grad_scale: 4.0 +2023-03-09 17:36:16,023 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=84940.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:36:18,420 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1904, 4.1053, 3.9657, 4.0945, 4.1219, 3.6969, 4.1125, 3.9620], + device='cuda:0'), covar=tensor([0.0536, 0.0853, 0.1276, 0.0820, 0.0708, 0.0491, 0.0523, 0.1090], + device='cuda:0'), in_proj_covar=tensor([0.0512, 0.0576, 0.0724, 0.0450, 0.0474, 0.0526, 0.0560, 0.0701], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 17:36:25,249 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2201, 4.1477, 3.9680, 4.1319, 4.1606, 3.6807, 4.1329, 3.9888], + device='cuda:0'), covar=tensor([0.0542, 0.0795, 0.1348, 0.0806, 0.0680, 0.0508, 0.0567, 0.1124], + device='cuda:0'), in_proj_covar=tensor([0.0513, 0.0577, 0.0724, 0.0450, 0.0475, 0.0526, 0.0561, 0.0702], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 17:36:28,806 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=84951.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:36:36,639 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7792, 3.6187, 5.0705, 2.8701, 4.5216, 2.5706, 3.0042, 1.7123], + device='cuda:0'), covar=tensor([0.1284, 0.1023, 0.0188, 0.1026, 0.0475, 0.2717, 0.2830, 0.2397], + device='cuda:0'), in_proj_covar=tensor([0.0227, 0.0249, 0.0210, 0.0204, 0.0264, 0.0276, 0.0333, 0.0244], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 17:37:05,448 INFO [train.py:898] (0/4) Epoch 24, batch 1400, loss[loss=0.1622, simple_loss=0.2569, pruned_loss=0.03376, over 18489.00 frames. ], tot_loss[loss=0.1602, simple_loss=0.251, pruned_loss=0.03475, over 3600486.12 frames. ], batch size: 53, lr: 4.70e-03, grad_scale: 4.0 +2023-03-09 17:37:12,987 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=84989.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:37:16,420 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=84992.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:37:18,346 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.714e+02 2.377e+02 2.923e+02 3.851e+02 7.184e+02, threshold=5.846e+02, percent-clipped=2.0 +2023-03-09 17:37:39,030 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85012.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:38:03,913 INFO [train.py:898] (0/4) Epoch 24, batch 1450, loss[loss=0.1469, simple_loss=0.2365, pruned_loss=0.02863, over 18502.00 frames. ], tot_loss[loss=0.1608, simple_loss=0.2516, pruned_loss=0.03502, over 3601120.60 frames. ], batch size: 47, lr: 4.70e-03, grad_scale: 4.0 +2023-03-09 17:38:12,403 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=85040.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:38:51,807 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-09 17:39:01,459 INFO [train.py:898] (0/4) Epoch 24, batch 1500, loss[loss=0.1752, simple_loss=0.2703, pruned_loss=0.04003, over 18287.00 frames. ], tot_loss[loss=0.1608, simple_loss=0.2518, pruned_loss=0.03494, over 3605348.93 frames. ], batch size: 57, lr: 4.70e-03, grad_scale: 4.0 +2023-03-09 17:39:14,114 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.797e+02 2.574e+02 3.009e+02 3.655e+02 5.814e+02, threshold=6.018e+02, percent-clipped=0.0 +2023-03-09 17:39:22,478 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7599, 2.4111, 2.6421, 2.8772, 3.2636, 5.0121, 4.8241, 3.3406], + device='cuda:0'), covar=tensor([0.1854, 0.2450, 0.3114, 0.1801, 0.2522, 0.0205, 0.0369, 0.1055], + device='cuda:0'), in_proj_covar=tensor([0.0312, 0.0353, 0.0393, 0.0284, 0.0394, 0.0252, 0.0300, 0.0263], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 17:39:28,913 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85107.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:39:37,840 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85115.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:39:42,591 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-09 17:39:59,272 INFO [train.py:898] (0/4) Epoch 24, batch 1550, loss[loss=0.1809, simple_loss=0.2814, pruned_loss=0.04021, over 18310.00 frames. ], tot_loss[loss=0.1605, simple_loss=0.2515, pruned_loss=0.03478, over 3602791.20 frames. ], batch size: 57, lr: 4.70e-03, grad_scale: 4.0 +2023-03-09 17:40:25,778 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85156.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:40:39,404 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85168.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 17:40:48,460 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85176.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:40:55,192 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85181.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:40:57,799 INFO [train.py:898] (0/4) Epoch 24, batch 1600, loss[loss=0.1409, simple_loss=0.231, pruned_loss=0.02542, over 18418.00 frames. ], tot_loss[loss=0.1594, simple_loss=0.2501, pruned_loss=0.03441, over 3607212.67 frames. ], batch size: 48, lr: 4.69e-03, grad_scale: 8.0 +2023-03-09 17:41:01,948 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-09 17:41:10,345 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.996e+02 2.511e+02 2.791e+02 3.550e+02 5.356e+02, threshold=5.582e+02, percent-clipped=0.0 +2023-03-09 17:41:51,992 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=85229.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:41:52,222 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1451, 3.1105, 3.0433, 2.7952, 3.0099, 2.4613, 2.4819, 3.1113], + device='cuda:0'), covar=tensor([0.0081, 0.0104, 0.0091, 0.0138, 0.0099, 0.0196, 0.0214, 0.0085], + device='cuda:0'), in_proj_covar=tensor([0.0145, 0.0166, 0.0139, 0.0191, 0.0148, 0.0181, 0.0187, 0.0125], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 17:41:57,473 INFO [train.py:898] (0/4) Epoch 24, batch 1650, loss[loss=0.1474, simple_loss=0.2361, pruned_loss=0.02939, over 18423.00 frames. ], tot_loss[loss=0.1603, simple_loss=0.2512, pruned_loss=0.03471, over 3580904.89 frames. ], batch size: 48, lr: 4.69e-03, grad_scale: 8.0 +2023-03-09 17:42:09,374 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-09 17:42:55,791 INFO [train.py:898] (0/4) Epoch 24, batch 1700, loss[loss=0.1595, simple_loss=0.2537, pruned_loss=0.03262, over 18490.00 frames. ], tot_loss[loss=0.1606, simple_loss=0.2513, pruned_loss=0.03491, over 3578779.77 frames. ], batch size: 53, lr: 4.69e-03, grad_scale: 8.0 +2023-03-09 17:42:57,893 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6982, 3.5883, 4.9227, 4.2824, 3.4178, 2.9163, 4.3792, 5.1352], + device='cuda:0'), covar=tensor([0.0846, 0.1446, 0.0226, 0.0441, 0.0944, 0.1324, 0.0398, 0.0254], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0280, 0.0164, 0.0185, 0.0195, 0.0195, 0.0199, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 17:43:03,303 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85289.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:43:08,656 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.937e+02 2.815e+02 3.263e+02 3.900e+02 7.200e+02, threshold=6.526e+02, percent-clipped=2.0 +2023-03-09 17:43:23,861 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85307.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:43:41,493 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5933, 2.2633, 2.5139, 2.5614, 3.0144, 4.6138, 4.5436, 3.2409], + device='cuda:0'), covar=tensor([0.1998, 0.2595, 0.3053, 0.1989, 0.2617, 0.0298, 0.0412, 0.1042], + device='cuda:0'), in_proj_covar=tensor([0.0314, 0.0356, 0.0395, 0.0285, 0.0396, 0.0254, 0.0301, 0.0265], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 17:43:49,014 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0828, 5.4700, 5.5090, 5.5261, 4.9531, 5.4210, 4.8427, 5.4049], + device='cuda:0'), covar=tensor([0.0220, 0.0320, 0.0188, 0.0362, 0.0436, 0.0220, 0.1113, 0.0272], + device='cuda:0'), in_proj_covar=tensor([0.0221, 0.0268, 0.0262, 0.0342, 0.0279, 0.0275, 0.0312, 0.0267], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0007, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 17:43:52,176 INFO [train.py:898] (0/4) Epoch 24, batch 1750, loss[loss=0.1614, simple_loss=0.2598, pruned_loss=0.03153, over 18348.00 frames. ], tot_loss[loss=0.1608, simple_loss=0.2515, pruned_loss=0.03501, over 3596655.78 frames. ], batch size: 55, lr: 4.69e-03, grad_scale: 8.0 +2023-03-09 17:43:57,788 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=85337.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:44:22,510 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-09 17:44:48,391 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6811, 4.0233, 2.3824, 4.0266, 5.1055, 2.4928, 3.6554, 3.8741], + device='cuda:0'), covar=tensor([0.0256, 0.1381, 0.1737, 0.0681, 0.0097, 0.1305, 0.0772, 0.0806], + device='cuda:0'), in_proj_covar=tensor([0.0175, 0.0274, 0.0206, 0.0201, 0.0134, 0.0185, 0.0219, 0.0227], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 17:44:49,062 INFO [train.py:898] (0/4) Epoch 24, batch 1800, loss[loss=0.1487, simple_loss=0.246, pruned_loss=0.02572, over 18303.00 frames. ], tot_loss[loss=0.1605, simple_loss=0.2511, pruned_loss=0.03495, over 3600420.57 frames. ], batch size: 54, lr: 4.69e-03, grad_scale: 8.0 +2023-03-09 17:45:01,616 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5721, 6.0556, 5.6446, 5.8940, 5.6859, 5.5549, 6.1791, 6.1025], + device='cuda:0'), covar=tensor([0.1155, 0.0822, 0.0456, 0.0740, 0.1408, 0.0688, 0.0549, 0.0758], + device='cuda:0'), in_proj_covar=tensor([0.0624, 0.0547, 0.0391, 0.0573, 0.0768, 0.0560, 0.0782, 0.0594], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 17:45:02,451 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.840e+02 2.661e+02 3.005e+02 3.822e+02 6.057e+02, threshold=6.010e+02, percent-clipped=0.0 +2023-03-09 17:45:10,982 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7888, 4.0716, 2.3248, 4.0410, 5.1077, 2.5158, 3.6619, 3.8622], + device='cuda:0'), covar=tensor([0.0187, 0.1087, 0.1675, 0.0620, 0.0085, 0.1217, 0.0727, 0.0732], + device='cuda:0'), in_proj_covar=tensor([0.0175, 0.0275, 0.0206, 0.0202, 0.0134, 0.0186, 0.0220, 0.0228], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 17:45:27,404 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8285, 5.0830, 2.5671, 4.9340, 4.7726, 5.1023, 4.9131, 2.5455], + device='cuda:0'), covar=tensor([0.0233, 0.0073, 0.0878, 0.0102, 0.0091, 0.0085, 0.0106, 0.1081], + device='cuda:0'), in_proj_covar=tensor([0.0088, 0.0081, 0.0096, 0.0095, 0.0086, 0.0076, 0.0085, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 17:45:42,389 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85428.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:45:47,766 INFO [train.py:898] (0/4) Epoch 24, batch 1850, loss[loss=0.1299, simple_loss=0.2207, pruned_loss=0.01953, over 18360.00 frames. ], tot_loss[loss=0.1608, simple_loss=0.2513, pruned_loss=0.0351, over 3594685.47 frames. ], batch size: 42, lr: 4.69e-03, grad_scale: 8.0 +2023-03-09 17:46:05,458 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0903, 5.5370, 5.2513, 5.3674, 5.1891, 5.0323, 5.6258, 5.5566], + device='cuda:0'), covar=tensor([0.1119, 0.0793, 0.0620, 0.0734, 0.1344, 0.0737, 0.0611, 0.0743], + device='cuda:0'), in_proj_covar=tensor([0.0622, 0.0544, 0.0389, 0.0571, 0.0765, 0.0559, 0.0780, 0.0593], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 17:46:15,454 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85456.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:46:23,636 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85463.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 17:46:32,417 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85471.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:46:41,638 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.10 vs. limit=5.0 +2023-03-09 17:46:45,574 INFO [train.py:898] (0/4) Epoch 24, batch 1900, loss[loss=0.1463, simple_loss=0.2302, pruned_loss=0.03119, over 18420.00 frames. ], tot_loss[loss=0.161, simple_loss=0.2515, pruned_loss=0.03531, over 3600780.25 frames. ], batch size: 43, lr: 4.69e-03, grad_scale: 8.0 +2023-03-09 17:46:45,935 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85483.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:46:52,919 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85489.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:46:58,608 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.851e+02 2.648e+02 3.187e+02 3.710e+02 7.319e+02, threshold=6.375e+02, percent-clipped=2.0 +2023-03-09 17:47:10,740 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=85504.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:47:16,672 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-09 17:47:40,138 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85530.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:47:43,283 INFO [train.py:898] (0/4) Epoch 24, batch 1950, loss[loss=0.174, simple_loss=0.266, pruned_loss=0.04099, over 12647.00 frames. ], tot_loss[loss=0.161, simple_loss=0.2513, pruned_loss=0.03539, over 3581262.13 frames. ], batch size: 130, lr: 4.68e-03, grad_scale: 8.0 +2023-03-09 17:47:51,353 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8882, 4.6921, 4.6868, 3.5185, 3.9066, 3.5227, 2.8133, 2.8241], + device='cuda:0'), covar=tensor([0.0234, 0.0136, 0.0073, 0.0309, 0.0318, 0.0251, 0.0684, 0.0773], + device='cuda:0'), in_proj_covar=tensor([0.0072, 0.0061, 0.0065, 0.0069, 0.0092, 0.0068, 0.0078, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 17:47:56,645 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85544.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:48:18,487 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.86 vs. limit=5.0 +2023-03-09 17:48:28,461 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8612, 4.5702, 4.5603, 3.4430, 3.8607, 3.5060, 2.6212, 2.6048], + device='cuda:0'), covar=tensor([0.0224, 0.0156, 0.0081, 0.0305, 0.0332, 0.0229, 0.0761, 0.0891], + device='cuda:0'), in_proj_covar=tensor([0.0073, 0.0061, 0.0065, 0.0069, 0.0091, 0.0068, 0.0078, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 17:48:41,755 INFO [train.py:898] (0/4) Epoch 24, batch 2000, loss[loss=0.1505, simple_loss=0.2249, pruned_loss=0.03806, over 16862.00 frames. ], tot_loss[loss=0.1608, simple_loss=0.2508, pruned_loss=0.03542, over 3584905.06 frames. ], batch size: 37, lr: 4.68e-03, grad_scale: 8.0 +2023-03-09 17:48:50,862 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85591.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:48:54,264 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.6222, 6.1502, 5.6559, 6.0197, 5.7958, 5.6552, 6.2716, 6.1691], + device='cuda:0'), covar=tensor([0.1228, 0.0674, 0.0458, 0.0711, 0.1330, 0.0703, 0.0502, 0.0685], + device='cuda:0'), in_proj_covar=tensor([0.0629, 0.0547, 0.0394, 0.0575, 0.0773, 0.0567, 0.0786, 0.0600], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 17:48:55,651 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.729e+02 2.662e+02 3.281e+02 4.045e+02 7.861e+02, threshold=6.561e+02, percent-clipped=1.0 +2023-03-09 17:49:10,613 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85607.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:49:25,016 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0495, 3.9151, 5.1586, 4.6511, 3.6932, 3.2852, 4.7551, 5.4574], + device='cuda:0'), covar=tensor([0.0776, 0.1455, 0.0271, 0.0363, 0.0827, 0.1102, 0.0307, 0.0205], + device='cuda:0'), in_proj_covar=tensor([0.0155, 0.0284, 0.0167, 0.0188, 0.0198, 0.0198, 0.0201, 0.0208], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 17:49:31,495 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85625.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:49:40,100 INFO [train.py:898] (0/4) Epoch 24, batch 2050, loss[loss=0.1647, simple_loss=0.2573, pruned_loss=0.03607, over 17799.00 frames. ], tot_loss[loss=0.1606, simple_loss=0.2506, pruned_loss=0.03534, over 3577587.81 frames. ], batch size: 70, lr: 4.68e-03, grad_scale: 8.0 +2023-03-09 17:50:06,779 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=85655.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:50:08,590 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.70 vs. limit=5.0 +2023-03-09 17:50:38,499 INFO [train.py:898] (0/4) Epoch 24, batch 2100, loss[loss=0.1749, simple_loss=0.2644, pruned_loss=0.04268, over 18386.00 frames. ], tot_loss[loss=0.1608, simple_loss=0.2507, pruned_loss=0.03543, over 3576192.61 frames. ], batch size: 52, lr: 4.68e-03, grad_scale: 8.0 +2023-03-09 17:50:42,252 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85686.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 17:50:52,032 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.890e+02 2.606e+02 3.124e+02 4.074e+02 9.215e+02, threshold=6.247e+02, percent-clipped=3.0 +2023-03-09 17:51:30,720 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7279, 4.5924, 4.7821, 4.4719, 4.5008, 4.6669, 4.9079, 4.7944], + device='cuda:0'), covar=tensor([0.0106, 0.0109, 0.0115, 0.0150, 0.0104, 0.0212, 0.0096, 0.0142], + device='cuda:0'), in_proj_covar=tensor([0.0097, 0.0072, 0.0077, 0.0097, 0.0077, 0.0107, 0.0090, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 17:51:37,109 INFO [train.py:898] (0/4) Epoch 24, batch 2150, loss[loss=0.1358, simple_loss=0.2241, pruned_loss=0.02374, over 18412.00 frames. ], tot_loss[loss=0.1605, simple_loss=0.2507, pruned_loss=0.03514, over 3582568.51 frames. ], batch size: 42, lr: 4.68e-03, grad_scale: 8.0 +2023-03-09 17:51:52,280 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.77 vs. limit=5.0 +2023-03-09 17:52:12,580 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85763.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:52:17,532 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.33 vs. limit=5.0 +2023-03-09 17:52:21,488 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=85771.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:52:35,302 INFO [train.py:898] (0/4) Epoch 24, batch 2200, loss[loss=0.1494, simple_loss=0.2418, pruned_loss=0.02853, over 18237.00 frames. ], tot_loss[loss=0.161, simple_loss=0.2511, pruned_loss=0.03543, over 3581492.08 frames. ], batch size: 60, lr: 4.68e-03, grad_scale: 4.0 +2023-03-09 17:52:36,738 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85784.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:52:49,955 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.809e+02 2.630e+02 3.062e+02 3.735e+02 6.133e+02, threshold=6.125e+02, percent-clipped=0.0 +2023-03-09 17:53:08,602 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=85811.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:53:17,772 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=85819.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:53:33,939 INFO [train.py:898] (0/4) Epoch 24, batch 2250, loss[loss=0.1593, simple_loss=0.2577, pruned_loss=0.03044, over 18401.00 frames. ], tot_loss[loss=0.1605, simple_loss=0.2506, pruned_loss=0.0352, over 3589760.02 frames. ], batch size: 52, lr: 4.68e-03, grad_scale: 4.0 +2023-03-09 17:53:39,966 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8256, 3.7913, 3.6526, 3.2778, 3.5240, 2.9751, 2.8943, 3.8538], + device='cuda:0'), covar=tensor([0.0058, 0.0106, 0.0082, 0.0159, 0.0106, 0.0180, 0.0214, 0.0065], + device='cuda:0'), in_proj_covar=tensor([0.0148, 0.0168, 0.0141, 0.0194, 0.0150, 0.0183, 0.0190, 0.0126], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 17:53:40,994 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85839.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:53:53,602 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9308, 4.9542, 5.0713, 4.7504, 4.8057, 4.7805, 5.1064, 5.0732], + device='cuda:0'), covar=tensor([0.0071, 0.0064, 0.0060, 0.0111, 0.0058, 0.0163, 0.0081, 0.0099], + device='cuda:0'), in_proj_covar=tensor([0.0097, 0.0071, 0.0077, 0.0096, 0.0077, 0.0107, 0.0090, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 17:54:23,424 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85875.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:54:33,025 INFO [train.py:898] (0/4) Epoch 24, batch 2300, loss[loss=0.1596, simple_loss=0.2492, pruned_loss=0.03501, over 18551.00 frames. ], tot_loss[loss=0.1609, simple_loss=0.2511, pruned_loss=0.03535, over 3590765.59 frames. ], batch size: 49, lr: 4.68e-03, grad_scale: 4.0 +2023-03-09 17:54:36,429 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85886.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:54:47,538 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.050e+02 2.637e+02 3.133e+02 3.716e+02 1.062e+03, threshold=6.266e+02, percent-clipped=3.0 +2023-03-09 17:54:50,226 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6630, 3.6315, 3.4872, 3.1440, 3.3723, 2.7517, 2.6131, 3.7350], + device='cuda:0'), covar=tensor([0.0073, 0.0098, 0.0089, 0.0155, 0.0113, 0.0216, 0.0255, 0.0061], + device='cuda:0'), in_proj_covar=tensor([0.0148, 0.0168, 0.0142, 0.0194, 0.0150, 0.0184, 0.0190, 0.0127], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 17:54:56,813 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4683, 3.2782, 2.0929, 4.2832, 2.8736, 3.7668, 2.0270, 3.5002], + device='cuda:0'), covar=tensor([0.0609, 0.0907, 0.1572, 0.0465, 0.0945, 0.0430, 0.1492, 0.0575], + device='cuda:0'), in_proj_covar=tensor([0.0218, 0.0227, 0.0191, 0.0289, 0.0195, 0.0269, 0.0202, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 17:55:27,630 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85930.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 17:55:30,548 INFO [train.py:898] (0/4) Epoch 24, batch 2350, loss[loss=0.1707, simple_loss=0.2587, pruned_loss=0.04136, over 18484.00 frames. ], tot_loss[loss=0.1612, simple_loss=0.2515, pruned_loss=0.0355, over 3589547.75 frames. ], batch size: 51, lr: 4.67e-03, grad_scale: 4.0 +2023-03-09 17:55:34,056 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85936.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:55:53,100 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-09 17:55:58,262 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=85957.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:56:26,809 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=85981.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 17:56:28,821 INFO [train.py:898] (0/4) Epoch 24, batch 2400, loss[loss=0.151, simple_loss=0.2432, pruned_loss=0.0294, over 18303.00 frames. ], tot_loss[loss=0.1602, simple_loss=0.2503, pruned_loss=0.03498, over 3602861.88 frames. ], batch size: 49, lr: 4.67e-03, grad_scale: 8.0 +2023-03-09 17:56:38,926 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=85991.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 17:56:44,281 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.063e+02 2.665e+02 3.240e+02 3.852e+02 9.011e+02, threshold=6.481e+02, percent-clipped=2.0 +2023-03-09 17:56:49,242 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-86000.pt +2023-03-09 17:56:56,477 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86002.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:57:15,737 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86018.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:57:32,757 INFO [train.py:898] (0/4) Epoch 24, batch 2450, loss[loss=0.1774, simple_loss=0.2729, pruned_loss=0.04093, over 16074.00 frames. ], tot_loss[loss=0.1601, simple_loss=0.2503, pruned_loss=0.035, over 3591180.62 frames. ], batch size: 94, lr: 4.67e-03, grad_scale: 8.0 +2023-03-09 17:57:54,716 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86052.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:58:07,254 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86063.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:58:14,911 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86069.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:58:31,029 INFO [train.py:898] (0/4) Epoch 24, batch 2500, loss[loss=0.1369, simple_loss=0.2267, pruned_loss=0.02355, over 18508.00 frames. ], tot_loss[loss=0.1604, simple_loss=0.2506, pruned_loss=0.03511, over 3579960.99 frames. ], batch size: 44, lr: 4.67e-03, grad_scale: 4.0 +2023-03-09 17:58:32,221 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86084.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:58:46,793 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.916e+02 2.583e+02 3.100e+02 3.790e+02 6.020e+02, threshold=6.201e+02, percent-clipped=0.0 +2023-03-09 17:59:05,026 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86113.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:59:25,901 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86130.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:59:27,923 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=86132.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:59:28,882 INFO [train.py:898] (0/4) Epoch 24, batch 2550, loss[loss=0.1715, simple_loss=0.2628, pruned_loss=0.04014, over 17155.00 frames. ], tot_loss[loss=0.1608, simple_loss=0.2508, pruned_loss=0.03534, over 3561844.74 frames. ], batch size: 78, lr: 4.67e-03, grad_scale: 4.0 +2023-03-09 17:59:35,705 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86139.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 17:59:47,192 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-09 18:00:26,684 INFO [train.py:898] (0/4) Epoch 24, batch 2600, loss[loss=0.149, simple_loss=0.244, pruned_loss=0.02704, over 18377.00 frames. ], tot_loss[loss=0.1598, simple_loss=0.25, pruned_loss=0.03481, over 3572987.62 frames. ], batch size: 50, lr: 4.67e-03, grad_scale: 4.0 +2023-03-09 18:00:30,412 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86186.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:00:31,282 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=86187.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:00:42,998 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.777e+02 2.596e+02 3.106e+02 3.737e+02 7.665e+02, threshold=6.211e+02, percent-clipped=2.0 +2023-03-09 18:00:54,713 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.27 vs. limit=2.0 +2023-03-09 18:01:23,328 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86231.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:01:25,435 INFO [train.py:898] (0/4) Epoch 24, batch 2650, loss[loss=0.1533, simple_loss=0.2542, pruned_loss=0.02619, over 18325.00 frames. ], tot_loss[loss=0.1595, simple_loss=0.2498, pruned_loss=0.03457, over 3591329.96 frames. ], batch size: 54, lr: 4.67e-03, grad_scale: 4.0 +2023-03-09 18:01:26,599 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=86234.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:01:40,403 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86246.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:02:22,525 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86281.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:02:24,585 INFO [train.py:898] (0/4) Epoch 24, batch 2700, loss[loss=0.1724, simple_loss=0.2713, pruned_loss=0.03679, over 18352.00 frames. ], tot_loss[loss=0.1593, simple_loss=0.2495, pruned_loss=0.03459, over 3575332.57 frames. ], batch size: 56, lr: 4.66e-03, grad_scale: 4.0 +2023-03-09 18:02:28,062 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86286.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 18:02:41,076 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.842e+02 2.486e+02 2.896e+02 3.686e+02 7.740e+02, threshold=5.792e+02, percent-clipped=4.0 +2023-03-09 18:02:52,805 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86307.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:02:54,033 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0264, 4.2314, 2.4830, 4.1105, 5.2547, 2.5659, 3.8953, 4.1186], + device='cuda:0'), covar=tensor([0.0182, 0.1118, 0.1621, 0.0613, 0.0087, 0.1198, 0.0632, 0.0677], + device='cuda:0'), in_proj_covar=tensor([0.0175, 0.0275, 0.0207, 0.0201, 0.0134, 0.0186, 0.0219, 0.0229], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 18:02:59,446 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86313.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:03:02,964 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4706, 5.4300, 5.0733, 5.4267, 5.4159, 4.7829, 5.2994, 5.0111], + device='cuda:0'), covar=tensor([0.0438, 0.0456, 0.1382, 0.0742, 0.0579, 0.0459, 0.0443, 0.1193], + device='cuda:0'), in_proj_covar=tensor([0.0511, 0.0578, 0.0723, 0.0445, 0.0471, 0.0526, 0.0557, 0.0693], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 18:03:17,871 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.13 vs. limit=2.0 +2023-03-09 18:03:18,480 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=86329.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:03:23,301 INFO [train.py:898] (0/4) Epoch 24, batch 2750, loss[loss=0.1609, simple_loss=0.2587, pruned_loss=0.03154, over 18560.00 frames. ], tot_loss[loss=0.1598, simple_loss=0.2503, pruned_loss=0.03466, over 3579788.17 frames. ], batch size: 54, lr: 4.66e-03, grad_scale: 4.0 +2023-03-09 18:03:52,791 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86358.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:04:21,055 INFO [train.py:898] (0/4) Epoch 24, batch 2800, loss[loss=0.172, simple_loss=0.2627, pruned_loss=0.04063, over 15831.00 frames. ], tot_loss[loss=0.1595, simple_loss=0.2501, pruned_loss=0.03441, over 3582454.92 frames. ], batch size: 94, lr: 4.66e-03, grad_scale: 8.0 +2023-03-09 18:04:37,475 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.618e+02 2.577e+02 3.017e+02 3.554e+02 5.358e+02, threshold=6.034e+02, percent-clipped=0.0 +2023-03-09 18:04:48,819 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0084, 4.6541, 4.7288, 3.5246, 3.8897, 3.5711, 2.7885, 2.9261], + device='cuda:0'), covar=tensor([0.0214, 0.0158, 0.0067, 0.0304, 0.0330, 0.0221, 0.0690, 0.0728], + device='cuda:0'), in_proj_covar=tensor([0.0072, 0.0062, 0.0065, 0.0069, 0.0091, 0.0068, 0.0078, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0004, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 18:04:50,984 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86408.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:05:10,657 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86425.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:05:19,742 INFO [train.py:898] (0/4) Epoch 24, batch 2850, loss[loss=0.1551, simple_loss=0.251, pruned_loss=0.02958, over 18630.00 frames. ], tot_loss[loss=0.1594, simple_loss=0.2498, pruned_loss=0.03448, over 3581216.13 frames. ], batch size: 52, lr: 4.66e-03, grad_scale: 8.0 +2023-03-09 18:05:29,157 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86440.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:05:34,641 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86445.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:05:59,001 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.46 vs. limit=2.0 +2023-03-09 18:06:18,031 INFO [train.py:898] (0/4) Epoch 24, batch 2900, loss[loss=0.164, simple_loss=0.2456, pruned_loss=0.04118, over 18267.00 frames. ], tot_loss[loss=0.1596, simple_loss=0.25, pruned_loss=0.03462, over 3580501.88 frames. ], batch size: 45, lr: 4.66e-03, grad_scale: 4.0 +2023-03-09 18:06:22,823 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.45 vs. limit=5.0 +2023-03-09 18:06:36,142 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.815e+02 2.792e+02 3.220e+02 3.955e+02 8.779e+02, threshold=6.440e+02, percent-clipped=4.0 +2023-03-09 18:06:39,822 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86501.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:06:45,431 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86506.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 18:07:14,360 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86531.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:07:14,448 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7183, 2.8525, 2.6757, 3.0057, 3.6689, 3.6457, 3.1977, 3.0679], + device='cuda:0'), covar=tensor([0.0182, 0.0333, 0.0536, 0.0379, 0.0184, 0.0175, 0.0365, 0.0387], + device='cuda:0'), in_proj_covar=tensor([0.0143, 0.0144, 0.0166, 0.0165, 0.0138, 0.0123, 0.0160, 0.0164], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 18:07:16,258 INFO [train.py:898] (0/4) Epoch 24, batch 2950, loss[loss=0.1841, simple_loss=0.2739, pruned_loss=0.04716, over 18142.00 frames. ], tot_loss[loss=0.159, simple_loss=0.2495, pruned_loss=0.03429, over 3590236.45 frames. ], batch size: 62, lr: 4.66e-03, grad_scale: 4.0 +2023-03-09 18:07:33,488 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4600, 3.2780, 2.2842, 4.2188, 2.9281, 3.9483, 2.3854, 3.6177], + device='cuda:0'), covar=tensor([0.0686, 0.0885, 0.1360, 0.0480, 0.0882, 0.0296, 0.1165, 0.0480], + device='cuda:0'), in_proj_covar=tensor([0.0221, 0.0230, 0.0194, 0.0293, 0.0197, 0.0274, 0.0205, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 18:08:10,165 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=86579.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:08:12,254 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86580.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:08:15,185 INFO [train.py:898] (0/4) Epoch 24, batch 3000, loss[loss=0.1439, simple_loss=0.2239, pruned_loss=0.03193, over 18493.00 frames. ], tot_loss[loss=0.1591, simple_loss=0.2497, pruned_loss=0.03423, over 3599100.14 frames. ], batch size: 44, lr: 4.66e-03, grad_scale: 4.0 +2023-03-09 18:08:15,188 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 18:08:27,162 INFO [train.py:932] (0/4) Epoch 24, validation: loss=0.1501, simple_loss=0.2489, pruned_loss=0.02569, over 944034.00 frames. +2023-03-09 18:08:27,162 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 18:08:31,430 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86586.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 18:08:44,045 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-09 18:08:44,128 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-03-09 18:08:44,562 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.708e+02 2.585e+02 3.043e+02 3.671e+02 7.667e+02, threshold=6.086e+02, percent-clipped=3.0 +2023-03-09 18:08:49,600 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86602.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:09:00,636 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6387, 2.2781, 2.5527, 2.6398, 3.1656, 4.7752, 4.7168, 3.3233], + device='cuda:0'), covar=tensor([0.2031, 0.2660, 0.3119, 0.2023, 0.2610, 0.0351, 0.0358, 0.1059], + device='cuda:0'), in_proj_covar=tensor([0.0313, 0.0354, 0.0395, 0.0285, 0.0394, 0.0254, 0.0299, 0.0265], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 18:09:02,666 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86613.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:09:06,118 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0113, 5.4641, 5.4628, 5.4458, 4.9242, 5.3456, 4.8114, 5.3556], + device='cuda:0'), covar=tensor([0.0258, 0.0308, 0.0182, 0.0407, 0.0367, 0.0229, 0.1066, 0.0322], + device='cuda:0'), in_proj_covar=tensor([0.0222, 0.0269, 0.0264, 0.0344, 0.0279, 0.0275, 0.0312, 0.0267], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 18:09:26,130 INFO [train.py:898] (0/4) Epoch 24, batch 3050, loss[loss=0.1694, simple_loss=0.2655, pruned_loss=0.03665, over 18488.00 frames. ], tot_loss[loss=0.1596, simple_loss=0.25, pruned_loss=0.03462, over 3590051.04 frames. ], batch size: 51, lr: 4.66e-03, grad_scale: 4.0 +2023-03-09 18:09:27,431 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=86634.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 18:09:35,894 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86641.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:09:55,828 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86658.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:09:59,176 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=86661.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:10:24,348 INFO [train.py:898] (0/4) Epoch 24, batch 3100, loss[loss=0.1711, simple_loss=0.2635, pruned_loss=0.03938, over 18629.00 frames. ], tot_loss[loss=0.1597, simple_loss=0.25, pruned_loss=0.03466, over 3600035.39 frames. ], batch size: 52, lr: 4.65e-03, grad_scale: 4.0 +2023-03-09 18:10:41,642 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.778e+02 2.436e+02 2.944e+02 3.607e+02 6.796e+02, threshold=5.887e+02, percent-clipped=2.0 +2023-03-09 18:10:50,942 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=86706.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:10:54,043 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86708.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:11:13,477 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86725.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:11:22,864 INFO [train.py:898] (0/4) Epoch 24, batch 3150, loss[loss=0.145, simple_loss=0.2399, pruned_loss=0.02508, over 18398.00 frames. ], tot_loss[loss=0.1594, simple_loss=0.2495, pruned_loss=0.03461, over 3604571.09 frames. ], batch size: 48, lr: 4.65e-03, grad_scale: 4.0 +2023-03-09 18:11:49,564 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=86756.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:12:09,022 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=86773.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:12:13,722 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-09 18:12:20,952 INFO [train.py:898] (0/4) Epoch 24, batch 3200, loss[loss=0.1711, simple_loss=0.2624, pruned_loss=0.03993, over 18570.00 frames. ], tot_loss[loss=0.1594, simple_loss=0.2497, pruned_loss=0.03456, over 3608364.24 frames. ], batch size: 54, lr: 4.65e-03, grad_scale: 8.0 +2023-03-09 18:12:22,604 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.4938, 2.1605, 2.3210, 2.4751, 2.6934, 4.7692, 4.6557, 3.3330], + device='cuda:0'), covar=tensor([0.2395, 0.3287, 0.4040, 0.2352, 0.3932, 0.0326, 0.0421, 0.1042], + device='cuda:0'), in_proj_covar=tensor([0.0313, 0.0354, 0.0394, 0.0284, 0.0394, 0.0254, 0.0300, 0.0265], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 18:12:35,982 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86796.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:12:37,983 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.596e+02 2.507e+02 2.980e+02 3.406e+02 8.664e+02, threshold=5.960e+02, percent-clipped=4.0 +2023-03-09 18:12:42,544 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86801.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 18:12:57,923 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86814.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:13:19,217 INFO [train.py:898] (0/4) Epoch 24, batch 3250, loss[loss=0.1462, simple_loss=0.2316, pruned_loss=0.03043, over 18498.00 frames. ], tot_loss[loss=0.1591, simple_loss=0.2494, pruned_loss=0.03442, over 3603005.45 frames. ], batch size: 47, lr: 4.65e-03, grad_scale: 8.0 +2023-03-09 18:13:51,289 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8054, 5.2554, 5.2507, 5.2211, 4.7448, 5.1099, 4.6229, 5.1511], + device='cuda:0'), covar=tensor([0.0240, 0.0311, 0.0181, 0.0442, 0.0413, 0.0244, 0.1049, 0.0311], + device='cuda:0'), in_proj_covar=tensor([0.0225, 0.0273, 0.0268, 0.0350, 0.0284, 0.0279, 0.0317, 0.0272], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0006, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 18:14:08,823 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86875.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:14:12,181 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=86878.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:14:14,978 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2278, 4.1673, 3.9760, 4.1459, 4.1598, 3.7442, 4.1678, 4.0019], + device='cuda:0'), covar=tensor([0.0508, 0.0789, 0.1391, 0.0816, 0.0740, 0.0475, 0.0516, 0.1063], + device='cuda:0'), in_proj_covar=tensor([0.0513, 0.0579, 0.0725, 0.0445, 0.0476, 0.0530, 0.0562, 0.0696], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0005, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 18:14:18,045 INFO [train.py:898] (0/4) Epoch 24, batch 3300, loss[loss=0.1623, simple_loss=0.2568, pruned_loss=0.03393, over 16108.00 frames. ], tot_loss[loss=0.1589, simple_loss=0.2494, pruned_loss=0.03421, over 3605645.54 frames. ], batch size: 94, lr: 4.65e-03, grad_scale: 8.0 +2023-03-09 18:14:35,713 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.783e+02 2.678e+02 3.309e+02 3.868e+02 1.091e+03, threshold=6.618e+02, percent-clipped=4.0 +2023-03-09 18:14:40,506 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=86902.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:14:50,233 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.12 vs. limit=5.0 +2023-03-09 18:15:16,762 INFO [train.py:898] (0/4) Epoch 24, batch 3350, loss[loss=0.1609, simple_loss=0.2574, pruned_loss=0.03221, over 18484.00 frames. ], tot_loss[loss=0.1589, simple_loss=0.2492, pruned_loss=0.03434, over 3608171.54 frames. ], batch size: 53, lr: 4.65e-03, grad_scale: 8.0 +2023-03-09 18:15:20,474 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=86936.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:15:24,108 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=86939.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:15:36,562 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=86950.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:16:14,777 INFO [train.py:898] (0/4) Epoch 24, batch 3400, loss[loss=0.1614, simple_loss=0.256, pruned_loss=0.0334, over 18479.00 frames. ], tot_loss[loss=0.1593, simple_loss=0.2498, pruned_loss=0.0344, over 3603298.87 frames. ], batch size: 53, lr: 4.65e-03, grad_scale: 8.0 +2023-03-09 18:16:32,665 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.872e+02 2.525e+02 2.880e+02 3.383e+02 6.366e+02, threshold=5.759e+02, percent-clipped=0.0 +2023-03-09 18:16:38,040 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87002.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 18:16:56,933 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-09 18:17:13,560 INFO [train.py:898] (0/4) Epoch 24, batch 3450, loss[loss=0.1453, simple_loss=0.2406, pruned_loss=0.02496, over 18631.00 frames. ], tot_loss[loss=0.1593, simple_loss=0.2497, pruned_loss=0.03445, over 3603470.03 frames. ], batch size: 52, lr: 4.64e-03, grad_scale: 8.0 +2023-03-09 18:17:19,980 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3875, 5.8524, 5.4357, 5.6969, 5.4709, 5.3695, 5.9620, 5.9087], + device='cuda:0'), covar=tensor([0.1307, 0.0822, 0.0541, 0.0755, 0.1495, 0.0720, 0.0578, 0.0737], + device='cuda:0'), in_proj_covar=tensor([0.0625, 0.0549, 0.0395, 0.0573, 0.0773, 0.0564, 0.0783, 0.0597], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 18:17:48,431 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87063.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 18:18:11,251 INFO [train.py:898] (0/4) Epoch 24, batch 3500, loss[loss=0.187, simple_loss=0.2781, pruned_loss=0.04798, over 18092.00 frames. ], tot_loss[loss=0.1595, simple_loss=0.2499, pruned_loss=0.0345, over 3604331.23 frames. ], batch size: 62, lr: 4.64e-03, grad_scale: 8.0 +2023-03-09 18:18:16,195 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7270, 3.5205, 4.9590, 2.9314, 4.2781, 2.5810, 3.0041, 1.6807], + device='cuda:0'), covar=tensor([0.1270, 0.1083, 0.0176, 0.0956, 0.0590, 0.2565, 0.2656, 0.2398], + device='cuda:0'), in_proj_covar=tensor([0.0225, 0.0248, 0.0212, 0.0203, 0.0262, 0.0277, 0.0332, 0.0242], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 18:18:26,399 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87096.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:18:28,425 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.855e+02 2.643e+02 3.056e+02 3.679e+02 7.050e+02, threshold=6.112e+02, percent-clipped=3.0 +2023-03-09 18:18:32,041 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87101.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 18:18:38,544 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0261, 5.0440, 5.0652, 4.8000, 4.7953, 4.9194, 5.1861, 5.1497], + device='cuda:0'), covar=tensor([0.0067, 0.0059, 0.0061, 0.0106, 0.0061, 0.0128, 0.0064, 0.0101], + device='cuda:0'), in_proj_covar=tensor([0.0097, 0.0071, 0.0077, 0.0096, 0.0077, 0.0107, 0.0090, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 18:18:56,432 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9178, 3.5326, 2.6610, 3.3043, 4.0071, 2.6469, 3.4147, 3.4415], + device='cuda:0'), covar=tensor([0.0280, 0.0897, 0.1311, 0.0683, 0.0170, 0.1055, 0.0619, 0.0689], + device='cuda:0'), in_proj_covar=tensor([0.0173, 0.0272, 0.0204, 0.0198, 0.0133, 0.0184, 0.0217, 0.0224], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 18:19:06,666 INFO [train.py:898] (0/4) Epoch 24, batch 3550, loss[loss=0.1376, simple_loss=0.2198, pruned_loss=0.02769, over 18162.00 frames. ], tot_loss[loss=0.1595, simple_loss=0.2498, pruned_loss=0.03461, over 3593978.45 frames. ], batch size: 44, lr: 4.64e-03, grad_scale: 8.0 +2023-03-09 18:19:14,485 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7250, 2.2376, 2.5634, 2.6145, 3.1215, 4.7420, 4.5908, 3.0870], + device='cuda:0'), covar=tensor([0.1891, 0.2704, 0.3038, 0.2043, 0.2633, 0.0250, 0.0393, 0.1142], + device='cuda:0'), in_proj_covar=tensor([0.0314, 0.0355, 0.0394, 0.0284, 0.0392, 0.0253, 0.0299, 0.0265], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 18:19:18,509 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=87144.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:19:23,970 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=87149.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:19:33,928 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.41 vs. limit=2.0 +2023-03-09 18:19:46,312 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87170.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:19:53,594 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.00 vs. limit=5.0 +2023-03-09 18:20:00,496 INFO [train.py:898] (0/4) Epoch 24, batch 3600, loss[loss=0.1471, simple_loss=0.2474, pruned_loss=0.02341, over 18378.00 frames. ], tot_loss[loss=0.1595, simple_loss=0.2499, pruned_loss=0.03453, over 3602146.28 frames. ], batch size: 50, lr: 4.64e-03, grad_scale: 8.0 +2023-03-09 18:20:01,419 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-09 18:20:16,548 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.830e+02 2.465e+02 3.001e+02 3.686e+02 6.054e+02, threshold=6.002e+02, percent-clipped=0.0 +2023-03-09 18:20:36,360 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-24.pt +2023-03-09 18:21:03,019 INFO [train.py:898] (0/4) Epoch 25, batch 0, loss[loss=0.1448, simple_loss=0.233, pruned_loss=0.02832, over 18514.00 frames. ], tot_loss[loss=0.1448, simple_loss=0.233, pruned_loss=0.02832, over 18514.00 frames. ], batch size: 47, lr: 4.54e-03, grad_scale: 8.0 +2023-03-09 18:21:03,022 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 18:21:14,848 INFO [train.py:932] (0/4) Epoch 25, validation: loss=0.1499, simple_loss=0.2489, pruned_loss=0.0255, over 944034.00 frames. +2023-03-09 18:21:14,849 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 18:21:30,504 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87230.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:21:32,950 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8333, 3.6449, 5.0125, 2.9648, 4.3319, 2.5632, 3.0287, 1.7331], + device='cuda:0'), covar=tensor([0.1202, 0.0940, 0.0142, 0.0979, 0.0510, 0.2734, 0.2656, 0.2381], + device='cuda:0'), in_proj_covar=tensor([0.0225, 0.0248, 0.0213, 0.0203, 0.0262, 0.0277, 0.0331, 0.0243], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 18:21:32,982 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6342, 2.2781, 2.5992, 2.5355, 3.1686, 4.6143, 4.5753, 3.2456], + device='cuda:0'), covar=tensor([0.1943, 0.2529, 0.2933, 0.2080, 0.2451, 0.0284, 0.0402, 0.1042], + device='cuda:0'), in_proj_covar=tensor([0.0312, 0.0352, 0.0391, 0.0282, 0.0389, 0.0252, 0.0297, 0.0263], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 18:21:34,878 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87234.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:21:37,255 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87236.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:22:14,610 INFO [train.py:898] (0/4) Epoch 25, batch 50, loss[loss=0.19, simple_loss=0.2729, pruned_loss=0.05351, over 12910.00 frames. ], tot_loss[loss=0.1578, simple_loss=0.2473, pruned_loss=0.03416, over 808161.76 frames. ], batch size: 129, lr: 4.54e-03, grad_scale: 8.0 +2023-03-09 18:22:34,837 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=87284.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:22:43,610 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87291.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:22:52,239 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.879e+02 2.489e+02 2.896e+02 3.619e+02 7.206e+02, threshold=5.791e+02, percent-clipped=1.0 +2023-03-09 18:23:14,358 INFO [train.py:898] (0/4) Epoch 25, batch 100, loss[loss=0.1687, simple_loss=0.2624, pruned_loss=0.03755, over 18368.00 frames. ], tot_loss[loss=0.1572, simple_loss=0.2473, pruned_loss=0.0336, over 1432925.77 frames. ], batch size: 55, lr: 4.54e-03, grad_scale: 8.0 +2023-03-09 18:23:19,320 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7332, 4.3353, 4.3076, 3.2815, 3.5471, 3.2401, 2.5696, 2.3887], + device='cuda:0'), covar=tensor([0.0217, 0.0171, 0.0092, 0.0330, 0.0360, 0.0238, 0.0731, 0.0853], + device='cuda:0'), in_proj_covar=tensor([0.0073, 0.0062, 0.0065, 0.0070, 0.0092, 0.0068, 0.0078, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 18:23:22,043 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-09 18:23:56,881 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-03-09 18:24:03,151 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87358.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 18:24:08,333 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87362.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:24:13,626 INFO [train.py:898] (0/4) Epoch 25, batch 150, loss[loss=0.1673, simple_loss=0.2616, pruned_loss=0.03649, over 16433.00 frames. ], tot_loss[loss=0.1582, simple_loss=0.248, pruned_loss=0.03415, over 1902551.13 frames. ], batch size: 94, lr: 4.54e-03, grad_scale: 8.0 +2023-03-09 18:24:49,024 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.772e+02 2.588e+02 3.217e+02 3.788e+02 6.268e+02, threshold=6.435e+02, percent-clipped=1.0 +2023-03-09 18:25:12,509 INFO [train.py:898] (0/4) Epoch 25, batch 200, loss[loss=0.17, simple_loss=0.2702, pruned_loss=0.03496, over 18363.00 frames. ], tot_loss[loss=0.1599, simple_loss=0.2498, pruned_loss=0.035, over 2277372.48 frames. ], batch size: 55, lr: 4.54e-03, grad_scale: 8.0 +2023-03-09 18:25:19,890 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87423.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:25:32,386 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.62 vs. limit=2.0 +2023-03-09 18:25:37,029 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5020, 5.4644, 5.0499, 5.4262, 5.4246, 4.9222, 5.3529, 4.9806], + device='cuda:0'), covar=tensor([0.0500, 0.0579, 0.1502, 0.0881, 0.0698, 0.0490, 0.0524, 0.1250], + device='cuda:0'), in_proj_covar=tensor([0.0508, 0.0577, 0.0718, 0.0443, 0.0471, 0.0526, 0.0558, 0.0690], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 18:25:42,680 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8372, 4.8974, 4.9673, 4.6423, 4.7160, 4.6797, 4.9966, 4.9900], + device='cuda:0'), covar=tensor([0.0079, 0.0073, 0.0059, 0.0121, 0.0060, 0.0172, 0.0099, 0.0121], + device='cuda:0'), in_proj_covar=tensor([0.0098, 0.0072, 0.0078, 0.0096, 0.0077, 0.0107, 0.0090, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 18:25:56,362 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2681, 5.3695, 5.6031, 5.6227, 5.2224, 6.1366, 5.7544, 5.3949], + device='cuda:0'), covar=tensor([0.1156, 0.0595, 0.0790, 0.0807, 0.1355, 0.0705, 0.0591, 0.1509], + device='cuda:0'), in_proj_covar=tensor([0.0360, 0.0294, 0.0317, 0.0322, 0.0328, 0.0430, 0.0287, 0.0422], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 18:26:04,009 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87461.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:26:10,404 INFO [train.py:898] (0/4) Epoch 25, batch 250, loss[loss=0.1502, simple_loss=0.2347, pruned_loss=0.03283, over 18390.00 frames. ], tot_loss[loss=0.1595, simple_loss=0.2495, pruned_loss=0.03476, over 2575647.48 frames. ], batch size: 50, lr: 4.54e-03, grad_scale: 8.0 +2023-03-09 18:26:14,705 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87470.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:26:18,681 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-09 18:26:46,938 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.865e+02 2.586e+02 2.892e+02 3.324e+02 7.017e+02, threshold=5.784e+02, percent-clipped=1.0 +2023-03-09 18:26:55,113 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87505.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:27:09,406 INFO [train.py:898] (0/4) Epoch 25, batch 300, loss[loss=0.1507, simple_loss=0.2441, pruned_loss=0.02863, over 18494.00 frames. ], tot_loss[loss=0.1592, simple_loss=0.2496, pruned_loss=0.03435, over 2799556.48 frames. ], batch size: 51, lr: 4.54e-03, grad_scale: 8.0 +2023-03-09 18:27:10,630 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=87518.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:27:16,030 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87522.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:27:29,431 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87534.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:27:53,263 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.93 vs. limit=2.0 +2023-03-09 18:28:07,738 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87566.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 18:28:08,433 INFO [train.py:898] (0/4) Epoch 25, batch 350, loss[loss=0.1494, simple_loss=0.2312, pruned_loss=0.03383, over 17714.00 frames. ], tot_loss[loss=0.1585, simple_loss=0.2488, pruned_loss=0.03411, over 2976458.23 frames. ], batch size: 39, lr: 4.53e-03, grad_scale: 8.0 +2023-03-09 18:28:26,387 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=87582.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:28:28,915 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87584.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:28:30,947 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87586.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:28:34,502 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87589.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:28:44,690 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.762e+02 2.689e+02 3.189e+02 3.817e+02 6.961e+02, threshold=6.379e+02, percent-clipped=1.0 +2023-03-09 18:28:53,133 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=87605.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:29:06,923 INFO [train.py:898] (0/4) Epoch 25, batch 400, loss[loss=0.1546, simple_loss=0.2466, pruned_loss=0.03135, over 17745.00 frames. ], tot_loss[loss=0.1585, simple_loss=0.2488, pruned_loss=0.0341, over 3121922.17 frames. ], batch size: 70, lr: 4.53e-03, grad_scale: 8.0 +2023-03-09 18:29:22,203 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7636, 3.7646, 3.6496, 3.2887, 3.5273, 2.9154, 2.9356, 3.7480], + device='cuda:0'), covar=tensor([0.0070, 0.0088, 0.0078, 0.0121, 0.0098, 0.0188, 0.0204, 0.0078], + device='cuda:0'), in_proj_covar=tensor([0.0149, 0.0167, 0.0141, 0.0191, 0.0148, 0.0183, 0.0188, 0.0127], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 18:29:34,801 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6855, 3.4887, 4.8362, 4.2070, 3.2928, 2.9142, 4.1603, 4.9620], + device='cuda:0'), covar=tensor([0.0829, 0.1433, 0.0212, 0.0437, 0.0960, 0.1293, 0.0443, 0.0355], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0278, 0.0164, 0.0184, 0.0193, 0.0193, 0.0196, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 18:29:40,390 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87645.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 18:29:46,464 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87650.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:29:48,056 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-09 18:29:55,144 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87658.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 18:30:04,405 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=87666.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 18:30:05,653 INFO [train.py:898] (0/4) Epoch 25, batch 450, loss[loss=0.1757, simple_loss=0.2662, pruned_loss=0.04264, over 18284.00 frames. ], tot_loss[loss=0.1581, simple_loss=0.2481, pruned_loss=0.03406, over 3232101.15 frames. ], batch size: 57, lr: 4.53e-03, grad_scale: 8.0 +2023-03-09 18:30:13,275 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8359, 5.3319, 2.5546, 5.1947, 4.9686, 5.2845, 5.0596, 2.4934], + device='cuda:0'), covar=tensor([0.0251, 0.0095, 0.0922, 0.0089, 0.0084, 0.0104, 0.0131, 0.1241], + device='cuda:0'), in_proj_covar=tensor([0.0089, 0.0082, 0.0096, 0.0096, 0.0087, 0.0077, 0.0085, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 18:30:23,251 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-09 18:30:33,079 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.4881, 4.9246, 4.8900, 4.8896, 4.4488, 4.8156, 4.3658, 4.8487], + device='cuda:0'), covar=tensor([0.0280, 0.0297, 0.0218, 0.0473, 0.0357, 0.0231, 0.0941, 0.0272], + device='cuda:0'), in_proj_covar=tensor([0.0225, 0.0271, 0.0267, 0.0346, 0.0280, 0.0276, 0.0312, 0.0268], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 18:30:36,660 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6588, 3.0310, 4.4200, 3.8858, 2.9411, 4.6908, 3.9738, 2.9379], + device='cuda:0'), covar=tensor([0.0547, 0.1471, 0.0278, 0.0437, 0.1478, 0.0217, 0.0609, 0.0994], + device='cuda:0'), in_proj_covar=tensor([0.0219, 0.0245, 0.0226, 0.0170, 0.0230, 0.0220, 0.0257, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 18:30:36,681 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6890, 3.3912, 2.3955, 4.4213, 3.0831, 4.2470, 2.6200, 3.9272], + device='cuda:0'), covar=tensor([0.0613, 0.0851, 0.1369, 0.0461, 0.0911, 0.0207, 0.1077, 0.0405], + device='cuda:0'), in_proj_covar=tensor([0.0220, 0.0229, 0.0194, 0.0292, 0.0196, 0.0271, 0.0205, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 18:30:42,032 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.987e+02 2.570e+02 3.208e+02 4.155e+02 1.022e+03, threshold=6.417e+02, percent-clipped=4.0 +2023-03-09 18:30:51,742 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=87706.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 18:31:03,994 INFO [train.py:898] (0/4) Epoch 25, batch 500, loss[loss=0.1887, simple_loss=0.2737, pruned_loss=0.0519, over 18285.00 frames. ], tot_loss[loss=0.1589, simple_loss=0.2488, pruned_loss=0.03443, over 3305151.42 frames. ], batch size: 57, lr: 4.53e-03, grad_scale: 8.0 +2023-03-09 18:31:05,337 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87718.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:31:10,420 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1681, 5.6382, 5.2565, 5.4329, 5.2473, 5.0966, 5.7044, 5.6563], + device='cuda:0'), covar=tensor([0.1105, 0.0807, 0.0646, 0.0787, 0.1426, 0.0800, 0.0671, 0.0801], + device='cuda:0'), in_proj_covar=tensor([0.0624, 0.0545, 0.0394, 0.0574, 0.0770, 0.0568, 0.0781, 0.0597], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 18:31:19,627 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-09 18:31:34,954 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.71 vs. limit=2.0 +2023-03-09 18:31:37,988 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7780, 4.7702, 4.4572, 4.6946, 4.7077, 4.2035, 4.6316, 4.4334], + device='cuda:0'), covar=tensor([0.0482, 0.0485, 0.1356, 0.0787, 0.0596, 0.0481, 0.0512, 0.1125], + device='cuda:0'), in_proj_covar=tensor([0.0509, 0.0576, 0.0724, 0.0448, 0.0469, 0.0527, 0.0561, 0.0693], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 18:32:02,444 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.17 vs. limit=5.0 +2023-03-09 18:32:03,143 INFO [train.py:898] (0/4) Epoch 25, batch 550, loss[loss=0.1689, simple_loss=0.2573, pruned_loss=0.04029, over 17139.00 frames. ], tot_loss[loss=0.159, simple_loss=0.2487, pruned_loss=0.03462, over 3361614.40 frames. ], batch size: 78, lr: 4.53e-03, grad_scale: 8.0 +2023-03-09 18:32:15,773 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-09 18:32:39,325 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.934e+02 2.781e+02 3.111e+02 3.709e+02 9.396e+02, threshold=6.222e+02, percent-clipped=2.0 +2023-03-09 18:33:02,209 INFO [train.py:898] (0/4) Epoch 25, batch 600, loss[loss=0.156, simple_loss=0.2522, pruned_loss=0.02992, over 18281.00 frames. ], tot_loss[loss=0.1596, simple_loss=0.2495, pruned_loss=0.03479, over 3406438.54 frames. ], batch size: 54, lr: 4.53e-03, grad_scale: 8.0 +2023-03-09 18:33:02,382 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87817.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:33:54,508 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87861.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 18:34:01,456 INFO [train.py:898] (0/4) Epoch 25, batch 650, loss[loss=0.1318, simple_loss=0.2184, pruned_loss=0.02256, over 18428.00 frames. ], tot_loss[loss=0.1587, simple_loss=0.2482, pruned_loss=0.0346, over 3457010.12 frames. ], batch size: 43, lr: 4.53e-03, grad_scale: 8.0 +2023-03-09 18:34:21,775 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0069, 5.2391, 2.5645, 5.0778, 4.8344, 5.2083, 5.0090, 2.3776], + device='cuda:0'), covar=tensor([0.0247, 0.0128, 0.1034, 0.0128, 0.0128, 0.0157, 0.0172, 0.1592], + device='cuda:0'), in_proj_covar=tensor([0.0090, 0.0082, 0.0097, 0.0097, 0.0088, 0.0077, 0.0085, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 18:34:23,959 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=87886.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:34:38,407 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.755e+02 2.498e+02 2.822e+02 3.446e+02 7.564e+02, threshold=5.643e+02, percent-clipped=1.0 +2023-03-09 18:35:00,457 INFO [train.py:898] (0/4) Epoch 25, batch 700, loss[loss=0.148, simple_loss=0.2417, pruned_loss=0.02712, over 18254.00 frames. ], tot_loss[loss=0.1582, simple_loss=0.2481, pruned_loss=0.03417, over 3487545.15 frames. ], batch size: 47, lr: 4.53e-03, grad_scale: 8.0 +2023-03-09 18:35:04,557 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-09 18:35:19,942 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=87934.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:35:20,334 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.09 vs. limit=5.0 +2023-03-09 18:35:24,258 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5924, 3.2949, 2.2147, 4.3063, 3.0283, 4.0141, 2.3769, 3.7598], + device='cuda:0'), covar=tensor([0.0575, 0.0913, 0.1551, 0.0487, 0.0877, 0.0292, 0.1199, 0.0452], + device='cuda:0'), in_proj_covar=tensor([0.0221, 0.0230, 0.0194, 0.0293, 0.0196, 0.0271, 0.0206, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 18:35:28,952 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87940.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 18:35:34,671 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87945.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:35:52,381 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=87961.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 18:35:59,590 INFO [train.py:898] (0/4) Epoch 25, batch 750, loss[loss=0.1634, simple_loss=0.2564, pruned_loss=0.03523, over 18226.00 frames. ], tot_loss[loss=0.1578, simple_loss=0.2476, pruned_loss=0.03399, over 3511367.86 frames. ], batch size: 60, lr: 4.52e-03, grad_scale: 8.0 +2023-03-09 18:36:35,769 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.893e+02 2.719e+02 3.180e+02 3.674e+02 1.066e+03, threshold=6.360e+02, percent-clipped=6.0 +2023-03-09 18:36:38,334 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-88000.pt +2023-03-09 18:36:52,433 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-09 18:36:59,853 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5986, 2.7358, 2.3956, 2.9029, 3.5913, 3.4806, 3.0393, 2.7858], + device='cuda:0'), covar=tensor([0.0223, 0.0397, 0.0696, 0.0396, 0.0190, 0.0172, 0.0456, 0.0469], + device='cuda:0'), in_proj_covar=tensor([0.0142, 0.0143, 0.0166, 0.0164, 0.0136, 0.0123, 0.0159, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 18:37:01,709 INFO [train.py:898] (0/4) Epoch 25, batch 800, loss[loss=0.1706, simple_loss=0.2661, pruned_loss=0.03756, over 18396.00 frames. ], tot_loss[loss=0.159, simple_loss=0.2487, pruned_loss=0.03462, over 3508280.76 frames. ], batch size: 52, lr: 4.52e-03, grad_scale: 8.0 +2023-03-09 18:37:03,675 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88018.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:37:10,379 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8801, 3.4792, 4.5093, 3.8698, 2.9933, 4.8421, 4.0602, 3.2066], + device='cuda:0'), covar=tensor([0.0517, 0.1116, 0.0310, 0.0502, 0.1416, 0.0212, 0.0544, 0.0804], + device='cuda:0'), in_proj_covar=tensor([0.0219, 0.0245, 0.0228, 0.0171, 0.0229, 0.0220, 0.0257, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 18:37:58,616 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=88066.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:37:59,607 INFO [train.py:898] (0/4) Epoch 25, batch 850, loss[loss=0.163, simple_loss=0.2556, pruned_loss=0.03521, over 18282.00 frames. ], tot_loss[loss=0.1587, simple_loss=0.2483, pruned_loss=0.03458, over 3524755.06 frames. ], batch size: 57, lr: 4.52e-03, grad_scale: 8.0 +2023-03-09 18:38:25,124 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8170, 3.6927, 2.5308, 4.5335, 3.3951, 4.3958, 2.8728, 4.2313], + device='cuda:0'), covar=tensor([0.0608, 0.0748, 0.1257, 0.0520, 0.0765, 0.0296, 0.1021, 0.0362], + device='cuda:0'), in_proj_covar=tensor([0.0219, 0.0228, 0.0192, 0.0290, 0.0194, 0.0269, 0.0203, 0.0204], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 18:38:35,409 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.735e+02 2.499e+02 3.015e+02 3.591e+02 1.108e+03, threshold=6.031e+02, percent-clipped=1.0 +2023-03-09 18:38:39,188 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=88100.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:38:50,771 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.97 vs. limit=5.0 +2023-03-09 18:38:58,093 INFO [train.py:898] (0/4) Epoch 25, batch 900, loss[loss=0.1369, simple_loss=0.2184, pruned_loss=0.02775, over 18427.00 frames. ], tot_loss[loss=0.1587, simple_loss=0.2482, pruned_loss=0.0346, over 3549574.38 frames. ], batch size: 43, lr: 4.52e-03, grad_scale: 8.0 +2023-03-09 18:38:58,398 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88117.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:39:39,216 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5424, 4.9246, 4.8962, 4.9218, 4.4652, 4.8647, 4.3701, 4.8207], + device='cuda:0'), covar=tensor([0.0244, 0.0317, 0.0219, 0.0473, 0.0403, 0.0236, 0.1030, 0.0346], + device='cuda:0'), in_proj_covar=tensor([0.0227, 0.0273, 0.0270, 0.0349, 0.0283, 0.0279, 0.0315, 0.0271], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 18:39:50,461 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88161.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 18:39:50,563 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=88161.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:39:55,053 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=88165.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:39:57,056 INFO [train.py:898] (0/4) Epoch 25, batch 950, loss[loss=0.1652, simple_loss=0.2634, pruned_loss=0.03351, over 18348.00 frames. ], tot_loss[loss=0.1587, simple_loss=0.2487, pruned_loss=0.03434, over 3563350.94 frames. ], batch size: 55, lr: 4.52e-03, grad_scale: 8.0 +2023-03-09 18:40:33,072 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.747e+02 2.618e+02 3.009e+02 3.619e+02 7.498e+02, threshold=6.018e+02, percent-clipped=2.0 +2023-03-09 18:40:42,223 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6779, 3.6943, 3.5526, 3.1483, 3.3688, 2.6779, 2.7757, 3.6571], + device='cuda:0'), covar=tensor([0.0072, 0.0090, 0.0076, 0.0140, 0.0101, 0.0205, 0.0199, 0.0076], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0168, 0.0141, 0.0193, 0.0149, 0.0184, 0.0188, 0.0128], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 18:40:47,051 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=88209.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:40:55,846 INFO [train.py:898] (0/4) Epoch 25, batch 1000, loss[loss=0.1693, simple_loss=0.2609, pruned_loss=0.03888, over 18619.00 frames. ], tot_loss[loss=0.1589, simple_loss=0.2484, pruned_loss=0.03464, over 3564449.09 frames. ], batch size: 52, lr: 4.52e-03, grad_scale: 8.0 +2023-03-09 18:41:22,490 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88240.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:41:28,195 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88245.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:41:43,459 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6927, 3.9681, 2.2519, 3.9868, 5.0418, 2.4288, 3.8266, 3.9783], + device='cuda:0'), covar=tensor([0.0190, 0.1007, 0.1730, 0.0583, 0.0089, 0.1222, 0.0664, 0.0674], + device='cuda:0'), in_proj_covar=tensor([0.0178, 0.0279, 0.0208, 0.0203, 0.0136, 0.0188, 0.0223, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 18:41:48,146 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88261.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 18:41:50,368 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0252, 3.8317, 5.1267, 3.0979, 4.4499, 2.6819, 3.2172, 1.9917], + device='cuda:0'), covar=tensor([0.1067, 0.0850, 0.0146, 0.0890, 0.0558, 0.2603, 0.2652, 0.2073], + device='cuda:0'), in_proj_covar=tensor([0.0229, 0.0252, 0.0217, 0.0207, 0.0266, 0.0281, 0.0336, 0.0246], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 18:41:54,316 INFO [train.py:898] (0/4) Epoch 25, batch 1050, loss[loss=0.1654, simple_loss=0.2613, pruned_loss=0.03473, over 17114.00 frames. ], tot_loss[loss=0.1594, simple_loss=0.2492, pruned_loss=0.03477, over 3577272.62 frames. ], batch size: 78, lr: 4.52e-03, grad_scale: 8.0 +2023-03-09 18:42:18,435 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=88288.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:42:22,039 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=88291.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 18:42:24,097 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=88293.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:42:29,461 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.898e+02 2.574e+02 3.147e+02 3.590e+02 6.876e+02, threshold=6.293e+02, percent-clipped=1.0 +2023-03-09 18:42:42,897 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=88309.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:42:52,749 INFO [train.py:898] (0/4) Epoch 25, batch 1100, loss[loss=0.1634, simple_loss=0.2622, pruned_loss=0.03232, over 17734.00 frames. ], tot_loss[loss=0.1592, simple_loss=0.2491, pruned_loss=0.03464, over 3578226.29 frames. ], batch size: 70, lr: 4.52e-03, grad_scale: 8.0 +2023-03-09 18:43:32,787 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=88352.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 18:43:50,015 INFO [train.py:898] (0/4) Epoch 25, batch 1150, loss[loss=0.138, simple_loss=0.2224, pruned_loss=0.02677, over 18267.00 frames. ], tot_loss[loss=0.1587, simple_loss=0.2488, pruned_loss=0.03432, over 3594089.86 frames. ], batch size: 45, lr: 4.51e-03, grad_scale: 8.0 +2023-03-09 18:44:25,855 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.694e+02 2.731e+02 3.095e+02 3.646e+02 6.544e+02, threshold=6.189e+02, percent-clipped=1.0 +2023-03-09 18:44:26,066 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1946, 5.2845, 5.4718, 5.5856, 5.1641, 6.0473, 5.6931, 5.3759], + device='cuda:0'), covar=tensor([0.1169, 0.0630, 0.0771, 0.0777, 0.1390, 0.0673, 0.0637, 0.1543], + device='cuda:0'), in_proj_covar=tensor([0.0367, 0.0295, 0.0320, 0.0326, 0.0336, 0.0433, 0.0290, 0.0425], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 18:44:45,194 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6878, 3.3898, 4.5501, 3.9700, 3.1339, 2.9891, 4.0975, 4.7355], + device='cuda:0'), covar=tensor([0.0788, 0.1302, 0.0246, 0.0464, 0.0981, 0.1220, 0.0444, 0.0394], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0280, 0.0165, 0.0186, 0.0195, 0.0194, 0.0197, 0.0208], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 18:44:48,678 INFO [train.py:898] (0/4) Epoch 25, batch 1200, loss[loss=0.1752, simple_loss=0.2627, pruned_loss=0.04389, over 17746.00 frames. ], tot_loss[loss=0.1588, simple_loss=0.2491, pruned_loss=0.03422, over 3601393.46 frames. ], batch size: 70, lr: 4.51e-03, grad_scale: 8.0 +2023-03-09 18:45:00,546 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0168, 4.9870, 4.6637, 4.9205, 4.9180, 4.3936, 4.8591, 4.6397], + device='cuda:0'), covar=tensor([0.0418, 0.0504, 0.1381, 0.0729, 0.0598, 0.0457, 0.0456, 0.1115], + device='cuda:0'), in_proj_covar=tensor([0.0504, 0.0571, 0.0715, 0.0445, 0.0468, 0.0520, 0.0558, 0.0691], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 18:45:29,971 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.79 vs. limit=5.0 +2023-03-09 18:45:33,340 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-09 18:45:34,111 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=88456.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:45:46,401 INFO [train.py:898] (0/4) Epoch 25, batch 1250, loss[loss=0.1497, simple_loss=0.2368, pruned_loss=0.03136, over 18344.00 frames. ], tot_loss[loss=0.1588, simple_loss=0.2488, pruned_loss=0.03436, over 3582761.07 frames. ], batch size: 46, lr: 4.51e-03, grad_scale: 16.0 +2023-03-09 18:46:22,885 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.846e+02 2.624e+02 3.040e+02 3.727e+02 1.203e+03, threshold=6.079e+02, percent-clipped=2.0 +2023-03-09 18:46:44,553 INFO [train.py:898] (0/4) Epoch 25, batch 1300, loss[loss=0.152, simple_loss=0.2402, pruned_loss=0.03192, over 18242.00 frames. ], tot_loss[loss=0.1589, simple_loss=0.2491, pruned_loss=0.03431, over 3591080.20 frames. ], batch size: 47, lr: 4.51e-03, grad_scale: 16.0 +2023-03-09 18:47:00,699 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8055, 3.1211, 4.5751, 3.9034, 2.8111, 4.8507, 4.1299, 3.1160], + device='cuda:0'), covar=tensor([0.0519, 0.1446, 0.0286, 0.0468, 0.1679, 0.0208, 0.0520, 0.0953], + device='cuda:0'), in_proj_covar=tensor([0.0217, 0.0244, 0.0226, 0.0169, 0.0227, 0.0218, 0.0256, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 18:47:38,478 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-09 18:47:42,878 INFO [train.py:898] (0/4) Epoch 25, batch 1350, loss[loss=0.1561, simple_loss=0.2531, pruned_loss=0.02955, over 17930.00 frames. ], tot_loss[loss=0.1589, simple_loss=0.2493, pruned_loss=0.03427, over 3593062.87 frames. ], batch size: 65, lr: 4.51e-03, grad_scale: 16.0 +2023-03-09 18:48:19,806 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.775e+02 2.523e+02 2.984e+02 3.616e+02 6.843e+02, threshold=5.967e+02, percent-clipped=1.0 +2023-03-09 18:48:38,384 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8680, 4.0711, 2.3258, 4.0550, 5.1672, 2.7491, 3.6533, 3.7622], + device='cuda:0'), covar=tensor([0.0206, 0.1332, 0.1778, 0.0694, 0.0098, 0.1183, 0.0847, 0.0991], + device='cuda:0'), in_proj_covar=tensor([0.0178, 0.0279, 0.0208, 0.0204, 0.0137, 0.0189, 0.0222, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 18:48:41,212 INFO [train.py:898] (0/4) Epoch 25, batch 1400, loss[loss=0.1672, simple_loss=0.267, pruned_loss=0.0337, over 17733.00 frames. ], tot_loss[loss=0.159, simple_loss=0.2497, pruned_loss=0.03418, over 3597305.76 frames. ], batch size: 70, lr: 4.51e-03, grad_scale: 16.0 +2023-03-09 18:49:17,087 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=88647.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 18:49:39,309 INFO [train.py:898] (0/4) Epoch 25, batch 1450, loss[loss=0.1624, simple_loss=0.2575, pruned_loss=0.0336, over 18627.00 frames. ], tot_loss[loss=0.1591, simple_loss=0.25, pruned_loss=0.03412, over 3600503.06 frames. ], batch size: 52, lr: 4.51e-03, grad_scale: 8.0 +2023-03-09 18:50:17,346 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.835e+02 2.449e+02 2.886e+02 3.614e+02 7.165e+02, threshold=5.771e+02, percent-clipped=1.0 +2023-03-09 18:50:37,471 INFO [train.py:898] (0/4) Epoch 25, batch 1500, loss[loss=0.1515, simple_loss=0.2489, pruned_loss=0.02703, over 17940.00 frames. ], tot_loss[loss=0.1586, simple_loss=0.2495, pruned_loss=0.03384, over 3609681.24 frames. ], batch size: 65, lr: 4.51e-03, grad_scale: 8.0 +2023-03-09 18:51:08,663 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.41 vs. limit=2.0 +2023-03-09 18:51:23,355 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88756.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:51:25,944 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.03 vs. limit=5.0 +2023-03-09 18:51:35,258 INFO [train.py:898] (0/4) Epoch 25, batch 1550, loss[loss=0.1613, simple_loss=0.2588, pruned_loss=0.03193, over 18355.00 frames. ], tot_loss[loss=0.1575, simple_loss=0.2481, pruned_loss=0.03339, over 3613477.56 frames. ], batch size: 55, lr: 4.50e-03, grad_scale: 8.0 +2023-03-09 18:51:40,607 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-03-09 18:52:06,668 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-09 18:52:13,192 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.757e+02 2.636e+02 3.082e+02 3.617e+02 8.587e+02, threshold=6.165e+02, percent-clipped=5.0 +2023-03-09 18:52:20,014 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=88804.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:52:34,352 INFO [train.py:898] (0/4) Epoch 25, batch 1600, loss[loss=0.1475, simple_loss=0.2262, pruned_loss=0.03438, over 18386.00 frames. ], tot_loss[loss=0.1578, simple_loss=0.2484, pruned_loss=0.0336, over 3606320.40 frames. ], batch size: 42, lr: 4.50e-03, grad_scale: 8.0 +2023-03-09 18:52:55,327 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7311, 2.6624, 2.6258, 2.5259, 2.6039, 2.2685, 2.3927, 2.7218], + device='cuda:0'), covar=tensor([0.0096, 0.0110, 0.0092, 0.0120, 0.0109, 0.0164, 0.0180, 0.0078], + device='cuda:0'), in_proj_covar=tensor([0.0148, 0.0167, 0.0140, 0.0192, 0.0148, 0.0181, 0.0186, 0.0127], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 18:53:32,608 INFO [train.py:898] (0/4) Epoch 25, batch 1650, loss[loss=0.1435, simple_loss=0.2352, pruned_loss=0.02591, over 18279.00 frames. ], tot_loss[loss=0.1571, simple_loss=0.2476, pruned_loss=0.03332, over 3604464.49 frames. ], batch size: 49, lr: 4.50e-03, grad_scale: 8.0 +2023-03-09 18:54:09,629 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.772e+02 2.755e+02 3.165e+02 3.755e+02 1.372e+03, threshold=6.330e+02, percent-clipped=6.0 +2023-03-09 18:54:30,550 INFO [train.py:898] (0/4) Epoch 25, batch 1700, loss[loss=0.144, simple_loss=0.2291, pruned_loss=0.02946, over 18269.00 frames. ], tot_loss[loss=0.1573, simple_loss=0.2476, pruned_loss=0.03348, over 3607158.57 frames. ], batch size: 45, lr: 4.50e-03, grad_scale: 8.0 +2023-03-09 18:54:36,691 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-09 18:55:04,791 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=88947.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 18:55:28,681 INFO [train.py:898] (0/4) Epoch 25, batch 1750, loss[loss=0.1567, simple_loss=0.2504, pruned_loss=0.03149, over 17103.00 frames. ], tot_loss[loss=0.1582, simple_loss=0.2485, pruned_loss=0.03392, over 3603453.90 frames. ], batch size: 78, lr: 4.50e-03, grad_scale: 8.0 +2023-03-09 18:55:42,386 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9340, 5.3907, 3.1654, 5.2346, 5.1009, 5.3983, 5.2129, 2.8723], + device='cuda:0'), covar=tensor([0.0225, 0.0065, 0.0659, 0.0074, 0.0074, 0.0067, 0.0092, 0.0963], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0083, 0.0098, 0.0098, 0.0089, 0.0078, 0.0086, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 18:55:45,597 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8724, 3.6794, 2.5333, 4.5625, 3.3856, 4.4435, 2.8714, 4.3108], + device='cuda:0'), covar=tensor([0.0557, 0.0787, 0.1296, 0.0485, 0.0720, 0.0311, 0.1076, 0.0356], + device='cuda:0'), in_proj_covar=tensor([0.0219, 0.0230, 0.0193, 0.0292, 0.0195, 0.0269, 0.0205, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 18:56:00,686 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=88995.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 18:56:04,900 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.658e+02 2.478e+02 2.880e+02 3.518e+02 7.008e+02, threshold=5.760e+02, percent-clipped=1.0 +2023-03-09 18:56:27,121 INFO [train.py:898] (0/4) Epoch 25, batch 1800, loss[loss=0.1688, simple_loss=0.2703, pruned_loss=0.03361, over 18322.00 frames. ], tot_loss[loss=0.1585, simple_loss=0.2487, pruned_loss=0.03416, over 3582667.17 frames. ], batch size: 54, lr: 4.50e-03, grad_scale: 4.0 +2023-03-09 18:56:27,393 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.6175, 5.5665, 5.2319, 5.5217, 5.5305, 4.9381, 5.4326, 5.1492], + device='cuda:0'), covar=tensor([0.0418, 0.0397, 0.1318, 0.0789, 0.0604, 0.0422, 0.0416, 0.1012], + device='cuda:0'), in_proj_covar=tensor([0.0514, 0.0581, 0.0721, 0.0453, 0.0476, 0.0526, 0.0565, 0.0702], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 18:56:34,263 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1656, 3.1149, 3.0104, 2.7518, 2.9737, 2.4388, 2.4875, 3.1313], + device='cuda:0'), covar=tensor([0.0076, 0.0115, 0.0099, 0.0139, 0.0107, 0.0214, 0.0221, 0.0086], + device='cuda:0'), in_proj_covar=tensor([0.0147, 0.0167, 0.0139, 0.0191, 0.0147, 0.0180, 0.0186, 0.0127], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 18:56:38,797 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89027.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:57:25,676 INFO [train.py:898] (0/4) Epoch 25, batch 1850, loss[loss=0.1498, simple_loss=0.2442, pruned_loss=0.02771, over 18494.00 frames. ], tot_loss[loss=0.1585, simple_loss=0.2488, pruned_loss=0.03407, over 3581007.86 frames. ], batch size: 53, lr: 4.50e-03, grad_scale: 4.0 +2023-03-09 18:57:28,362 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7821, 2.9946, 2.7388, 3.1218, 3.8093, 3.7527, 3.3508, 3.0667], + device='cuda:0'), covar=tensor([0.0164, 0.0278, 0.0506, 0.0331, 0.0161, 0.0146, 0.0338, 0.0389], + device='cuda:0'), in_proj_covar=tensor([0.0143, 0.0143, 0.0166, 0.0165, 0.0139, 0.0123, 0.0160, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 18:57:42,949 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8140, 4.8826, 4.9800, 4.6299, 4.6088, 4.6382, 4.9886, 5.0469], + device='cuda:0'), covar=tensor([0.0068, 0.0062, 0.0053, 0.0105, 0.0074, 0.0163, 0.0072, 0.0079], + device='cuda:0'), in_proj_covar=tensor([0.0097, 0.0072, 0.0078, 0.0096, 0.0077, 0.0107, 0.0089, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 18:57:49,916 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89088.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:58:03,151 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.649e+02 2.662e+02 3.134e+02 3.843e+02 1.322e+03, threshold=6.269e+02, percent-clipped=2.0 +2023-03-09 18:58:11,345 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9055, 3.7343, 4.9881, 4.4878, 3.4343, 3.1878, 4.5249, 5.2886], + device='cuda:0'), covar=tensor([0.0763, 0.1409, 0.0249, 0.0406, 0.0937, 0.1114, 0.0362, 0.0254], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0278, 0.0164, 0.0184, 0.0194, 0.0192, 0.0197, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 18:58:15,068 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1149, 5.5132, 5.5349, 5.4993, 4.9988, 5.4079, 4.8812, 5.4374], + device='cuda:0'), covar=tensor([0.0212, 0.0233, 0.0155, 0.0406, 0.0362, 0.0244, 0.0959, 0.0268], + device='cuda:0'), in_proj_covar=tensor([0.0228, 0.0272, 0.0272, 0.0348, 0.0284, 0.0282, 0.0316, 0.0273], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 18:58:23,378 INFO [train.py:898] (0/4) Epoch 25, batch 1900, loss[loss=0.1501, simple_loss=0.2478, pruned_loss=0.0262, over 18305.00 frames. ], tot_loss[loss=0.1583, simple_loss=0.2486, pruned_loss=0.03397, over 3575657.96 frames. ], batch size: 54, lr: 4.50e-03, grad_scale: 4.0 +2023-03-09 18:59:22,368 INFO [train.py:898] (0/4) Epoch 25, batch 1950, loss[loss=0.2026, simple_loss=0.2839, pruned_loss=0.06065, over 11936.00 frames. ], tot_loss[loss=0.1591, simple_loss=0.2492, pruned_loss=0.03448, over 3561208.19 frames. ], batch size: 130, lr: 4.49e-03, grad_scale: 4.0 +2023-03-09 18:59:25,410 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89169.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 18:59:49,533 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2871, 5.8384, 5.3429, 5.6273, 5.4163, 5.2822, 5.9013, 5.8874], + device='cuda:0'), covar=tensor([0.1324, 0.0810, 0.0587, 0.0771, 0.1533, 0.0828, 0.0681, 0.0722], + device='cuda:0'), in_proj_covar=tensor([0.0622, 0.0542, 0.0395, 0.0574, 0.0766, 0.0565, 0.0781, 0.0591], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 18:59:50,789 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6316, 2.2963, 2.5506, 2.6869, 2.9788, 4.6419, 4.5096, 3.1674], + device='cuda:0'), covar=tensor([0.2016, 0.2626, 0.3107, 0.1976, 0.2674, 0.0297, 0.0408, 0.1116], + device='cuda:0'), in_proj_covar=tensor([0.0319, 0.0358, 0.0400, 0.0288, 0.0395, 0.0257, 0.0302, 0.0267], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 19:00:00,686 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.906e+02 2.686e+02 3.109e+02 3.749e+02 7.120e+02, threshold=6.217e+02, percent-clipped=3.0 +2023-03-09 19:00:20,454 INFO [train.py:898] (0/4) Epoch 25, batch 2000, loss[loss=0.1776, simple_loss=0.2705, pruned_loss=0.04231, over 18492.00 frames. ], tot_loss[loss=0.1591, simple_loss=0.2493, pruned_loss=0.03442, over 3576027.14 frames. ], batch size: 51, lr: 4.49e-03, grad_scale: 8.0 +2023-03-09 19:00:36,490 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89230.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:01:18,543 INFO [train.py:898] (0/4) Epoch 25, batch 2050, loss[loss=0.1503, simple_loss=0.2322, pruned_loss=0.03426, over 18257.00 frames. ], tot_loss[loss=0.1594, simple_loss=0.2499, pruned_loss=0.03443, over 3585497.38 frames. ], batch size: 45, lr: 4.49e-03, grad_scale: 8.0 +2023-03-09 19:01:51,629 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3602, 5.2210, 5.6345, 5.6049, 5.2927, 6.1262, 5.7394, 5.4520], + device='cuda:0'), covar=tensor([0.1068, 0.0670, 0.0649, 0.0645, 0.1406, 0.0680, 0.0603, 0.1578], + device='cuda:0'), in_proj_covar=tensor([0.0363, 0.0295, 0.0318, 0.0324, 0.0335, 0.0434, 0.0288, 0.0423], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 19:01:57,830 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.761e+02 2.663e+02 3.151e+02 3.828e+02 7.716e+02, threshold=6.301e+02, percent-clipped=2.0 +2023-03-09 19:02:17,077 INFO [train.py:898] (0/4) Epoch 25, batch 2100, loss[loss=0.1711, simple_loss=0.2639, pruned_loss=0.0392, over 17824.00 frames. ], tot_loss[loss=0.1594, simple_loss=0.2498, pruned_loss=0.03446, over 3592897.56 frames. ], batch size: 70, lr: 4.49e-03, grad_scale: 8.0 +2023-03-09 19:02:38,604 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-09 19:02:59,019 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4414, 5.3956, 5.0431, 5.3266, 5.3344, 4.8029, 5.2625, 4.9485], + device='cuda:0'), covar=tensor([0.0411, 0.0432, 0.1319, 0.0815, 0.0501, 0.0435, 0.0416, 0.1192], + device='cuda:0'), in_proj_covar=tensor([0.0512, 0.0578, 0.0721, 0.0453, 0.0474, 0.0525, 0.0565, 0.0699], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 19:03:10,182 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7782, 3.1549, 2.5460, 3.1232, 3.8482, 3.6917, 3.3744, 3.2959], + device='cuda:0'), covar=tensor([0.0226, 0.0264, 0.0641, 0.0306, 0.0159, 0.0163, 0.0319, 0.0282], + device='cuda:0'), in_proj_covar=tensor([0.0144, 0.0142, 0.0165, 0.0163, 0.0137, 0.0123, 0.0159, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 19:03:15,258 INFO [train.py:898] (0/4) Epoch 25, batch 2150, loss[loss=0.1527, simple_loss=0.2372, pruned_loss=0.03409, over 18419.00 frames. ], tot_loss[loss=0.1595, simple_loss=0.2501, pruned_loss=0.03446, over 3585044.98 frames. ], batch size: 48, lr: 4.49e-03, grad_scale: 8.0 +2023-03-09 19:03:24,295 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89374.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:03:35,640 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89383.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:03:54,951 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.788e+02 2.675e+02 3.172e+02 3.693e+02 8.174e+02, threshold=6.344e+02, percent-clipped=3.0 +2023-03-09 19:04:15,088 INFO [train.py:898] (0/4) Epoch 25, batch 2200, loss[loss=0.1679, simple_loss=0.26, pruned_loss=0.0379, over 18579.00 frames. ], tot_loss[loss=0.1591, simple_loss=0.2495, pruned_loss=0.03439, over 3588619.15 frames. ], batch size: 54, lr: 4.49e-03, grad_scale: 8.0 +2023-03-09 19:04:25,101 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89425.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:04:36,862 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89435.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:04:41,855 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89439.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 19:05:13,786 INFO [train.py:898] (0/4) Epoch 25, batch 2250, loss[loss=0.1586, simple_loss=0.2518, pruned_loss=0.03271, over 18364.00 frames. ], tot_loss[loss=0.159, simple_loss=0.2491, pruned_loss=0.03441, over 3586574.74 frames. ], batch size: 50, lr: 4.49e-03, grad_scale: 8.0 +2023-03-09 19:05:27,771 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7270, 4.0249, 2.3350, 3.9935, 5.0059, 2.6605, 3.7583, 3.8805], + device='cuda:0'), covar=tensor([0.0197, 0.1057, 0.1623, 0.0610, 0.0098, 0.1121, 0.0661, 0.0696], + device='cuda:0'), in_proj_covar=tensor([0.0176, 0.0274, 0.0205, 0.0199, 0.0135, 0.0184, 0.0219, 0.0227], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0002, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 19:05:28,922 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1360, 5.1565, 5.2210, 4.9479, 4.9556, 4.9677, 5.2886, 5.3224], + device='cuda:0'), covar=tensor([0.0061, 0.0061, 0.0052, 0.0104, 0.0055, 0.0157, 0.0079, 0.0101], + device='cuda:0'), in_proj_covar=tensor([0.0099, 0.0073, 0.0079, 0.0098, 0.0078, 0.0108, 0.0091, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 19:05:38,325 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89486.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:05:53,609 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.616e+02 2.577e+02 2.992e+02 3.589e+02 8.023e+02, threshold=5.985e+02, percent-clipped=1.0 +2023-03-09 19:05:53,877 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89500.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 19:06:04,245 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89509.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:06:12,817 INFO [train.py:898] (0/4) Epoch 25, batch 2300, loss[loss=0.1531, simple_loss=0.2314, pruned_loss=0.0374, over 18461.00 frames. ], tot_loss[loss=0.1591, simple_loss=0.2494, pruned_loss=0.03444, over 3583860.87 frames. ], batch size: 43, lr: 4.49e-03, grad_scale: 8.0 +2023-03-09 19:06:22,134 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89525.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:06:40,891 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0331, 3.7177, 5.1434, 2.9251, 4.5629, 2.6352, 3.0743, 1.8703], + device='cuda:0'), covar=tensor([0.1105, 0.0938, 0.0190, 0.1001, 0.0474, 0.2652, 0.2941, 0.2300], + device='cuda:0'), in_proj_covar=tensor([0.0226, 0.0249, 0.0214, 0.0203, 0.0261, 0.0277, 0.0332, 0.0243], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 19:06:53,686 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-09 19:07:11,312 INFO [train.py:898] (0/4) Epoch 25, batch 2350, loss[loss=0.174, simple_loss=0.266, pruned_loss=0.04101, over 17767.00 frames. ], tot_loss[loss=0.1592, simple_loss=0.2495, pruned_loss=0.03444, over 3566349.56 frames. ], batch size: 70, lr: 4.48e-03, grad_scale: 8.0 +2023-03-09 19:07:15,040 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89570.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:07:51,625 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.602e+02 2.634e+02 3.010e+02 3.571e+02 1.034e+03, threshold=6.019e+02, percent-clipped=2.0 +2023-03-09 19:08:09,864 INFO [train.py:898] (0/4) Epoch 25, batch 2400, loss[loss=0.1601, simple_loss=0.2496, pruned_loss=0.03535, over 18332.00 frames. ], tot_loss[loss=0.1593, simple_loss=0.2497, pruned_loss=0.03442, over 3574117.36 frames. ], batch size: 46, lr: 4.48e-03, grad_scale: 8.0 +2023-03-09 19:08:46,706 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8666, 3.8905, 5.0908, 2.9834, 4.5353, 2.6017, 3.0813, 1.8524], + device='cuda:0'), covar=tensor([0.1231, 0.0893, 0.0196, 0.0960, 0.0548, 0.2747, 0.2844, 0.2349], + device='cuda:0'), in_proj_covar=tensor([0.0227, 0.0249, 0.0216, 0.0204, 0.0262, 0.0278, 0.0333, 0.0244], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 19:09:07,612 INFO [train.py:898] (0/4) Epoch 25, batch 2450, loss[loss=0.1796, simple_loss=0.2746, pruned_loss=0.0423, over 18373.00 frames. ], tot_loss[loss=0.1592, simple_loss=0.2496, pruned_loss=0.03439, over 3578179.47 frames. ], batch size: 52, lr: 4.48e-03, grad_scale: 4.0 +2023-03-09 19:09:23,406 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89680.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:09:25,913 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7311, 3.0333, 4.5547, 3.7557, 2.8746, 4.7389, 3.9820, 3.0978], + device='cuda:0'), covar=tensor([0.0482, 0.1397, 0.0221, 0.0468, 0.1549, 0.0207, 0.0539, 0.0916], + device='cuda:0'), in_proj_covar=tensor([0.0219, 0.0245, 0.0228, 0.0170, 0.0228, 0.0220, 0.0257, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 19:09:26,939 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=89683.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:09:49,621 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.687e+02 2.547e+02 3.016e+02 3.415e+02 5.579e+02, threshold=6.031e+02, percent-clipped=0.0 +2023-03-09 19:10:02,345 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0493, 5.5305, 5.5293, 5.5288, 4.9710, 5.4200, 4.9427, 5.4119], + device='cuda:0'), covar=tensor([0.0223, 0.0223, 0.0169, 0.0366, 0.0378, 0.0199, 0.0884, 0.0285], + device='cuda:0'), in_proj_covar=tensor([0.0230, 0.0275, 0.0275, 0.0351, 0.0286, 0.0284, 0.0316, 0.0276], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 19:10:06,502 INFO [train.py:898] (0/4) Epoch 25, batch 2500, loss[loss=0.1549, simple_loss=0.2421, pruned_loss=0.03389, over 18306.00 frames. ], tot_loss[loss=0.1593, simple_loss=0.2498, pruned_loss=0.03438, over 3581569.32 frames. ], batch size: 49, lr: 4.48e-03, grad_scale: 4.0 +2023-03-09 19:10:21,654 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89730.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:10:22,817 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=89731.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:10:34,995 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89741.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:10:52,997 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-09 19:11:04,383 INFO [train.py:898] (0/4) Epoch 25, batch 2550, loss[loss=0.1669, simple_loss=0.2623, pruned_loss=0.03573, over 18490.00 frames. ], tot_loss[loss=0.1593, simple_loss=0.2497, pruned_loss=0.03448, over 3590409.23 frames. ], batch size: 51, lr: 4.48e-03, grad_scale: 4.0 +2023-03-09 19:11:20,624 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89781.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:11:37,076 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89795.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 19:11:45,425 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.598e+02 2.460e+02 2.989e+02 3.597e+02 5.800e+02, threshold=5.978e+02, percent-clipped=0.0 +2023-03-09 19:12:03,016 INFO [train.py:898] (0/4) Epoch 25, batch 2600, loss[loss=0.17, simple_loss=0.2714, pruned_loss=0.03428, over 18499.00 frames. ], tot_loss[loss=0.1597, simple_loss=0.25, pruned_loss=0.03468, over 3573336.14 frames. ], batch size: 53, lr: 4.48e-03, grad_scale: 4.0 +2023-03-09 19:12:12,956 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=89825.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:12:19,297 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.70 vs. limit=2.0 +2023-03-09 19:12:59,239 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=89865.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:12:59,382 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=89865.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:13:01,342 INFO [train.py:898] (0/4) Epoch 25, batch 2650, loss[loss=0.1286, simple_loss=0.207, pruned_loss=0.02513, over 18416.00 frames. ], tot_loss[loss=0.1598, simple_loss=0.2501, pruned_loss=0.03469, over 3574227.64 frames. ], batch size: 43, lr: 4.48e-03, grad_scale: 4.0 +2023-03-09 19:13:08,332 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=89873.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:13:41,748 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.822e+02 2.660e+02 3.253e+02 3.912e+02 1.582e+03, threshold=6.506e+02, percent-clipped=3.0 +2023-03-09 19:13:59,663 INFO [train.py:898] (0/4) Epoch 25, batch 2700, loss[loss=0.1935, simple_loss=0.2793, pruned_loss=0.05385, over 13087.00 frames. ], tot_loss[loss=0.1598, simple_loss=0.25, pruned_loss=0.03478, over 3574769.69 frames. ], batch size: 130, lr: 4.48e-03, grad_scale: 4.0 +2023-03-09 19:14:04,525 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7854, 5.2721, 5.2642, 5.2028, 4.7267, 5.1343, 4.6746, 5.1375], + device='cuda:0'), covar=tensor([0.0294, 0.0278, 0.0177, 0.0446, 0.0417, 0.0247, 0.1007, 0.0331], + device='cuda:0'), in_proj_covar=tensor([0.0229, 0.0275, 0.0273, 0.0351, 0.0285, 0.0281, 0.0316, 0.0275], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 19:14:08,492 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.36 vs. limit=2.0 +2023-03-09 19:14:09,981 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=89926.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 19:14:24,806 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0223, 3.8159, 5.1838, 4.6579, 3.5862, 3.2674, 4.7037, 5.4116], + device='cuda:0'), covar=tensor([0.0734, 0.1459, 0.0184, 0.0346, 0.0844, 0.1051, 0.0295, 0.0236], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0283, 0.0167, 0.0187, 0.0197, 0.0195, 0.0200, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 19:14:58,044 INFO [train.py:898] (0/4) Epoch 25, batch 2750, loss[loss=0.1661, simple_loss=0.2596, pruned_loss=0.03628, over 18334.00 frames. ], tot_loss[loss=0.1594, simple_loss=0.2495, pruned_loss=0.03461, over 3570930.35 frames. ], batch size: 55, lr: 4.47e-03, grad_scale: 4.0 +2023-03-09 19:15:07,254 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3687, 3.2660, 2.2306, 4.0212, 2.8385, 3.6755, 2.3955, 3.5205], + device='cuda:0'), covar=tensor([0.0633, 0.0860, 0.1416, 0.0507, 0.0882, 0.0369, 0.1184, 0.0464], + device='cuda:0'), in_proj_covar=tensor([0.0220, 0.0229, 0.0193, 0.0291, 0.0195, 0.0269, 0.0204, 0.0205], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 19:15:36,637 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-90000.pt +2023-03-09 19:15:43,121 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.861e+02 2.579e+02 2.972e+02 3.453e+02 7.499e+02, threshold=5.944e+02, percent-clipped=1.0 +2023-03-09 19:16:00,865 INFO [train.py:898] (0/4) Epoch 25, batch 2800, loss[loss=0.1383, simple_loss=0.2287, pruned_loss=0.02394, over 18358.00 frames. ], tot_loss[loss=0.1587, simple_loss=0.249, pruned_loss=0.03419, over 3573021.57 frames. ], batch size: 46, lr: 4.47e-03, grad_scale: 8.0 +2023-03-09 19:16:04,606 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3571, 4.3853, 4.4566, 4.1970, 4.2495, 4.2028, 4.4620, 4.5159], + device='cuda:0'), covar=tensor([0.0087, 0.0079, 0.0069, 0.0124, 0.0072, 0.0187, 0.0096, 0.0100], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0074, 0.0079, 0.0098, 0.0078, 0.0109, 0.0091, 0.0090], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 19:16:15,883 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90030.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:16:22,531 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90036.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:16:39,180 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-03-09 19:16:57,953 INFO [train.py:898] (0/4) Epoch 25, batch 2850, loss[loss=0.1959, simple_loss=0.2845, pruned_loss=0.05364, over 18029.00 frames. ], tot_loss[loss=0.159, simple_loss=0.2494, pruned_loss=0.03433, over 3586286.92 frames. ], batch size: 65, lr: 4.47e-03, grad_scale: 8.0 +2023-03-09 19:17:11,249 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=90078.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:17:12,848 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.99 vs. limit=5.0 +2023-03-09 19:17:14,752 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90081.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:17:30,983 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90095.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 19:17:38,643 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.912e+02 2.603e+02 3.094e+02 3.729e+02 6.556e+02, threshold=6.188e+02, percent-clipped=2.0 +2023-03-09 19:17:56,099 INFO [train.py:898] (0/4) Epoch 25, batch 2900, loss[loss=0.1658, simple_loss=0.2596, pruned_loss=0.03604, over 18497.00 frames. ], tot_loss[loss=0.159, simple_loss=0.2495, pruned_loss=0.03424, over 3592497.22 frames. ], batch size: 51, lr: 4.47e-03, grad_scale: 8.0 +2023-03-09 19:18:10,697 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=90129.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:18:11,996 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90130.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:18:26,753 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=90143.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 19:18:52,949 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90165.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:18:54,881 INFO [train.py:898] (0/4) Epoch 25, batch 2950, loss[loss=0.1486, simple_loss=0.241, pruned_loss=0.02809, over 18487.00 frames. ], tot_loss[loss=0.1587, simple_loss=0.249, pruned_loss=0.03417, over 3586426.35 frames. ], batch size: 51, lr: 4.47e-03, grad_scale: 8.0 +2023-03-09 19:18:59,789 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.4666, 2.3057, 2.4078, 2.5806, 3.0865, 4.7493, 4.6564, 3.2375], + device='cuda:0'), covar=tensor([0.2212, 0.2751, 0.3420, 0.2068, 0.2678, 0.0270, 0.0404, 0.1082], + device='cuda:0'), in_proj_covar=tensor([0.0319, 0.0359, 0.0402, 0.0289, 0.0396, 0.0259, 0.0302, 0.0268], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 19:19:04,415 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.57 vs. limit=5.0 +2023-03-09 19:19:23,412 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90191.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:19:36,171 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.680e+02 2.539e+02 2.924e+02 3.398e+02 8.084e+02, threshold=5.847e+02, percent-clipped=3.0 +2023-03-09 19:19:36,924 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-09 19:19:48,871 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=90213.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:19:53,270 INFO [train.py:898] (0/4) Epoch 25, batch 3000, loss[loss=0.1434, simple_loss=0.2205, pruned_loss=0.03312, over 17698.00 frames. ], tot_loss[loss=0.1585, simple_loss=0.2489, pruned_loss=0.03408, over 3596304.18 frames. ], batch size: 39, lr: 4.47e-03, grad_scale: 8.0 +2023-03-09 19:19:53,272 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 19:20:05,345 INFO [train.py:932] (0/4) Epoch 25, validation: loss=0.1501, simple_loss=0.2485, pruned_loss=0.02584, over 944034.00 frames. +2023-03-09 19:20:05,346 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 19:20:09,989 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90221.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 19:20:22,481 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90232.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:21:03,798 INFO [train.py:898] (0/4) Epoch 25, batch 3050, loss[loss=0.1812, simple_loss=0.2688, pruned_loss=0.04679, over 18138.00 frames. ], tot_loss[loss=0.1588, simple_loss=0.2493, pruned_loss=0.03417, over 3597151.22 frames. ], batch size: 62, lr: 4.47e-03, grad_scale: 8.0 +2023-03-09 19:21:22,252 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3938, 2.7542, 2.4380, 2.7883, 3.5132, 3.4204, 3.0574, 2.8222], + device='cuda:0'), covar=tensor([0.0188, 0.0280, 0.0563, 0.0356, 0.0186, 0.0159, 0.0372, 0.0352], + device='cuda:0'), in_proj_covar=tensor([0.0142, 0.0142, 0.0165, 0.0163, 0.0139, 0.0122, 0.0159, 0.0161], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 19:21:34,106 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90293.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:21:41,487 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9292, 5.3790, 5.3461, 5.3488, 4.8998, 5.2767, 4.7262, 5.2646], + device='cuda:0'), covar=tensor([0.0231, 0.0251, 0.0176, 0.0427, 0.0394, 0.0219, 0.1056, 0.0321], + device='cuda:0'), in_proj_covar=tensor([0.0226, 0.0273, 0.0270, 0.0348, 0.0283, 0.0280, 0.0314, 0.0272], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 19:21:44,552 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.813e+02 2.699e+02 3.154e+02 3.887e+02 1.496e+03, threshold=6.309e+02, percent-clipped=8.0 +2023-03-09 19:22:02,750 INFO [train.py:898] (0/4) Epoch 25, batch 3100, loss[loss=0.1598, simple_loss=0.2475, pruned_loss=0.03601, over 18400.00 frames. ], tot_loss[loss=0.1599, simple_loss=0.2505, pruned_loss=0.03465, over 3597744.70 frames. ], batch size: 48, lr: 4.47e-03, grad_scale: 8.0 +2023-03-09 19:22:24,065 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90336.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:22:42,232 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90351.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:23:00,970 INFO [train.py:898] (0/4) Epoch 25, batch 3150, loss[loss=0.1548, simple_loss=0.2505, pruned_loss=0.0295, over 18494.00 frames. ], tot_loss[loss=0.1602, simple_loss=0.251, pruned_loss=0.03475, over 3576729.25 frames. ], batch size: 51, lr: 4.46e-03, grad_scale: 4.0 +2023-03-09 19:23:20,279 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=90384.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:23:42,933 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.817e+02 2.713e+02 3.340e+02 4.151e+02 5.769e+02, threshold=6.681e+02, percent-clipped=0.0 +2023-03-09 19:23:53,113 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90412.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:23:59,428 INFO [train.py:898] (0/4) Epoch 25, batch 3200, loss[loss=0.179, simple_loss=0.2729, pruned_loss=0.04252, over 18361.00 frames. ], tot_loss[loss=0.1599, simple_loss=0.2504, pruned_loss=0.03467, over 3577342.22 frames. ], batch size: 56, lr: 4.46e-03, grad_scale: 8.0 +2023-03-09 19:24:18,773 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2334, 5.2108, 5.5641, 5.5024, 5.1361, 5.9443, 5.6683, 5.2564], + device='cuda:0'), covar=tensor([0.1204, 0.0673, 0.0706, 0.0973, 0.1365, 0.0796, 0.0687, 0.1643], + device='cuda:0'), in_proj_covar=tensor([0.0367, 0.0298, 0.0325, 0.0327, 0.0338, 0.0440, 0.0293, 0.0431], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 19:24:48,282 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90459.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:24:57,502 INFO [train.py:898] (0/4) Epoch 25, batch 3250, loss[loss=0.1444, simple_loss=0.2363, pruned_loss=0.02627, over 18388.00 frames. ], tot_loss[loss=0.1601, simple_loss=0.2504, pruned_loss=0.03488, over 3564673.49 frames. ], batch size: 50, lr: 4.46e-03, grad_scale: 8.0 +2023-03-09 19:25:19,133 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90486.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:25:31,107 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90496.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:25:38,661 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.740e+02 2.752e+02 3.132e+02 3.659e+02 1.386e+03, threshold=6.263e+02, percent-clipped=2.0 +2023-03-09 19:25:54,832 INFO [train.py:898] (0/4) Epoch 25, batch 3300, loss[loss=0.1368, simple_loss=0.2228, pruned_loss=0.02541, over 18495.00 frames. ], tot_loss[loss=0.1592, simple_loss=0.2493, pruned_loss=0.03457, over 3578522.86 frames. ], batch size: 44, lr: 4.46e-03, grad_scale: 8.0 +2023-03-09 19:25:58,871 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.47 vs. limit=2.0 +2023-03-09 19:25:59,729 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90520.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:26:00,800 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90521.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 19:26:42,181 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90557.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:26:52,886 INFO [train.py:898] (0/4) Epoch 25, batch 3350, loss[loss=0.1535, simple_loss=0.242, pruned_loss=0.0325, over 18421.00 frames. ], tot_loss[loss=0.1601, simple_loss=0.2506, pruned_loss=0.03477, over 3568218.57 frames. ], batch size: 48, lr: 4.46e-03, grad_scale: 8.0 +2023-03-09 19:26:55,352 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=90569.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:27:17,325 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90588.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:27:34,779 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.869e+02 2.603e+02 3.023e+02 3.526e+02 7.217e+02, threshold=6.047e+02, percent-clipped=2.0 +2023-03-09 19:27:42,704 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-09 19:27:50,758 INFO [train.py:898] (0/4) Epoch 25, batch 3400, loss[loss=0.1623, simple_loss=0.2504, pruned_loss=0.03711, over 18552.00 frames. ], tot_loss[loss=0.1595, simple_loss=0.2499, pruned_loss=0.03455, over 3577357.25 frames. ], batch size: 49, lr: 4.46e-03, grad_scale: 8.0 +2023-03-09 19:28:17,992 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5391, 6.1579, 5.5793, 5.9859, 5.7712, 5.5586, 6.2466, 6.1744], + device='cuda:0'), covar=tensor([0.1115, 0.0666, 0.0418, 0.0681, 0.1299, 0.0763, 0.0559, 0.0642], + device='cuda:0'), in_proj_covar=tensor([0.0627, 0.0557, 0.0404, 0.0580, 0.0778, 0.0575, 0.0795, 0.0599], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 19:28:49,671 INFO [train.py:898] (0/4) Epoch 25, batch 3450, loss[loss=0.1848, simple_loss=0.2765, pruned_loss=0.04649, over 18312.00 frames. ], tot_loss[loss=0.1584, simple_loss=0.2489, pruned_loss=0.03395, over 3578006.19 frames. ], batch size: 54, lr: 4.46e-03, grad_scale: 8.0 +2023-03-09 19:28:50,490 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.90 vs. limit=5.0 +2023-03-09 19:29:17,409 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9085, 3.9314, 3.7381, 3.4054, 3.7255, 3.1095, 3.0815, 3.9620], + device='cuda:0'), covar=tensor([0.0084, 0.0098, 0.0085, 0.0133, 0.0088, 0.0187, 0.0202, 0.0060], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0170, 0.0143, 0.0194, 0.0150, 0.0185, 0.0188, 0.0130], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 19:29:31,655 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.836e+02 2.461e+02 2.925e+02 3.622e+02 7.224e+02, threshold=5.850e+02, percent-clipped=2.0 +2023-03-09 19:29:37,284 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90707.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:29:48,682 INFO [train.py:898] (0/4) Epoch 25, batch 3500, loss[loss=0.1437, simple_loss=0.2357, pruned_loss=0.02589, over 18364.00 frames. ], tot_loss[loss=0.1579, simple_loss=0.2484, pruned_loss=0.03375, over 3589521.50 frames. ], batch size: 46, lr: 4.46e-03, grad_scale: 8.0 +2023-03-09 19:29:56,895 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6401, 3.5585, 4.8197, 4.1780, 3.2030, 2.8524, 4.3167, 5.1042], + device='cuda:0'), covar=tensor([0.0870, 0.1643, 0.0238, 0.0476, 0.1038, 0.1313, 0.0421, 0.0359], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0282, 0.0166, 0.0185, 0.0195, 0.0194, 0.0199, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 19:29:58,914 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7242, 3.7435, 3.6265, 3.2322, 3.5924, 2.9031, 2.8635, 3.7699], + device='cuda:0'), covar=tensor([0.0084, 0.0102, 0.0090, 0.0142, 0.0088, 0.0200, 0.0220, 0.0080], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0169, 0.0142, 0.0193, 0.0149, 0.0184, 0.0188, 0.0129], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 19:30:44,126 INFO [train.py:898] (0/4) Epoch 25, batch 3550, loss[loss=0.1711, simple_loss=0.2634, pruned_loss=0.0394, over 18076.00 frames. ], tot_loss[loss=0.1581, simple_loss=0.2486, pruned_loss=0.03382, over 3594955.47 frames. ], batch size: 62, lr: 4.45e-03, grad_scale: 8.0 +2023-03-09 19:30:55,381 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-09 19:31:04,086 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90786.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:31:22,415 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.855e+02 2.570e+02 3.030e+02 3.622e+02 8.486e+02, threshold=6.060e+02, percent-clipped=3.0 +2023-03-09 19:31:35,409 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90815.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:31:37,538 INFO [train.py:898] (0/4) Epoch 25, batch 3600, loss[loss=0.1319, simple_loss=0.2122, pruned_loss=0.02585, over 17686.00 frames. ], tot_loss[loss=0.1577, simple_loss=0.2478, pruned_loss=0.03379, over 3582032.00 frames. ], batch size: 39, lr: 4.45e-03, grad_scale: 8.0 +2023-03-09 19:31:56,435 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=90834.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:31:58,697 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6864, 3.0641, 4.3959, 3.6617, 2.6461, 4.6209, 3.8866, 3.0446], + device='cuda:0'), covar=tensor([0.0481, 0.1238, 0.0260, 0.0492, 0.1586, 0.0220, 0.0589, 0.0851], + device='cuda:0'), in_proj_covar=tensor([0.0215, 0.0243, 0.0226, 0.0169, 0.0226, 0.0217, 0.0254, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 19:32:11,868 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=90849.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:32:13,980 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-25.pt +2023-03-09 19:32:40,256 INFO [train.py:898] (0/4) Epoch 26, batch 0, loss[loss=0.1636, simple_loss=0.2614, pruned_loss=0.03286, over 18567.00 frames. ], tot_loss[loss=0.1636, simple_loss=0.2614, pruned_loss=0.03286, over 18567.00 frames. ], batch size: 54, lr: 4.36e-03, grad_scale: 8.0 +2023-03-09 19:32:40,258 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 19:32:52,309 INFO [train.py:932] (0/4) Epoch 26, validation: loss=0.1501, simple_loss=0.2487, pruned_loss=0.02573, over 944034.00 frames. +2023-03-09 19:32:52,310 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 19:32:53,456 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=90852.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:33:17,671 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.37 vs. limit=2.0 +2023-03-09 19:33:31,443 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-09 19:33:35,290 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=90888.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:33:49,734 INFO [train.py:898] (0/4) Epoch 26, batch 50, loss[loss=0.1661, simple_loss=0.2589, pruned_loss=0.03667, over 17000.00 frames. ], tot_loss[loss=0.1581, simple_loss=0.2481, pruned_loss=0.03408, over 814222.31 frames. ], batch size: 78, lr: 4.36e-03, grad_scale: 8.0 +2023-03-09 19:33:52,048 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.759e+02 2.622e+02 3.196e+02 4.102e+02 7.514e+02, threshold=6.391e+02, percent-clipped=4.0 +2023-03-09 19:34:01,674 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=90910.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:34:27,515 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9691, 3.2963, 4.6164, 3.8667, 2.7825, 4.8447, 4.1239, 3.1890], + device='cuda:0'), covar=tensor([0.0403, 0.1194, 0.0247, 0.0458, 0.1563, 0.0207, 0.0494, 0.0887], + device='cuda:0'), in_proj_covar=tensor([0.0214, 0.0242, 0.0225, 0.0169, 0.0224, 0.0216, 0.0253, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 19:34:31,656 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=90936.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:34:44,416 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5655, 2.3571, 2.5337, 2.6008, 3.2090, 4.8009, 4.7332, 3.1917], + device='cuda:0'), covar=tensor([0.2005, 0.2633, 0.3146, 0.2036, 0.2471, 0.0264, 0.0370, 0.1129], + device='cuda:0'), in_proj_covar=tensor([0.0318, 0.0356, 0.0400, 0.0286, 0.0392, 0.0256, 0.0298, 0.0266], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 19:34:48,576 INFO [train.py:898] (0/4) Epoch 26, batch 100, loss[loss=0.1553, simple_loss=0.2529, pruned_loss=0.02882, over 18475.00 frames. ], tot_loss[loss=0.1589, simple_loss=0.2498, pruned_loss=0.034, over 1436465.27 frames. ], batch size: 53, lr: 4.36e-03, grad_scale: 8.0 +2023-03-09 19:35:03,419 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9831, 5.3259, 2.5958, 5.1768, 5.0917, 5.3884, 5.2774, 2.6331], + device='cuda:0'), covar=tensor([0.0218, 0.0056, 0.0900, 0.0074, 0.0071, 0.0070, 0.0070, 0.1089], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0085, 0.0099, 0.0100, 0.0090, 0.0081, 0.0088, 0.0100], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 19:35:47,460 INFO [train.py:898] (0/4) Epoch 26, batch 150, loss[loss=0.1428, simple_loss=0.2274, pruned_loss=0.02909, over 17646.00 frames. ], tot_loss[loss=0.1575, simple_loss=0.248, pruned_loss=0.03347, over 1923966.53 frames. ], batch size: 39, lr: 4.36e-03, grad_scale: 8.0 +2023-03-09 19:35:49,740 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.912e+02 2.712e+02 3.182e+02 3.692e+02 5.749e+02, threshold=6.364e+02, percent-clipped=0.0 +2023-03-09 19:35:54,572 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91007.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:36:06,110 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0342, 5.4359, 5.4400, 5.4257, 4.8768, 5.3195, 4.8147, 5.3415], + device='cuda:0'), covar=tensor([0.0233, 0.0292, 0.0185, 0.0412, 0.0432, 0.0228, 0.0984, 0.0310], + device='cuda:0'), in_proj_covar=tensor([0.0228, 0.0271, 0.0270, 0.0347, 0.0282, 0.0280, 0.0314, 0.0274], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 19:36:30,939 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9197, 3.6560, 5.1134, 2.9981, 4.4361, 2.6446, 3.1641, 1.8200], + device='cuda:0'), covar=tensor([0.1141, 0.0983, 0.0148, 0.0945, 0.0495, 0.2600, 0.2690, 0.2288], + device='cuda:0'), in_proj_covar=tensor([0.0227, 0.0249, 0.0217, 0.0205, 0.0261, 0.0276, 0.0333, 0.0242], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 19:36:31,099 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-09 19:36:34,674 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.40 vs. limit=2.0 +2023-03-09 19:36:46,107 INFO [train.py:898] (0/4) Epoch 26, batch 200, loss[loss=0.1532, simple_loss=0.2436, pruned_loss=0.03139, over 18572.00 frames. ], tot_loss[loss=0.1574, simple_loss=0.2478, pruned_loss=0.03353, over 2303757.14 frames. ], batch size: 54, lr: 4.36e-03, grad_scale: 8.0 +2023-03-09 19:36:50,853 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=91055.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:37:43,759 INFO [train.py:898] (0/4) Epoch 26, batch 250, loss[loss=0.1419, simple_loss=0.2231, pruned_loss=0.03034, over 18437.00 frames. ], tot_loss[loss=0.1578, simple_loss=0.2483, pruned_loss=0.03366, over 2595017.37 frames. ], batch size: 43, lr: 4.36e-03, grad_scale: 8.0 +2023-03-09 19:37:45,994 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.600e+02 2.590e+02 3.099e+02 3.572e+02 5.022e+02, threshold=6.197e+02, percent-clipped=0.0 +2023-03-09 19:37:59,500 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91115.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:38:11,457 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5093, 5.5260, 5.0288, 5.4531, 5.4712, 4.8478, 5.3647, 4.9789], + device='cuda:0'), covar=tensor([0.0540, 0.0516, 0.1612, 0.0892, 0.0699, 0.0520, 0.0548, 0.1222], + device='cuda:0'), in_proj_covar=tensor([0.0505, 0.0573, 0.0713, 0.0446, 0.0471, 0.0520, 0.0553, 0.0686], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 19:38:41,414 INFO [train.py:898] (0/4) Epoch 26, batch 300, loss[loss=0.1481, simple_loss=0.2246, pruned_loss=0.03584, over 17698.00 frames. ], tot_loss[loss=0.1573, simple_loss=0.2478, pruned_loss=0.03337, over 2824815.52 frames. ], batch size: 39, lr: 4.36e-03, grad_scale: 8.0 +2023-03-09 19:38:42,772 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91152.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:38:55,109 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=91163.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:39:15,527 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4117, 6.0867, 5.6741, 5.8423, 5.6374, 5.3998, 6.1222, 6.0741], + device='cuda:0'), covar=tensor([0.1235, 0.0667, 0.0386, 0.0679, 0.1396, 0.0712, 0.0613, 0.0697], + device='cuda:0'), in_proj_covar=tensor([0.0625, 0.0550, 0.0402, 0.0576, 0.0773, 0.0570, 0.0786, 0.0598], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 19:39:39,114 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=91200.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:39:40,111 INFO [train.py:898] (0/4) Epoch 26, batch 350, loss[loss=0.1584, simple_loss=0.2519, pruned_loss=0.03245, over 16163.00 frames. ], tot_loss[loss=0.1569, simple_loss=0.2472, pruned_loss=0.03326, over 2991488.76 frames. ], batch size: 94, lr: 4.36e-03, grad_scale: 8.0 +2023-03-09 19:39:42,442 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.808e+02 2.557e+02 3.001e+02 3.591e+02 6.784e+02, threshold=6.003e+02, percent-clipped=1.0 +2023-03-09 19:39:44,861 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=91205.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:40:00,582 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2456, 5.2445, 4.8701, 5.1634, 5.2071, 4.5411, 5.0924, 4.8192], + device='cuda:0'), covar=tensor([0.0475, 0.0413, 0.1252, 0.0790, 0.0537, 0.0466, 0.0430, 0.1109], + device='cuda:0'), in_proj_covar=tensor([0.0509, 0.0577, 0.0718, 0.0450, 0.0475, 0.0523, 0.0557, 0.0691], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 19:40:38,704 INFO [train.py:898] (0/4) Epoch 26, batch 400, loss[loss=0.1339, simple_loss=0.2161, pruned_loss=0.02585, over 18256.00 frames. ], tot_loss[loss=0.1569, simple_loss=0.247, pruned_loss=0.0334, over 3126628.73 frames. ], batch size: 45, lr: 4.35e-03, grad_scale: 8.0 +2023-03-09 19:41:00,464 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8751, 5.3305, 2.7236, 5.1597, 5.0514, 5.3688, 5.1931, 2.7706], + device='cuda:0'), covar=tensor([0.0238, 0.0055, 0.0815, 0.0073, 0.0074, 0.0058, 0.0076, 0.0966], + device='cuda:0'), in_proj_covar=tensor([0.0091, 0.0084, 0.0098, 0.0099, 0.0089, 0.0079, 0.0087, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 19:41:18,966 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91285.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:41:37,885 INFO [train.py:898] (0/4) Epoch 26, batch 450, loss[loss=0.1558, simple_loss=0.2459, pruned_loss=0.03283, over 18388.00 frames. ], tot_loss[loss=0.1565, simple_loss=0.2469, pruned_loss=0.03301, over 3227653.55 frames. ], batch size: 50, lr: 4.35e-03, grad_scale: 8.0 +2023-03-09 19:41:40,034 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.596e+02 2.481e+02 2.873e+02 3.390e+02 6.237e+02, threshold=5.746e+02, percent-clipped=1.0 +2023-03-09 19:42:05,106 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8463, 5.3775, 2.7988, 5.1991, 5.1329, 5.4165, 5.2403, 2.7623], + device='cuda:0'), covar=tensor([0.0260, 0.0071, 0.0779, 0.0085, 0.0074, 0.0067, 0.0082, 0.1010], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0084, 0.0097, 0.0099, 0.0089, 0.0079, 0.0087, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 19:42:13,162 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-09 19:42:29,754 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=91346.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:42:35,722 INFO [train.py:898] (0/4) Epoch 26, batch 500, loss[loss=0.1368, simple_loss=0.2203, pruned_loss=0.02665, over 18276.00 frames. ], tot_loss[loss=0.1568, simple_loss=0.2474, pruned_loss=0.03309, over 3302399.71 frames. ], batch size: 45, lr: 4.35e-03, grad_scale: 8.0 +2023-03-09 19:43:32,649 INFO [train.py:898] (0/4) Epoch 26, batch 550, loss[loss=0.1488, simple_loss=0.2479, pruned_loss=0.02488, over 18340.00 frames. ], tot_loss[loss=0.1566, simple_loss=0.247, pruned_loss=0.03313, over 3372848.05 frames. ], batch size: 55, lr: 4.35e-03, grad_scale: 8.0 +2023-03-09 19:43:35,350 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.676e+02 2.412e+02 2.843e+02 3.584e+02 8.267e+02, threshold=5.686e+02, percent-clipped=4.0 +2023-03-09 19:43:36,267 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-09 19:43:52,701 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7643, 3.6224, 4.9330, 2.8332, 4.3059, 2.5250, 3.0600, 1.7010], + device='cuda:0'), covar=tensor([0.1314, 0.1013, 0.0179, 0.1044, 0.0536, 0.2712, 0.2730, 0.2424], + device='cuda:0'), in_proj_covar=tensor([0.0230, 0.0252, 0.0219, 0.0207, 0.0265, 0.0279, 0.0338, 0.0244], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 19:44:11,348 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5812, 5.3632, 5.7244, 5.8290, 5.3852, 6.2554, 5.8985, 5.6014], + device='cuda:0'), covar=tensor([0.0951, 0.0611, 0.0686, 0.0700, 0.1372, 0.0684, 0.0631, 0.1533], + device='cuda:0'), in_proj_covar=tensor([0.0364, 0.0298, 0.0322, 0.0324, 0.0337, 0.0435, 0.0291, 0.0426], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 19:44:30,263 INFO [train.py:898] (0/4) Epoch 26, batch 600, loss[loss=0.1781, simple_loss=0.2661, pruned_loss=0.04508, over 17193.00 frames. ], tot_loss[loss=0.1565, simple_loss=0.2467, pruned_loss=0.03315, over 3429639.72 frames. ], batch size: 78, lr: 4.35e-03, grad_scale: 8.0 +2023-03-09 19:44:32,202 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7797, 3.4908, 4.7908, 2.9175, 4.1519, 2.5959, 3.0254, 1.7548], + device='cuda:0'), covar=tensor([0.1323, 0.1103, 0.0208, 0.1050, 0.0595, 0.2630, 0.2746, 0.2534], + device='cuda:0'), in_proj_covar=tensor([0.0230, 0.0252, 0.0219, 0.0207, 0.0265, 0.0279, 0.0338, 0.0244], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 19:44:34,294 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7942, 2.9678, 2.8784, 3.0024, 3.8186, 3.6983, 3.4036, 3.1648], + device='cuda:0'), covar=tensor([0.0157, 0.0293, 0.0496, 0.0375, 0.0168, 0.0170, 0.0343, 0.0363], + device='cuda:0'), in_proj_covar=tensor([0.0143, 0.0143, 0.0165, 0.0162, 0.0139, 0.0123, 0.0160, 0.0162], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 19:44:58,103 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.60 vs. limit=2.0 +2023-03-09 19:45:29,313 INFO [train.py:898] (0/4) Epoch 26, batch 650, loss[loss=0.1488, simple_loss=0.2471, pruned_loss=0.02525, over 18620.00 frames. ], tot_loss[loss=0.1568, simple_loss=0.2471, pruned_loss=0.03332, over 3474699.28 frames. ], batch size: 52, lr: 4.35e-03, grad_scale: 8.0 +2023-03-09 19:45:32,528 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.747e+02 2.481e+02 2.888e+02 3.491e+02 6.758e+02, threshold=5.775e+02, percent-clipped=2.0 +2023-03-09 19:45:34,087 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91505.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:46:16,603 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91542.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 19:46:20,127 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8126, 3.7056, 5.0071, 2.8854, 4.4120, 2.6347, 3.0674, 1.7282], + device='cuda:0'), covar=tensor([0.1218, 0.0961, 0.0162, 0.1023, 0.0502, 0.2639, 0.2687, 0.2327], + device='cuda:0'), in_proj_covar=tensor([0.0230, 0.0253, 0.0221, 0.0208, 0.0265, 0.0279, 0.0339, 0.0245], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 19:46:27,363 INFO [train.py:898] (0/4) Epoch 26, batch 700, loss[loss=0.1556, simple_loss=0.2522, pruned_loss=0.02948, over 17769.00 frames. ], tot_loss[loss=0.1574, simple_loss=0.2478, pruned_loss=0.03349, over 3505894.89 frames. ], batch size: 70, lr: 4.35e-03, grad_scale: 4.0 +2023-03-09 19:46:29,758 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=91553.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:47:00,398 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.6928, 5.6783, 5.3095, 5.6354, 5.6234, 4.9930, 5.5504, 5.2528], + device='cuda:0'), covar=tensor([0.0390, 0.0375, 0.1271, 0.0668, 0.0496, 0.0367, 0.0383, 0.0979], + device='cuda:0'), in_proj_covar=tensor([0.0507, 0.0576, 0.0717, 0.0449, 0.0473, 0.0521, 0.0557, 0.0691], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 19:47:04,321 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-09 19:47:26,223 INFO [train.py:898] (0/4) Epoch 26, batch 750, loss[loss=0.1506, simple_loss=0.233, pruned_loss=0.03412, over 18246.00 frames. ], tot_loss[loss=0.1576, simple_loss=0.2479, pruned_loss=0.03365, over 3508582.14 frames. ], batch size: 45, lr: 4.35e-03, grad_scale: 4.0 +2023-03-09 19:47:29,407 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=91603.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 19:47:30,095 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.568e+02 2.592e+02 2.993e+02 3.569e+02 6.310e+02, threshold=5.987e+02, percent-clipped=2.0 +2023-03-09 19:48:12,469 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=91641.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:48:16,630 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.23 vs. limit=2.0 +2023-03-09 19:48:23,828 INFO [train.py:898] (0/4) Epoch 26, batch 800, loss[loss=0.1318, simple_loss=0.2179, pruned_loss=0.02289, over 18274.00 frames. ], tot_loss[loss=0.1577, simple_loss=0.248, pruned_loss=0.03368, over 3525160.73 frames. ], batch size: 45, lr: 4.35e-03, grad_scale: 8.0 +2023-03-09 19:49:22,518 INFO [train.py:898] (0/4) Epoch 26, batch 850, loss[loss=0.159, simple_loss=0.252, pruned_loss=0.03301, over 18406.00 frames. ], tot_loss[loss=0.158, simple_loss=0.2482, pruned_loss=0.03393, over 3540976.32 frames. ], batch size: 50, lr: 4.34e-03, grad_scale: 8.0 +2023-03-09 19:49:25,717 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.771e+02 2.531e+02 3.031e+02 3.445e+02 6.326e+02, threshold=6.062e+02, percent-clipped=1.0 +2023-03-09 19:50:20,616 INFO [train.py:898] (0/4) Epoch 26, batch 900, loss[loss=0.1547, simple_loss=0.2514, pruned_loss=0.02897, over 18493.00 frames. ], tot_loss[loss=0.1582, simple_loss=0.2487, pruned_loss=0.03389, over 3559874.62 frames. ], batch size: 53, lr: 4.34e-03, grad_scale: 8.0 +2023-03-09 19:51:19,512 INFO [train.py:898] (0/4) Epoch 26, batch 950, loss[loss=0.1396, simple_loss=0.2241, pruned_loss=0.02756, over 18375.00 frames. ], tot_loss[loss=0.1584, simple_loss=0.2491, pruned_loss=0.03385, over 3574370.62 frames. ], batch size: 46, lr: 4.34e-03, grad_scale: 8.0 +2023-03-09 19:51:19,874 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8316, 3.9066, 3.6807, 3.3878, 3.6013, 2.9974, 3.0241, 3.8600], + device='cuda:0'), covar=tensor([0.0076, 0.0096, 0.0092, 0.0130, 0.0108, 0.0188, 0.0203, 0.0073], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0173, 0.0145, 0.0196, 0.0154, 0.0188, 0.0192, 0.0132], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 19:51:22,552 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.812e+02 2.439e+02 3.074e+02 3.438e+02 1.468e+03, threshold=6.149e+02, percent-clipped=4.0 +2023-03-09 19:51:37,884 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8358, 3.7079, 5.0264, 4.5805, 3.3402, 3.0576, 4.4843, 5.2416], + device='cuda:0'), covar=tensor([0.0840, 0.1732, 0.0205, 0.0380, 0.0999, 0.1269, 0.0391, 0.0243], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0283, 0.0167, 0.0187, 0.0196, 0.0195, 0.0201, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 19:52:18,267 INFO [train.py:898] (0/4) Epoch 26, batch 1000, loss[loss=0.1556, simple_loss=0.2478, pruned_loss=0.03171, over 18483.00 frames. ], tot_loss[loss=0.1583, simple_loss=0.2491, pruned_loss=0.03375, over 3577943.57 frames. ], batch size: 51, lr: 4.34e-03, grad_scale: 8.0 +2023-03-09 19:53:05,492 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91891.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:53:13,088 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=91898.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 19:53:16,256 INFO [train.py:898] (0/4) Epoch 26, batch 1050, loss[loss=0.1492, simple_loss=0.2482, pruned_loss=0.02511, over 18285.00 frames. ], tot_loss[loss=0.1582, simple_loss=0.249, pruned_loss=0.0337, over 3586011.41 frames. ], batch size: 49, lr: 4.34e-03, grad_scale: 8.0 +2023-03-09 19:53:19,766 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.941e+02 2.632e+02 3.014e+02 3.529e+02 5.169e+02, threshold=6.027e+02, percent-clipped=0.0 +2023-03-09 19:53:25,816 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9297, 4.2493, 4.1864, 4.2474, 3.8603, 4.1652, 3.8502, 4.1707], + device='cuda:0'), covar=tensor([0.0277, 0.0328, 0.0266, 0.0542, 0.0382, 0.0259, 0.0854, 0.0347], + device='cuda:0'), in_proj_covar=tensor([0.0230, 0.0276, 0.0276, 0.0354, 0.0288, 0.0285, 0.0318, 0.0279], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0006, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 19:54:03,932 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=91941.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:54:07,163 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.6485, 6.1710, 5.6078, 6.0167, 5.8132, 5.6461, 6.3040, 6.2146], + device='cuda:0'), covar=tensor([0.1140, 0.0755, 0.0479, 0.0621, 0.1365, 0.0638, 0.0522, 0.0648], + device='cuda:0'), in_proj_covar=tensor([0.0624, 0.0554, 0.0400, 0.0576, 0.0772, 0.0570, 0.0784, 0.0598], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 19:54:14,902 INFO [train.py:898] (0/4) Epoch 26, batch 1100, loss[loss=0.1347, simple_loss=0.2228, pruned_loss=0.02331, over 18255.00 frames. ], tot_loss[loss=0.158, simple_loss=0.2487, pruned_loss=0.03369, over 3576772.94 frames. ], batch size: 47, lr: 4.34e-03, grad_scale: 8.0 +2023-03-09 19:54:16,435 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=91952.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:54:19,642 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91955.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:54:50,166 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=91981.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 19:54:51,983 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4903, 3.3640, 3.3306, 2.9208, 3.1404, 2.5666, 2.6182, 3.2659], + device='cuda:0'), covar=tensor([0.0087, 0.0151, 0.0113, 0.0174, 0.0151, 0.0273, 0.0265, 0.0126], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0175, 0.0146, 0.0197, 0.0155, 0.0188, 0.0192, 0.0132], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 19:54:59,716 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=91989.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:55:09,274 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8219, 3.7255, 5.1011, 3.0131, 4.4743, 2.6326, 3.1036, 1.8137], + device='cuda:0'), covar=tensor([0.1276, 0.1023, 0.0140, 0.0936, 0.0507, 0.2557, 0.2715, 0.2334], + device='cuda:0'), in_proj_covar=tensor([0.0229, 0.0253, 0.0220, 0.0207, 0.0267, 0.0279, 0.0339, 0.0244], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 19:55:12,605 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-92000.pt +2023-03-09 19:55:17,948 INFO [train.py:898] (0/4) Epoch 26, batch 1150, loss[loss=0.1647, simple_loss=0.2579, pruned_loss=0.0358, over 16139.00 frames. ], tot_loss[loss=0.1584, simple_loss=0.2491, pruned_loss=0.03387, over 3572967.61 frames. ], batch size: 94, lr: 4.34e-03, grad_scale: 8.0 +2023-03-09 19:55:19,503 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3402, 2.6914, 2.3851, 2.6736, 3.3878, 3.3380, 2.9841, 2.7161], + device='cuda:0'), covar=tensor([0.0222, 0.0326, 0.0656, 0.0469, 0.0247, 0.0197, 0.0438, 0.0437], + device='cuda:0'), in_proj_covar=tensor([0.0145, 0.0144, 0.0167, 0.0164, 0.0141, 0.0125, 0.0161, 0.0164], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 19:55:21,343 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.947e+02 2.558e+02 3.039e+02 3.805e+02 7.483e+02, threshold=6.077e+02, percent-clipped=1.0 +2023-03-09 19:55:34,881 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92016.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 19:55:38,442 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8736, 3.7368, 5.0830, 4.6433, 3.5821, 3.1486, 4.5479, 5.3104], + device='cuda:0'), covar=tensor([0.0822, 0.1586, 0.0217, 0.0349, 0.0881, 0.1236, 0.0398, 0.0344], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0283, 0.0167, 0.0187, 0.0196, 0.0196, 0.0201, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 19:56:06,534 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92042.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 19:56:16,298 INFO [train.py:898] (0/4) Epoch 26, batch 1200, loss[loss=0.1617, simple_loss=0.2569, pruned_loss=0.03326, over 18294.00 frames. ], tot_loss[loss=0.1583, simple_loss=0.2489, pruned_loss=0.0339, over 3579812.56 frames. ], batch size: 54, lr: 4.34e-03, grad_scale: 8.0 +2023-03-09 19:56:50,578 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9604, 5.4431, 2.6383, 5.3037, 5.1840, 5.4685, 5.3019, 2.6979], + device='cuda:0'), covar=tensor([0.0226, 0.0051, 0.0837, 0.0070, 0.0060, 0.0064, 0.0071, 0.1005], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0084, 0.0098, 0.0099, 0.0089, 0.0080, 0.0087, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 19:57:14,321 INFO [train.py:898] (0/4) Epoch 26, batch 1250, loss[loss=0.1636, simple_loss=0.2605, pruned_loss=0.03334, over 18571.00 frames. ], tot_loss[loss=0.1586, simple_loss=0.2493, pruned_loss=0.03396, over 3578640.03 frames. ], batch size: 54, lr: 4.33e-03, grad_scale: 8.0 +2023-03-09 19:57:17,634 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.939e+02 2.685e+02 3.194e+02 3.689e+02 7.845e+02, threshold=6.387e+02, percent-clipped=2.0 +2023-03-09 19:58:12,743 INFO [train.py:898] (0/4) Epoch 26, batch 1300, loss[loss=0.1488, simple_loss=0.228, pruned_loss=0.03482, over 18361.00 frames. ], tot_loss[loss=0.1581, simple_loss=0.2485, pruned_loss=0.0338, over 3586540.62 frames. ], batch size: 42, lr: 4.33e-03, grad_scale: 8.0 +2023-03-09 19:58:19,710 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4281, 3.9194, 3.8883, 3.1319, 3.3873, 3.1040, 2.5283, 2.3899], + device='cuda:0'), covar=tensor([0.0248, 0.0158, 0.0115, 0.0308, 0.0344, 0.0265, 0.0669, 0.0821], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0063, 0.0068, 0.0070, 0.0092, 0.0071, 0.0079, 0.0086], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0006, 0.0005, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 19:58:42,668 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7276, 3.7361, 2.3018, 4.6323, 3.2451, 4.1949, 2.6826, 4.1382], + device='cuda:0'), covar=tensor([0.0575, 0.0757, 0.1509, 0.0465, 0.0781, 0.0344, 0.1193, 0.0455], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0231, 0.0195, 0.0295, 0.0196, 0.0271, 0.0207, 0.0206], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 19:58:46,214 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.4960, 2.1305, 2.3626, 2.5205, 2.7336, 4.5771, 4.5435, 3.1858], + device='cuda:0'), covar=tensor([0.2479, 0.3493, 0.3922, 0.2331, 0.3902, 0.0366, 0.0442, 0.1184], + device='cuda:0'), in_proj_covar=tensor([0.0321, 0.0358, 0.0403, 0.0288, 0.0394, 0.0259, 0.0301, 0.0268], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 19:59:07,902 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92198.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 19:59:11,336 INFO [train.py:898] (0/4) Epoch 26, batch 1350, loss[loss=0.1649, simple_loss=0.2538, pruned_loss=0.03797, over 18345.00 frames. ], tot_loss[loss=0.1581, simple_loss=0.2485, pruned_loss=0.03386, over 3589469.56 frames. ], batch size: 56, lr: 4.33e-03, grad_scale: 8.0 +2023-03-09 19:59:14,646 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.584e+02 2.419e+02 2.885e+02 3.522e+02 6.118e+02, threshold=5.770e+02, percent-clipped=0.0 +2023-03-09 19:59:26,229 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8179, 3.7009, 4.9219, 4.4012, 3.4038, 2.9301, 4.3496, 5.2374], + device='cuda:0'), covar=tensor([0.0778, 0.1456, 0.0245, 0.0410, 0.0923, 0.1267, 0.0432, 0.0206], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0283, 0.0168, 0.0187, 0.0197, 0.0196, 0.0201, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 19:59:31,611 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2854, 3.2688, 2.1445, 3.9768, 2.7437, 3.7207, 2.4720, 3.5377], + device='cuda:0'), covar=tensor([0.0685, 0.0855, 0.1442, 0.0569, 0.0900, 0.0293, 0.1138, 0.0461], + device='cuda:0'), in_proj_covar=tensor([0.0225, 0.0232, 0.0196, 0.0297, 0.0197, 0.0272, 0.0208, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 20:00:03,381 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=92246.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 20:00:04,391 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92247.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:00:09,206 INFO [train.py:898] (0/4) Epoch 26, batch 1400, loss[loss=0.1524, simple_loss=0.2384, pruned_loss=0.03323, over 18426.00 frames. ], tot_loss[loss=0.1579, simple_loss=0.2481, pruned_loss=0.03391, over 3583447.07 frames. ], batch size: 43, lr: 4.33e-03, grad_scale: 8.0 +2023-03-09 20:00:21,857 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7848, 3.1260, 4.4706, 3.7028, 2.9374, 4.7684, 3.9668, 3.1138], + device='cuda:0'), covar=tensor([0.0471, 0.1412, 0.0343, 0.0528, 0.1527, 0.0196, 0.0600, 0.0934], + device='cuda:0'), in_proj_covar=tensor([0.0219, 0.0246, 0.0231, 0.0172, 0.0228, 0.0221, 0.0259, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 20:01:06,842 INFO [train.py:898] (0/4) Epoch 26, batch 1450, loss[loss=0.1758, simple_loss=0.2658, pruned_loss=0.04288, over 18249.00 frames. ], tot_loss[loss=0.1579, simple_loss=0.248, pruned_loss=0.0339, over 3593491.25 frames. ], batch size: 60, lr: 4.33e-03, grad_scale: 8.0 +2023-03-09 20:01:10,175 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.801e+02 2.582e+02 3.137e+02 3.833e+02 9.171e+02, threshold=6.274e+02, percent-clipped=8.0 +2023-03-09 20:01:14,622 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7083, 4.0975, 2.3761, 4.1024, 5.0610, 2.6052, 3.8934, 3.9770], + device='cuda:0'), covar=tensor([0.0221, 0.1090, 0.1627, 0.0580, 0.0096, 0.1183, 0.0618, 0.0663], + device='cuda:0'), in_proj_covar=tensor([0.0180, 0.0279, 0.0207, 0.0201, 0.0138, 0.0186, 0.0222, 0.0228], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 20:01:18,983 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92311.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:01:48,292 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92337.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 20:02:04,832 INFO [train.py:898] (0/4) Epoch 26, batch 1500, loss[loss=0.1724, simple_loss=0.2727, pruned_loss=0.03605, over 17903.00 frames. ], tot_loss[loss=0.1576, simple_loss=0.2482, pruned_loss=0.03352, over 3595344.05 frames. ], batch size: 70, lr: 4.33e-03, grad_scale: 8.0 +2023-03-09 20:03:03,800 INFO [train.py:898] (0/4) Epoch 26, batch 1550, loss[loss=0.1501, simple_loss=0.2446, pruned_loss=0.02785, over 18504.00 frames. ], tot_loss[loss=0.158, simple_loss=0.2486, pruned_loss=0.03374, over 3605148.63 frames. ], batch size: 47, lr: 4.33e-03, grad_scale: 8.0 +2023-03-09 20:03:07,217 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.867e+02 2.584e+02 2.933e+02 3.549e+02 6.992e+02, threshold=5.866e+02, percent-clipped=1.0 +2023-03-09 20:04:01,753 INFO [train.py:898] (0/4) Epoch 26, batch 1600, loss[loss=0.1536, simple_loss=0.2492, pruned_loss=0.02899, over 18371.00 frames. ], tot_loss[loss=0.1568, simple_loss=0.2472, pruned_loss=0.03322, over 3607285.35 frames. ], batch size: 56, lr: 4.33e-03, grad_scale: 8.0 +2023-03-09 20:04:33,947 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9432, 3.2422, 4.6645, 4.0211, 3.1558, 4.9022, 4.2304, 3.2495], + device='cuda:0'), covar=tensor([0.0423, 0.1315, 0.0297, 0.0422, 0.1338, 0.0203, 0.0488, 0.0931], + device='cuda:0'), in_proj_covar=tensor([0.0219, 0.0244, 0.0229, 0.0171, 0.0226, 0.0219, 0.0256, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 20:04:38,709 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.82 vs. limit=2.0 +2023-03-09 20:04:59,855 INFO [train.py:898] (0/4) Epoch 26, batch 1650, loss[loss=0.152, simple_loss=0.2283, pruned_loss=0.03782, over 18380.00 frames. ], tot_loss[loss=0.1565, simple_loss=0.2466, pruned_loss=0.03325, over 3589443.05 frames. ], batch size: 42, lr: 4.33e-03, grad_scale: 8.0 +2023-03-09 20:05:02,960 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.649e+02 2.477e+02 2.880e+02 3.587e+02 5.533e+02, threshold=5.760e+02, percent-clipped=0.0 +2023-03-09 20:05:53,372 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92547.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:05:58,142 INFO [train.py:898] (0/4) Epoch 26, batch 1700, loss[loss=0.1468, simple_loss=0.2351, pruned_loss=0.02924, over 18291.00 frames. ], tot_loss[loss=0.1568, simple_loss=0.2471, pruned_loss=0.03328, over 3603583.19 frames. ], batch size: 49, lr: 4.32e-03, grad_scale: 8.0 +2023-03-09 20:06:48,745 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=92595.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:06:55,618 INFO [train.py:898] (0/4) Epoch 26, batch 1750, loss[loss=0.1498, simple_loss=0.252, pruned_loss=0.02374, over 18579.00 frames. ], tot_loss[loss=0.1573, simple_loss=0.2481, pruned_loss=0.03326, over 3602461.19 frames. ], batch size: 54, lr: 4.32e-03, grad_scale: 8.0 +2023-03-09 20:06:59,646 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.605e+02 2.546e+02 2.994e+02 3.589e+02 6.308e+02, threshold=5.987e+02, percent-clipped=1.0 +2023-03-09 20:07:00,132 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0510, 3.7199, 5.0535, 3.0107, 4.4943, 2.6642, 3.0605, 1.9085], + device='cuda:0'), covar=tensor([0.1166, 0.1009, 0.0176, 0.0984, 0.0502, 0.2722, 0.2975, 0.2338], + device='cuda:0'), in_proj_covar=tensor([0.0228, 0.0251, 0.0219, 0.0206, 0.0266, 0.0277, 0.0336, 0.0244], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 20:07:08,069 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92611.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:07:25,029 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92625.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:07:38,450 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=92637.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 20:07:54,310 INFO [train.py:898] (0/4) Epoch 26, batch 1800, loss[loss=0.1801, simple_loss=0.265, pruned_loss=0.04759, over 12458.00 frames. ], tot_loss[loss=0.1572, simple_loss=0.248, pruned_loss=0.03321, over 3595733.82 frames. ], batch size: 129, lr: 4.32e-03, grad_scale: 8.0 +2023-03-09 20:08:04,020 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=92659.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:08:35,015 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=92685.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 20:08:36,155 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92686.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:08:52,536 INFO [train.py:898] (0/4) Epoch 26, batch 1850, loss[loss=0.1528, simple_loss=0.2444, pruned_loss=0.03065, over 18481.00 frames. ], tot_loss[loss=0.1579, simple_loss=0.2487, pruned_loss=0.03349, over 3587812.25 frames. ], batch size: 51, lr: 4.32e-03, grad_scale: 8.0 +2023-03-09 20:08:55,754 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.027e+02 2.840e+02 3.330e+02 3.901e+02 1.111e+03, threshold=6.660e+02, percent-clipped=3.0 +2023-03-09 20:09:16,037 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92720.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:09:51,312 INFO [train.py:898] (0/4) Epoch 26, batch 1900, loss[loss=0.1756, simple_loss=0.2675, pruned_loss=0.04186, over 18482.00 frames. ], tot_loss[loss=0.158, simple_loss=0.2488, pruned_loss=0.03358, over 3586842.27 frames. ], batch size: 51, lr: 4.32e-03, grad_scale: 8.0 +2023-03-09 20:10:28,235 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92781.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:10:50,692 INFO [train.py:898] (0/4) Epoch 26, batch 1950, loss[loss=0.1282, simple_loss=0.2115, pruned_loss=0.02249, over 18457.00 frames. ], tot_loss[loss=0.1578, simple_loss=0.2483, pruned_loss=0.0336, over 3578107.63 frames. ], batch size: 43, lr: 4.32e-03, grad_scale: 8.0 +2023-03-09 20:10:54,074 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.627e+02 2.403e+02 2.804e+02 3.512e+02 6.323e+02, threshold=5.608e+02, percent-clipped=0.0 +2023-03-09 20:11:02,708 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-09 20:11:26,597 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92831.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:11:40,648 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.60 vs. limit=2.0 +2023-03-09 20:11:49,093 INFO [train.py:898] (0/4) Epoch 26, batch 2000, loss[loss=0.1519, simple_loss=0.2459, pruned_loss=0.02891, over 18612.00 frames. ], tot_loss[loss=0.1572, simple_loss=0.2474, pruned_loss=0.0335, over 3590546.33 frames. ], batch size: 52, lr: 4.32e-03, grad_scale: 8.0 +2023-03-09 20:12:38,376 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92892.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:12:45,999 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92899.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:12:47,829 INFO [train.py:898] (0/4) Epoch 26, batch 2050, loss[loss=0.1514, simple_loss=0.2467, pruned_loss=0.02808, over 18374.00 frames. ], tot_loss[loss=0.1572, simple_loss=0.2474, pruned_loss=0.03348, over 3593260.49 frames. ], batch size: 46, lr: 4.32e-03, grad_scale: 8.0 +2023-03-09 20:12:51,214 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.874e+02 2.574e+02 3.005e+02 3.405e+02 1.130e+03, threshold=6.011e+02, percent-clipped=2.0 +2023-03-09 20:13:23,167 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92931.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:13:45,946 INFO [train.py:898] (0/4) Epoch 26, batch 2100, loss[loss=0.164, simple_loss=0.2491, pruned_loss=0.03948, over 18252.00 frames. ], tot_loss[loss=0.1577, simple_loss=0.2479, pruned_loss=0.0337, over 3587270.84 frames. ], batch size: 45, lr: 4.31e-03, grad_scale: 8.0 +2023-03-09 20:13:56,493 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92960.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:14:13,916 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=92975.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:14:21,749 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=92981.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:14:34,849 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=92992.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:14:44,672 INFO [train.py:898] (0/4) Epoch 26, batch 2150, loss[loss=0.1384, simple_loss=0.2204, pruned_loss=0.0282, over 18388.00 frames. ], tot_loss[loss=0.1577, simple_loss=0.2478, pruned_loss=0.03373, over 3561690.06 frames. ], batch size: 42, lr: 4.31e-03, grad_scale: 8.0 +2023-03-09 20:14:48,050 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.894e+02 2.711e+02 3.209e+02 3.707e+02 5.473e+02, threshold=6.417e+02, percent-clipped=0.0 +2023-03-09 20:14:55,195 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93010.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:15:26,076 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93036.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:15:26,118 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93036.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:15:43,265 INFO [train.py:898] (0/4) Epoch 26, batch 2200, loss[loss=0.1616, simple_loss=0.2519, pruned_loss=0.03565, over 18255.00 frames. ], tot_loss[loss=0.1576, simple_loss=0.2478, pruned_loss=0.03372, over 3565144.37 frames. ], batch size: 60, lr: 4.31e-03, grad_scale: 8.0 +2023-03-09 20:15:48,091 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1383, 5.1543, 5.2042, 4.9371, 4.9749, 4.9843, 5.2892, 5.2721], + device='cuda:0'), covar=tensor([0.0060, 0.0067, 0.0058, 0.0104, 0.0060, 0.0140, 0.0086, 0.0112], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0074, 0.0079, 0.0098, 0.0079, 0.0108, 0.0091, 0.0090], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 20:16:06,493 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93071.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:16:12,765 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93076.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:16:15,345 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.44 vs. limit=2.0 +2023-03-09 20:16:38,240 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93097.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 20:16:42,339 INFO [train.py:898] (0/4) Epoch 26, batch 2250, loss[loss=0.1598, simple_loss=0.2588, pruned_loss=0.03044, over 18632.00 frames. ], tot_loss[loss=0.1572, simple_loss=0.2476, pruned_loss=0.03338, over 3582246.05 frames. ], batch size: 52, lr: 4.31e-03, grad_scale: 8.0 +2023-03-09 20:16:45,629 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.732e+02 2.458e+02 2.849e+02 3.441e+02 5.512e+02, threshold=5.698e+02, percent-clipped=0.0 +2023-03-09 20:16:58,276 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93115.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:17:20,372 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9694, 4.1552, 2.7233, 4.2363, 5.2078, 2.7771, 3.9778, 4.0722], + device='cuda:0'), covar=tensor([0.0175, 0.1133, 0.1443, 0.0576, 0.0091, 0.1121, 0.0634, 0.0652], + device='cuda:0'), in_proj_covar=tensor([0.0179, 0.0276, 0.0206, 0.0199, 0.0138, 0.0186, 0.0221, 0.0228], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 20:17:41,073 INFO [train.py:898] (0/4) Epoch 26, batch 2300, loss[loss=0.1939, simple_loss=0.2741, pruned_loss=0.05689, over 13155.00 frames. ], tot_loss[loss=0.1575, simple_loss=0.248, pruned_loss=0.0335, over 3574468.51 frames. ], batch size: 130, lr: 4.31e-03, grad_scale: 8.0 +2023-03-09 20:18:09,608 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93176.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:18:22,389 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93187.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:18:33,117 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0761, 5.1813, 5.2646, 5.3338, 5.0170, 5.8821, 5.5313, 5.1423], + device='cuda:0'), covar=tensor([0.1162, 0.0711, 0.0832, 0.0859, 0.1428, 0.0707, 0.0680, 0.1606], + device='cuda:0'), in_proj_covar=tensor([0.0369, 0.0304, 0.0326, 0.0330, 0.0343, 0.0441, 0.0296, 0.0432], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 20:18:40,068 INFO [train.py:898] (0/4) Epoch 26, batch 2350, loss[loss=0.1389, simple_loss=0.2194, pruned_loss=0.02913, over 18037.00 frames. ], tot_loss[loss=0.1577, simple_loss=0.2481, pruned_loss=0.0336, over 3587719.60 frames. ], batch size: 40, lr: 4.31e-03, grad_scale: 8.0 +2023-03-09 20:18:43,316 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.656e+02 2.468e+02 2.905e+02 3.361e+02 5.906e+02, threshold=5.811e+02, percent-clipped=1.0 +2023-03-09 20:19:38,035 INFO [train.py:898] (0/4) Epoch 26, batch 2400, loss[loss=0.1589, simple_loss=0.2533, pruned_loss=0.03227, over 18305.00 frames. ], tot_loss[loss=0.1579, simple_loss=0.2484, pruned_loss=0.03371, over 3592071.11 frames. ], batch size: 54, lr: 4.31e-03, grad_scale: 8.0 +2023-03-09 20:19:42,639 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93255.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:19:54,493 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.32 vs. limit=5.0 +2023-03-09 20:19:55,313 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7362, 2.3913, 2.6172, 2.7255, 3.2138, 4.8437, 4.6236, 3.3770], + device='cuda:0'), covar=tensor([0.2028, 0.2689, 0.3301, 0.2023, 0.2687, 0.0251, 0.0429, 0.1047], + device='cuda:0'), in_proj_covar=tensor([0.0324, 0.0361, 0.0404, 0.0287, 0.0395, 0.0262, 0.0301, 0.0270], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 20:20:11,738 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93281.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:20:18,997 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93287.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:20:34,697 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.11 vs. limit=5.0 +2023-03-09 20:20:36,268 INFO [train.py:898] (0/4) Epoch 26, batch 2450, loss[loss=0.1661, simple_loss=0.2572, pruned_loss=0.03751, over 16242.00 frames. ], tot_loss[loss=0.1581, simple_loss=0.2486, pruned_loss=0.03378, over 3582073.58 frames. ], batch size: 94, lr: 4.31e-03, grad_scale: 8.0 +2023-03-09 20:20:39,771 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.978e+02 2.528e+02 2.957e+02 3.512e+02 5.428e+02, threshold=5.913e+02, percent-clipped=0.0 +2023-03-09 20:20:54,133 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.23 vs. limit=5.0 +2023-03-09 20:21:08,323 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=93329.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:21:10,632 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93331.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:21:11,876 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3824, 5.3567, 4.9775, 5.2964, 5.3092, 4.7190, 5.1925, 4.9563], + device='cuda:0'), covar=tensor([0.0505, 0.0473, 0.1537, 0.0831, 0.0552, 0.0428, 0.0490, 0.1063], + device='cuda:0'), in_proj_covar=tensor([0.0513, 0.0578, 0.0724, 0.0450, 0.0473, 0.0523, 0.0561, 0.0695], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 20:21:33,885 INFO [train.py:898] (0/4) Epoch 26, batch 2500, loss[loss=0.1372, simple_loss=0.225, pruned_loss=0.02469, over 18551.00 frames. ], tot_loss[loss=0.1569, simple_loss=0.2474, pruned_loss=0.03325, over 3593741.78 frames. ], batch size: 49, lr: 4.31e-03, grad_scale: 8.0 +2023-03-09 20:21:52,365 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93366.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:22:00,219 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2153, 5.1992, 4.8075, 5.1667, 5.1625, 4.5617, 5.0216, 4.8486], + device='cuda:0'), covar=tensor([0.0506, 0.0519, 0.1515, 0.0754, 0.0599, 0.0444, 0.0514, 0.1084], + device='cuda:0'), in_proj_covar=tensor([0.0514, 0.0579, 0.0724, 0.0451, 0.0473, 0.0525, 0.0563, 0.0696], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 20:22:03,727 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93376.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:22:22,156 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93392.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 20:22:32,980 INFO [train.py:898] (0/4) Epoch 26, batch 2550, loss[loss=0.1315, simple_loss=0.2176, pruned_loss=0.02267, over 18424.00 frames. ], tot_loss[loss=0.1569, simple_loss=0.2473, pruned_loss=0.03327, over 3589173.42 frames. ], batch size: 43, lr: 4.30e-03, grad_scale: 8.0 +2023-03-09 20:22:36,875 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.774e+02 2.484e+02 3.027e+02 3.757e+02 6.388e+02, threshold=6.054e+02, percent-clipped=1.0 +2023-03-09 20:23:00,378 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=93424.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:23:31,986 INFO [train.py:898] (0/4) Epoch 26, batch 2600, loss[loss=0.169, simple_loss=0.2643, pruned_loss=0.03682, over 18495.00 frames. ], tot_loss[loss=0.157, simple_loss=0.2474, pruned_loss=0.03326, over 3586775.93 frames. ], batch size: 53, lr: 4.30e-03, grad_scale: 8.0 +2023-03-09 20:23:42,728 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8111, 3.2367, 3.9475, 2.8677, 3.6625, 2.5709, 2.7107, 2.2986], + device='cuda:0'), covar=tensor([0.1084, 0.0954, 0.0322, 0.0793, 0.0617, 0.2322, 0.2363, 0.1750], + device='cuda:0'), in_proj_covar=tensor([0.0226, 0.0248, 0.0218, 0.0204, 0.0263, 0.0274, 0.0333, 0.0241], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 20:23:55,853 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=93471.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:23:57,531 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.21 vs. limit=2.0 +2023-03-09 20:24:14,423 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93487.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:24:30,091 INFO [train.py:898] (0/4) Epoch 26, batch 2650, loss[loss=0.1288, simple_loss=0.2115, pruned_loss=0.02307, over 18496.00 frames. ], tot_loss[loss=0.157, simple_loss=0.2472, pruned_loss=0.03339, over 3573459.59 frames. ], batch size: 44, lr: 4.30e-03, grad_scale: 8.0 +2023-03-09 20:24:34,070 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.541e+02 2.517e+02 2.981e+02 3.680e+02 1.070e+03, threshold=5.961e+02, percent-clipped=2.0 +2023-03-09 20:25:09,832 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=93535.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:25:27,947 INFO [train.py:898] (0/4) Epoch 26, batch 2700, loss[loss=0.157, simple_loss=0.2545, pruned_loss=0.02974, over 17038.00 frames. ], tot_loss[loss=0.1569, simple_loss=0.2472, pruned_loss=0.03327, over 3583606.60 frames. ], batch size: 78, lr: 4.30e-03, grad_scale: 16.0 +2023-03-09 20:25:32,729 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93555.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:25:38,182 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9227, 3.7303, 5.0410, 4.5596, 3.4031, 3.1590, 4.4621, 5.3076], + device='cuda:0'), covar=tensor([0.0777, 0.1596, 0.0235, 0.0361, 0.0998, 0.1160, 0.0391, 0.0286], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0279, 0.0167, 0.0184, 0.0193, 0.0195, 0.0198, 0.0208], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 20:26:03,757 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5610, 2.2583, 2.4841, 2.5123, 3.0327, 4.4310, 4.4350, 3.1865], + device='cuda:0'), covar=tensor([0.2123, 0.2809, 0.3021, 0.2131, 0.2592, 0.0360, 0.0424, 0.1105], + device='cuda:0'), in_proj_covar=tensor([0.0323, 0.0359, 0.0404, 0.0286, 0.0393, 0.0262, 0.0300, 0.0270], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 20:26:10,311 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93587.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:26:11,644 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5588, 3.3811, 2.2568, 4.3969, 3.0303, 4.0735, 2.4393, 3.8340], + device='cuda:0'), covar=tensor([0.0653, 0.0865, 0.1462, 0.0430, 0.0856, 0.0372, 0.1234, 0.0447], + device='cuda:0'), in_proj_covar=tensor([0.0222, 0.0231, 0.0195, 0.0294, 0.0197, 0.0271, 0.0206, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 20:26:18,675 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.54 vs. limit=5.0 +2023-03-09 20:26:26,586 INFO [train.py:898] (0/4) Epoch 26, batch 2750, loss[loss=0.1819, simple_loss=0.2727, pruned_loss=0.04555, over 12217.00 frames. ], tot_loss[loss=0.1576, simple_loss=0.248, pruned_loss=0.03361, over 3569716.98 frames. ], batch size: 130, lr: 4.30e-03, grad_scale: 16.0 +2023-03-09 20:26:28,998 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=93603.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:26:29,919 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.721e+02 2.428e+02 2.838e+02 3.417e+02 1.067e+03, threshold=5.676e+02, percent-clipped=2.0 +2023-03-09 20:26:58,964 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5704, 6.0860, 5.6357, 5.9280, 5.7494, 5.5436, 6.1576, 6.1009], + device='cuda:0'), covar=tensor([0.1098, 0.0737, 0.0412, 0.0654, 0.1276, 0.0716, 0.0578, 0.0679], + device='cuda:0'), in_proj_covar=tensor([0.0628, 0.0554, 0.0398, 0.0574, 0.0773, 0.0571, 0.0785, 0.0598], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 20:27:02,340 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93631.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:27:06,745 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=93635.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:27:25,265 INFO [train.py:898] (0/4) Epoch 26, batch 2800, loss[loss=0.1423, simple_loss=0.2211, pruned_loss=0.03175, over 17227.00 frames. ], tot_loss[loss=0.1573, simple_loss=0.2475, pruned_loss=0.03353, over 3580274.57 frames. ], batch size: 38, lr: 4.30e-03, grad_scale: 16.0 +2023-03-09 20:27:43,716 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93666.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:27:58,593 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=93679.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:28:13,481 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93692.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:28:23,639 INFO [train.py:898] (0/4) Epoch 26, batch 2850, loss[loss=0.1986, simple_loss=0.274, pruned_loss=0.0616, over 12607.00 frames. ], tot_loss[loss=0.1571, simple_loss=0.2473, pruned_loss=0.03342, over 3582954.87 frames. ], batch size: 132, lr: 4.30e-03, grad_scale: 16.0 +2023-03-09 20:28:27,049 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.715e+02 2.471e+02 3.010e+02 3.579e+02 6.367e+02, threshold=6.020e+02, percent-clipped=3.0 +2023-03-09 20:28:38,554 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=93714.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:28:54,346 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-09 20:29:00,447 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6135, 4.2254, 4.1520, 3.2250, 3.5920, 3.3408, 2.5286, 2.4268], + device='cuda:0'), covar=tensor([0.0275, 0.0180, 0.0107, 0.0336, 0.0384, 0.0258, 0.0733, 0.0873], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0064, 0.0068, 0.0071, 0.0092, 0.0070, 0.0079, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 20:29:07,606 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-09 20:29:09,307 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=93740.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:29:21,126 INFO [train.py:898] (0/4) Epoch 26, batch 2900, loss[loss=0.1292, simple_loss=0.2106, pruned_loss=0.02384, over 18389.00 frames. ], tot_loss[loss=0.1571, simple_loss=0.2475, pruned_loss=0.0333, over 3585615.98 frames. ], batch size: 42, lr: 4.30e-03, grad_scale: 16.0 +2023-03-09 20:29:21,362 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1395, 5.2485, 5.3987, 5.4618, 5.0853, 5.9749, 5.5955, 5.2285], + device='cuda:0'), covar=tensor([0.1108, 0.0647, 0.0688, 0.0816, 0.1539, 0.0696, 0.0756, 0.1790], + device='cuda:0'), in_proj_covar=tensor([0.0366, 0.0302, 0.0322, 0.0330, 0.0341, 0.0437, 0.0293, 0.0430], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 20:29:37,146 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.76 vs. limit=2.0 +2023-03-09 20:29:44,545 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93770.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:29:45,657 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=93771.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:29:46,037 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.49 vs. limit=2.0 +2023-03-09 20:30:19,938 INFO [train.py:898] (0/4) Epoch 26, batch 2950, loss[loss=0.1588, simple_loss=0.2481, pruned_loss=0.03475, over 18404.00 frames. ], tot_loss[loss=0.1574, simple_loss=0.2481, pruned_loss=0.03331, over 3596706.84 frames. ], batch size: 50, lr: 4.30e-03, grad_scale: 16.0 +2023-03-09 20:30:23,986 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.766e+02 2.413e+02 2.765e+02 3.459e+02 6.770e+02, threshold=5.531e+02, percent-clipped=1.0 +2023-03-09 20:30:41,049 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=93819.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:30:56,178 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=93831.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:31:01,022 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-09 20:31:18,478 INFO [train.py:898] (0/4) Epoch 26, batch 3000, loss[loss=0.1436, simple_loss=0.224, pruned_loss=0.03159, over 18517.00 frames. ], tot_loss[loss=0.1566, simple_loss=0.2473, pruned_loss=0.03295, over 3604533.30 frames. ], batch size: 44, lr: 4.29e-03, grad_scale: 16.0 +2023-03-09 20:31:18,480 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 20:31:30,355 INFO [train.py:932] (0/4) Epoch 26, validation: loss=0.15, simple_loss=0.2481, pruned_loss=0.02599, over 944034.00 frames. +2023-03-09 20:31:30,356 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 20:32:28,919 INFO [train.py:898] (0/4) Epoch 26, batch 3050, loss[loss=0.1419, simple_loss=0.2206, pruned_loss=0.0316, over 18434.00 frames. ], tot_loss[loss=0.1561, simple_loss=0.2468, pruned_loss=0.03275, over 3611512.03 frames. ], batch size: 43, lr: 4.29e-03, grad_scale: 16.0 +2023-03-09 20:32:30,373 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9140, 3.8553, 3.8179, 3.3306, 3.6395, 2.9679, 3.0390, 3.8925], + device='cuda:0'), covar=tensor([0.0056, 0.0077, 0.0070, 0.0137, 0.0085, 0.0193, 0.0179, 0.0070], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0172, 0.0145, 0.0194, 0.0152, 0.0186, 0.0190, 0.0131], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 20:32:32,222 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.657e+02 2.555e+02 3.073e+02 3.527e+02 8.787e+02, threshold=6.146e+02, percent-clipped=2.0 +2023-03-09 20:32:35,493 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-09 20:33:08,569 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7499, 2.4601, 2.6955, 2.8123, 3.1896, 4.8507, 4.8856, 3.4150], + device='cuda:0'), covar=tensor([0.2012, 0.2480, 0.3151, 0.1857, 0.2525, 0.0255, 0.0331, 0.1007], + device='cuda:0'), in_proj_covar=tensor([0.0322, 0.0358, 0.0402, 0.0285, 0.0392, 0.0261, 0.0300, 0.0268], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 20:33:18,610 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8878, 4.5606, 4.5762, 3.5598, 3.8039, 3.5466, 2.8235, 2.6875], + device='cuda:0'), covar=tensor([0.0229, 0.0162, 0.0087, 0.0297, 0.0362, 0.0236, 0.0670, 0.0803], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0063, 0.0067, 0.0071, 0.0091, 0.0070, 0.0078, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 20:33:27,970 INFO [train.py:898] (0/4) Epoch 26, batch 3100, loss[loss=0.1521, simple_loss=0.2332, pruned_loss=0.03549, over 17664.00 frames. ], tot_loss[loss=0.1563, simple_loss=0.247, pruned_loss=0.03283, over 3595654.19 frames. ], batch size: 39, lr: 4.29e-03, grad_scale: 16.0 +2023-03-09 20:33:34,374 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=3.02 vs. limit=5.0 +2023-03-09 20:34:03,824 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=93981.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:34:25,888 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-94000.pt +2023-03-09 20:34:31,188 INFO [train.py:898] (0/4) Epoch 26, batch 3150, loss[loss=0.1516, simple_loss=0.2388, pruned_loss=0.03222, over 18260.00 frames. ], tot_loss[loss=0.1558, simple_loss=0.2466, pruned_loss=0.03251, over 3599414.34 frames. ], batch size: 47, lr: 4.29e-03, grad_scale: 16.0 +2023-03-09 20:34:34,547 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.522e+02 2.445e+02 2.866e+02 3.470e+02 6.892e+02, threshold=5.732e+02, percent-clipped=1.0 +2023-03-09 20:34:55,543 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.67 vs. limit=2.0 +2023-03-09 20:35:19,130 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=94042.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:35:29,134 INFO [train.py:898] (0/4) Epoch 26, batch 3200, loss[loss=0.167, simple_loss=0.2598, pruned_loss=0.03716, over 18381.00 frames. ], tot_loss[loss=0.1562, simple_loss=0.2468, pruned_loss=0.03283, over 3591078.57 frames. ], batch size: 50, lr: 4.29e-03, grad_scale: 16.0 +2023-03-09 20:36:27,601 INFO [train.py:898] (0/4) Epoch 26, batch 3250, loss[loss=0.1375, simple_loss=0.2215, pruned_loss=0.02671, over 17625.00 frames. ], tot_loss[loss=0.1562, simple_loss=0.2467, pruned_loss=0.03288, over 3593193.32 frames. ], batch size: 39, lr: 4.29e-03, grad_scale: 16.0 +2023-03-09 20:36:31,044 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.794e+02 2.553e+02 3.024e+02 3.724e+02 8.014e+02, threshold=6.047e+02, percent-clipped=2.0 +2023-03-09 20:36:41,524 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.81 vs. limit=2.0 +2023-03-09 20:36:42,534 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-09 20:36:57,633 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94126.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:37:26,333 INFO [train.py:898] (0/4) Epoch 26, batch 3300, loss[loss=0.1387, simple_loss=0.2326, pruned_loss=0.02245, over 18245.00 frames. ], tot_loss[loss=0.1558, simple_loss=0.2462, pruned_loss=0.03266, over 3595185.29 frames. ], batch size: 45, lr: 4.29e-03, grad_scale: 16.0 +2023-03-09 20:37:59,480 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6198, 3.5985, 3.4568, 3.1333, 3.4014, 2.7589, 2.7204, 3.6377], + device='cuda:0'), covar=tensor([0.0078, 0.0104, 0.0100, 0.0146, 0.0099, 0.0212, 0.0251, 0.0083], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0174, 0.0145, 0.0196, 0.0153, 0.0187, 0.0192, 0.0132], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 20:38:24,387 INFO [train.py:898] (0/4) Epoch 26, batch 3350, loss[loss=0.1354, simple_loss=0.2184, pruned_loss=0.02624, over 17677.00 frames. ], tot_loss[loss=0.1562, simple_loss=0.2465, pruned_loss=0.03293, over 3590528.80 frames. ], batch size: 39, lr: 4.29e-03, grad_scale: 8.0 +2023-03-09 20:38:28,979 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.721e+02 2.569e+02 3.071e+02 3.653e+02 6.732e+02, threshold=6.142e+02, percent-clipped=1.0 +2023-03-09 20:38:50,226 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=94223.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 20:38:54,993 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-09 20:39:23,052 INFO [train.py:898] (0/4) Epoch 26, batch 3400, loss[loss=0.1452, simple_loss=0.2241, pruned_loss=0.03316, over 18402.00 frames. ], tot_loss[loss=0.1564, simple_loss=0.2471, pruned_loss=0.03289, over 3589461.99 frames. ], batch size: 42, lr: 4.29e-03, grad_scale: 8.0 +2023-03-09 20:40:02,575 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=94284.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 20:40:10,743 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1632, 5.6869, 5.2802, 5.4766, 5.3515, 5.1208, 5.7492, 5.7078], + device='cuda:0'), covar=tensor([0.1287, 0.0855, 0.0635, 0.0810, 0.1355, 0.0754, 0.0624, 0.0763], + device='cuda:0'), in_proj_covar=tensor([0.0632, 0.0561, 0.0404, 0.0582, 0.0780, 0.0577, 0.0794, 0.0603], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 20:40:21,758 INFO [train.py:898] (0/4) Epoch 26, batch 3450, loss[loss=0.1592, simple_loss=0.2547, pruned_loss=0.03181, over 18488.00 frames. ], tot_loss[loss=0.1564, simple_loss=0.247, pruned_loss=0.03294, over 3593122.47 frames. ], batch size: 53, lr: 4.28e-03, grad_scale: 8.0 +2023-03-09 20:40:26,279 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.946e+02 2.477e+02 2.921e+02 3.603e+02 6.240e+02, threshold=5.842e+02, percent-clipped=1.0 +2023-03-09 20:41:04,281 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94337.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:41:17,096 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=94348.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:41:20,132 INFO [train.py:898] (0/4) Epoch 26, batch 3500, loss[loss=0.1505, simple_loss=0.2505, pruned_loss=0.02524, over 17044.00 frames. ], tot_loss[loss=0.1569, simple_loss=0.2477, pruned_loss=0.03301, over 3591208.71 frames. ], batch size: 78, lr: 4.28e-03, grad_scale: 8.0 +2023-03-09 20:42:09,863 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9082, 3.8602, 3.7401, 3.2946, 3.6453, 3.0018, 2.9264, 3.9093], + device='cuda:0'), covar=tensor([0.0069, 0.0092, 0.0089, 0.0135, 0.0101, 0.0191, 0.0222, 0.0077], + device='cuda:0'), in_proj_covar=tensor([0.0152, 0.0172, 0.0144, 0.0195, 0.0153, 0.0186, 0.0190, 0.0130], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 20:42:15,619 INFO [train.py:898] (0/4) Epoch 26, batch 3550, loss[loss=0.1794, simple_loss=0.2772, pruned_loss=0.04078, over 17041.00 frames. ], tot_loss[loss=0.1575, simple_loss=0.2483, pruned_loss=0.03335, over 3580307.93 frames. ], batch size: 78, lr: 4.28e-03, grad_scale: 8.0 +2023-03-09 20:42:20,514 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.890e+02 2.549e+02 2.986e+02 3.546e+02 5.816e+02, threshold=5.972e+02, percent-clipped=0.0 +2023-03-09 20:42:25,116 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=94409.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 20:42:43,007 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=94426.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:42:47,533 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6769, 2.4500, 2.5440, 2.7972, 3.1172, 4.9182, 4.8969, 3.1921], + device='cuda:0'), covar=tensor([0.2161, 0.2663, 0.3362, 0.1913, 0.2786, 0.0255, 0.0330, 0.1169], + device='cuda:0'), in_proj_covar=tensor([0.0325, 0.0361, 0.0407, 0.0287, 0.0397, 0.0263, 0.0304, 0.0270], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 20:42:57,306 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-03-09 20:43:09,311 INFO [train.py:898] (0/4) Epoch 26, batch 3600, loss[loss=0.151, simple_loss=0.2412, pruned_loss=0.03035, over 18288.00 frames. ], tot_loss[loss=0.1574, simple_loss=0.2481, pruned_loss=0.03338, over 3581489.66 frames. ], batch size: 49, lr: 4.28e-03, grad_scale: 8.0 +2023-03-09 20:43:10,763 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7726, 3.7344, 3.6216, 3.2473, 3.5269, 3.0084, 2.9156, 3.8016], + device='cuda:0'), covar=tensor([0.0066, 0.0104, 0.0088, 0.0135, 0.0101, 0.0183, 0.0214, 0.0058], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0173, 0.0145, 0.0196, 0.0153, 0.0187, 0.0191, 0.0131], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 20:43:34,191 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=94474.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:43:45,188 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-26.pt +2023-03-09 20:44:11,700 INFO [train.py:898] (0/4) Epoch 27, batch 0, loss[loss=0.1404, simple_loss=0.2321, pruned_loss=0.02429, over 18359.00 frames. ], tot_loss[loss=0.1404, simple_loss=0.2321, pruned_loss=0.02429, over 18359.00 frames. ], batch size: 46, lr: 4.20e-03, grad_scale: 8.0 +2023-03-09 20:44:11,703 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 20:44:23,643 INFO [train.py:932] (0/4) Epoch 27, validation: loss=0.1494, simple_loss=0.2481, pruned_loss=0.02532, over 944034.00 frames. +2023-03-09 20:44:23,644 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 20:44:31,925 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.54 vs. limit=2.0 +2023-03-09 20:44:49,666 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.900e+02 2.525e+02 3.009e+02 3.747e+02 9.938e+02, threshold=6.019e+02, percent-clipped=2.0 +2023-03-09 20:45:06,000 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.90 vs. limit=2.0 +2023-03-09 20:45:13,968 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.94 vs. limit=2.0 +2023-03-09 20:45:22,316 INFO [train.py:898] (0/4) Epoch 27, batch 50, loss[loss=0.1624, simple_loss=0.2576, pruned_loss=0.03362, over 18489.00 frames. ], tot_loss[loss=0.1556, simple_loss=0.2464, pruned_loss=0.03235, over 815954.97 frames. ], batch size: 51, lr: 4.20e-03, grad_scale: 8.0 +2023-03-09 20:46:13,890 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94579.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 20:46:20,482 INFO [train.py:898] (0/4) Epoch 27, batch 100, loss[loss=0.1566, simple_loss=0.2532, pruned_loss=0.03005, over 18261.00 frames. ], tot_loss[loss=0.1554, simple_loss=0.2461, pruned_loss=0.03238, over 1436979.19 frames. ], batch size: 57, lr: 4.20e-03, grad_scale: 8.0 +2023-03-09 20:46:29,205 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-09 20:46:44,575 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6917, 3.5468, 2.3019, 4.4939, 3.2225, 4.2847, 2.7056, 4.0110], + device='cuda:0'), covar=tensor([0.0689, 0.0870, 0.1588, 0.0547, 0.0865, 0.0296, 0.1223, 0.0426], + device='cuda:0'), in_proj_covar=tensor([0.0222, 0.0231, 0.0195, 0.0294, 0.0196, 0.0269, 0.0205, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 20:46:46,512 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.709e+02 2.657e+02 3.166e+02 3.640e+02 6.599e+02, threshold=6.333e+02, percent-clipped=4.0 +2023-03-09 20:47:19,783 INFO [train.py:898] (0/4) Epoch 27, batch 150, loss[loss=0.1387, simple_loss=0.2278, pruned_loss=0.0248, over 18370.00 frames. ], tot_loss[loss=0.1568, simple_loss=0.2476, pruned_loss=0.03294, over 1916167.89 frames. ], batch size: 46, lr: 4.19e-03, grad_scale: 8.0 +2023-03-09 20:47:22,222 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=94637.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:47:25,011 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.33 vs. limit=2.0 +2023-03-09 20:47:59,558 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.99 vs. limit=2.0 +2023-03-09 20:48:14,576 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-09 20:48:17,491 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-09 20:48:17,943 INFO [train.py:898] (0/4) Epoch 27, batch 200, loss[loss=0.1431, simple_loss=0.2353, pruned_loss=0.02541, over 18279.00 frames. ], tot_loss[loss=0.1572, simple_loss=0.2476, pruned_loss=0.03342, over 2283012.41 frames. ], batch size: 47, lr: 4.19e-03, grad_scale: 8.0 +2023-03-09 20:48:18,085 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=94685.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:48:39,866 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=94704.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 20:48:41,888 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.913e+02 2.647e+02 3.071e+02 3.764e+02 1.112e+03, threshold=6.143e+02, percent-clipped=3.0 +2023-03-09 20:49:07,675 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2844, 2.7989, 2.3737, 2.7109, 3.4498, 3.3447, 3.0041, 2.6906], + device='cuda:0'), covar=tensor([0.0232, 0.0292, 0.0590, 0.0405, 0.0242, 0.0174, 0.0394, 0.0434], + device='cuda:0'), in_proj_covar=tensor([0.0147, 0.0146, 0.0171, 0.0166, 0.0144, 0.0129, 0.0164, 0.0168], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 20:49:10,142 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.39 vs. limit=2.0 +2023-03-09 20:49:16,139 INFO [train.py:898] (0/4) Epoch 27, batch 250, loss[loss=0.145, simple_loss=0.2269, pruned_loss=0.03156, over 18494.00 frames. ], tot_loss[loss=0.1571, simple_loss=0.2474, pruned_loss=0.03339, over 2577388.17 frames. ], batch size: 44, lr: 4.19e-03, grad_scale: 8.0 +2023-03-09 20:49:59,142 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7075, 2.5884, 2.5480, 2.6789, 2.9690, 3.7480, 3.6997, 3.1423], + device='cuda:0'), covar=tensor([0.1736, 0.2209, 0.2713, 0.1809, 0.2221, 0.0486, 0.0576, 0.0867], + device='cuda:0'), in_proj_covar=tensor([0.0325, 0.0361, 0.0407, 0.0289, 0.0396, 0.0264, 0.0304, 0.0270], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 20:50:14,304 INFO [train.py:898] (0/4) Epoch 27, batch 300, loss[loss=0.1686, simple_loss=0.2632, pruned_loss=0.03704, over 15982.00 frames. ], tot_loss[loss=0.1574, simple_loss=0.2478, pruned_loss=0.0335, over 2791331.03 frames. ], batch size: 94, lr: 4.19e-03, grad_scale: 8.0 +2023-03-09 20:50:34,950 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7611, 5.5788, 5.2238, 5.4051, 4.9998, 5.2827, 5.7506, 5.5901], + device='cuda:0'), covar=tensor([0.2482, 0.1253, 0.0925, 0.1202, 0.2319, 0.1313, 0.0887, 0.1227], + device='cuda:0'), in_proj_covar=tensor([0.0631, 0.0560, 0.0405, 0.0579, 0.0777, 0.0577, 0.0793, 0.0601], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 20:50:38,014 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.940e+02 2.501e+02 2.879e+02 3.354e+02 8.790e+02, threshold=5.757e+02, percent-clipped=2.0 +2023-03-09 20:51:12,790 INFO [train.py:898] (0/4) Epoch 27, batch 350, loss[loss=0.1385, simple_loss=0.2282, pruned_loss=0.02439, over 18356.00 frames. ], tot_loss[loss=0.1574, simple_loss=0.2481, pruned_loss=0.03339, over 2967445.44 frames. ], batch size: 46, lr: 4.19e-03, grad_scale: 8.0 +2023-03-09 20:51:22,259 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7360, 4.1799, 2.4719, 4.0853, 5.1436, 2.5852, 3.8838, 4.1723], + device='cuda:0'), covar=tensor([0.0213, 0.0846, 0.1539, 0.0565, 0.0094, 0.1180, 0.0652, 0.0588], + device='cuda:0'), in_proj_covar=tensor([0.0182, 0.0278, 0.0207, 0.0200, 0.0139, 0.0186, 0.0221, 0.0230], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 20:52:01,044 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9385, 4.2479, 2.6574, 3.9959, 5.2566, 2.8329, 3.9809, 4.0111], + device='cuda:0'), covar=tensor([0.0196, 0.1088, 0.1554, 0.0742, 0.0092, 0.1204, 0.0626, 0.0736], + device='cuda:0'), in_proj_covar=tensor([0.0183, 0.0279, 0.0208, 0.0201, 0.0140, 0.0187, 0.0221, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 20:52:04,469 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=94879.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 20:52:11,467 INFO [train.py:898] (0/4) Epoch 27, batch 400, loss[loss=0.1687, simple_loss=0.2629, pruned_loss=0.03727, over 16035.00 frames. ], tot_loss[loss=0.1578, simple_loss=0.2489, pruned_loss=0.03338, over 3111969.52 frames. ], batch size: 94, lr: 4.19e-03, grad_scale: 8.0 +2023-03-09 20:52:35,456 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.622e+02 2.561e+02 2.993e+02 3.675e+02 9.257e+02, threshold=5.986e+02, percent-clipped=2.0 +2023-03-09 20:52:54,461 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7966, 5.3117, 2.6042, 5.1859, 5.0480, 5.3382, 5.1268, 2.5990], + device='cuda:0'), covar=tensor([0.0269, 0.0085, 0.0850, 0.0076, 0.0082, 0.0085, 0.0102, 0.0999], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0083, 0.0098, 0.0099, 0.0089, 0.0079, 0.0086, 0.0097], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 20:53:00,929 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=94927.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 20:53:10,430 INFO [train.py:898] (0/4) Epoch 27, batch 450, loss[loss=0.1509, simple_loss=0.2463, pruned_loss=0.02781, over 18295.00 frames. ], tot_loss[loss=0.1572, simple_loss=0.2482, pruned_loss=0.03312, over 3225672.85 frames. ], batch size: 57, lr: 4.19e-03, grad_scale: 8.0 +2023-03-09 20:53:11,989 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6896, 4.0114, 2.3145, 3.8921, 5.0384, 2.5607, 3.6955, 3.8342], + device='cuda:0'), covar=tensor([0.0214, 0.1083, 0.1707, 0.0685, 0.0108, 0.1230, 0.0681, 0.0741], + device='cuda:0'), in_proj_covar=tensor([0.0181, 0.0276, 0.0206, 0.0199, 0.0139, 0.0185, 0.0219, 0.0229], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 20:53:33,431 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-09 20:54:08,447 INFO [train.py:898] (0/4) Epoch 27, batch 500, loss[loss=0.1399, simple_loss=0.2311, pruned_loss=0.02433, over 18250.00 frames. ], tot_loss[loss=0.1569, simple_loss=0.2478, pruned_loss=0.03301, over 3316830.38 frames. ], batch size: 47, lr: 4.19e-03, grad_scale: 8.0 +2023-03-09 20:54:31,066 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95004.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:54:33,123 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.646e+02 2.379e+02 2.893e+02 3.506e+02 5.482e+02, threshold=5.786e+02, percent-clipped=0.0 +2023-03-09 20:55:06,356 INFO [train.py:898] (0/4) Epoch 27, batch 550, loss[loss=0.1735, simple_loss=0.2629, pruned_loss=0.042, over 18487.00 frames. ], tot_loss[loss=0.1571, simple_loss=0.2478, pruned_loss=0.03319, over 3380185.29 frames. ], batch size: 53, lr: 4.19e-03, grad_scale: 8.0 +2023-03-09 20:55:26,979 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=95052.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:56:04,278 INFO [train.py:898] (0/4) Epoch 27, batch 600, loss[loss=0.1485, simple_loss=0.2432, pruned_loss=0.02684, over 18286.00 frames. ], tot_loss[loss=0.157, simple_loss=0.2479, pruned_loss=0.03306, over 3438063.23 frames. ], batch size: 49, lr: 4.19e-03, grad_scale: 8.0 +2023-03-09 20:56:19,123 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9029, 5.4306, 2.7704, 5.3156, 5.2080, 5.4603, 5.3139, 2.7324], + device='cuda:0'), covar=tensor([0.0244, 0.0070, 0.0782, 0.0072, 0.0071, 0.0075, 0.0079, 0.0938], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0084, 0.0099, 0.0099, 0.0090, 0.0080, 0.0087, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 20:56:28,696 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95105.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 20:56:29,342 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.684e+02 2.616e+02 3.098e+02 3.741e+02 7.084e+02, threshold=6.196e+02, percent-clipped=4.0 +2023-03-09 20:56:57,768 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.78 vs. limit=5.0 +2023-03-09 20:57:02,881 INFO [train.py:898] (0/4) Epoch 27, batch 650, loss[loss=0.1746, simple_loss=0.259, pruned_loss=0.04505, over 12862.00 frames. ], tot_loss[loss=0.1568, simple_loss=0.2475, pruned_loss=0.03305, over 3471088.90 frames. ], batch size: 129, lr: 4.18e-03, grad_scale: 8.0 +2023-03-09 20:57:08,292 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5730, 6.1417, 5.7646, 5.9594, 5.7892, 5.5960, 6.2041, 6.1329], + device='cuda:0'), covar=tensor([0.1254, 0.0781, 0.0409, 0.0663, 0.1263, 0.0674, 0.0564, 0.0694], + device='cuda:0'), in_proj_covar=tensor([0.0633, 0.0557, 0.0402, 0.0577, 0.0778, 0.0575, 0.0790, 0.0601], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 20:57:40,181 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95166.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 20:58:02,040 INFO [train.py:898] (0/4) Epoch 27, batch 700, loss[loss=0.1482, simple_loss=0.2387, pruned_loss=0.02886, over 18500.00 frames. ], tot_loss[loss=0.1574, simple_loss=0.2481, pruned_loss=0.03332, over 3495406.21 frames. ], batch size: 51, lr: 4.18e-03, grad_scale: 8.0 +2023-03-09 20:58:11,759 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7217, 2.4091, 2.6588, 2.7208, 3.2387, 4.8560, 4.7927, 3.2965], + device='cuda:0'), covar=tensor([0.1968, 0.2532, 0.2987, 0.1969, 0.2467, 0.0245, 0.0335, 0.1098], + device='cuda:0'), in_proj_covar=tensor([0.0327, 0.0362, 0.0408, 0.0290, 0.0398, 0.0265, 0.0305, 0.0271], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0002, 0.0001], + device='cuda:0') +2023-03-09 20:58:28,145 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.730e+02 2.528e+02 2.858e+02 3.402e+02 6.027e+02, threshold=5.717e+02, percent-clipped=1.0 +2023-03-09 20:59:01,287 INFO [train.py:898] (0/4) Epoch 27, batch 750, loss[loss=0.1541, simple_loss=0.2462, pruned_loss=0.03096, over 18480.00 frames. ], tot_loss[loss=0.1571, simple_loss=0.2475, pruned_loss=0.03339, over 3519345.61 frames. ], batch size: 51, lr: 4.18e-03, grad_scale: 8.0 +2023-03-09 20:59:31,747 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95260.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 20:59:37,263 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2091, 5.2341, 5.2765, 4.9864, 5.0489, 5.0451, 5.3489, 5.3510], + device='cuda:0'), covar=tensor([0.0064, 0.0060, 0.0055, 0.0107, 0.0052, 0.0163, 0.0072, 0.0083], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0074, 0.0080, 0.0100, 0.0079, 0.0108, 0.0091, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 20:59:40,853 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95268.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:00:00,185 INFO [train.py:898] (0/4) Epoch 27, batch 800, loss[loss=0.1547, simple_loss=0.2459, pruned_loss=0.0318, over 18389.00 frames. ], tot_loss[loss=0.1569, simple_loss=0.2469, pruned_loss=0.03344, over 3527265.52 frames. ], batch size: 52, lr: 4.18e-03, grad_scale: 8.0 +2023-03-09 21:00:25,350 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.793e+02 2.511e+02 2.986e+02 3.454e+02 7.557e+02, threshold=5.973e+02, percent-clipped=4.0 +2023-03-09 21:00:43,452 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95321.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:00:52,344 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9639, 5.4453, 5.4708, 5.4822, 4.9176, 5.3962, 4.6376, 5.3303], + device='cuda:0'), covar=tensor([0.0292, 0.0349, 0.0234, 0.0431, 0.0438, 0.0249, 0.1439, 0.0396], + device='cuda:0'), in_proj_covar=tensor([0.0236, 0.0277, 0.0279, 0.0357, 0.0287, 0.0286, 0.0321, 0.0281], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 21:00:52,409 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95329.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:00:58,807 INFO [train.py:898] (0/4) Epoch 27, batch 850, loss[loss=0.1437, simple_loss=0.2318, pruned_loss=0.02777, over 18525.00 frames. ], tot_loss[loss=0.1569, simple_loss=0.2473, pruned_loss=0.03324, over 3550832.68 frames. ], batch size: 49, lr: 4.18e-03, grad_scale: 8.0 +2023-03-09 21:01:04,255 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95339.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:01:57,815 INFO [train.py:898] (0/4) Epoch 27, batch 900, loss[loss=0.1605, simple_loss=0.2485, pruned_loss=0.0363, over 18500.00 frames. ], tot_loss[loss=0.1568, simple_loss=0.2472, pruned_loss=0.03326, over 3561590.04 frames. ], batch size: 47, lr: 4.18e-03, grad_scale: 8.0 +2023-03-09 21:02:10,617 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5205, 2.3077, 2.4805, 2.5042, 2.9790, 4.4645, 4.3873, 3.0950], + device='cuda:0'), covar=tensor([0.2182, 0.2772, 0.3110, 0.2180, 0.2628, 0.0360, 0.0435, 0.1179], + device='cuda:0'), in_proj_covar=tensor([0.0327, 0.0362, 0.0408, 0.0290, 0.0397, 0.0265, 0.0305, 0.0270], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 21:02:16,704 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95400.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:02:24,175 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.646e+02 2.410e+02 2.846e+02 3.556e+02 5.953e+02, threshold=5.693e+02, percent-clipped=0.0 +2023-03-09 21:02:57,605 INFO [train.py:898] (0/4) Epoch 27, batch 950, loss[loss=0.1696, simple_loss=0.2601, pruned_loss=0.03957, over 18093.00 frames. ], tot_loss[loss=0.157, simple_loss=0.2475, pruned_loss=0.03322, over 3562486.26 frames. ], batch size: 62, lr: 4.18e-03, grad_scale: 8.0 +2023-03-09 21:03:28,549 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95461.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 21:03:55,823 INFO [train.py:898] (0/4) Epoch 27, batch 1000, loss[loss=0.1555, simple_loss=0.2458, pruned_loss=0.03258, over 18621.00 frames. ], tot_loss[loss=0.1578, simple_loss=0.2484, pruned_loss=0.03356, over 3557952.36 frames. ], batch size: 52, lr: 4.18e-03, grad_scale: 8.0 +2023-03-09 21:04:19,641 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.888e+02 2.625e+02 2.944e+02 3.804e+02 7.534e+02, threshold=5.888e+02, percent-clipped=3.0 +2023-03-09 21:04:40,492 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.6926, 6.1852, 5.6642, 6.0087, 5.8448, 5.6377, 6.2580, 6.2285], + device='cuda:0'), covar=tensor([0.1128, 0.0758, 0.0463, 0.0666, 0.1350, 0.0655, 0.0582, 0.0630], + device='cuda:0'), in_proj_covar=tensor([0.0630, 0.0557, 0.0402, 0.0577, 0.0775, 0.0578, 0.0787, 0.0599], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 21:04:53,990 INFO [train.py:898] (0/4) Epoch 27, batch 1050, loss[loss=0.1781, simple_loss=0.28, pruned_loss=0.03808, over 18274.00 frames. ], tot_loss[loss=0.158, simple_loss=0.2488, pruned_loss=0.03362, over 3567375.00 frames. ], batch size: 57, lr: 4.18e-03, grad_scale: 8.0 +2023-03-09 21:05:19,299 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.91 vs. limit=2.0 +2023-03-09 21:05:53,036 INFO [train.py:898] (0/4) Epoch 27, batch 1100, loss[loss=0.1566, simple_loss=0.2386, pruned_loss=0.03734, over 17625.00 frames. ], tot_loss[loss=0.1569, simple_loss=0.2474, pruned_loss=0.0332, over 3585040.57 frames. ], batch size: 39, lr: 4.17e-03, grad_scale: 8.0 +2023-03-09 21:06:17,708 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.642e+02 2.398e+02 2.884e+02 3.392e+02 9.491e+02, threshold=5.768e+02, percent-clipped=3.0 +2023-03-09 21:06:29,777 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95616.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:06:31,636 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95617.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:06:39,702 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95624.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:06:43,697 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2717, 5.2459, 4.9103, 5.1940, 5.2083, 4.5841, 5.1027, 4.8631], + device='cuda:0'), covar=tensor([0.0453, 0.0416, 0.1228, 0.0768, 0.0536, 0.0436, 0.0441, 0.1108], + device='cuda:0'), in_proj_covar=tensor([0.0515, 0.0584, 0.0725, 0.0450, 0.0473, 0.0530, 0.0564, 0.0701], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 21:06:52,364 INFO [train.py:898] (0/4) Epoch 27, batch 1150, loss[loss=0.1696, simple_loss=0.2627, pruned_loss=0.03826, over 17726.00 frames. ], tot_loss[loss=0.1569, simple_loss=0.2474, pruned_loss=0.03318, over 3593024.40 frames. ], batch size: 70, lr: 4.17e-03, grad_scale: 4.0 +2023-03-09 21:07:43,520 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95678.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:07:51,487 INFO [train.py:898] (0/4) Epoch 27, batch 1200, loss[loss=0.1698, simple_loss=0.2601, pruned_loss=0.03971, over 18353.00 frames. ], tot_loss[loss=0.1566, simple_loss=0.2474, pruned_loss=0.03293, over 3599618.18 frames. ], batch size: 56, lr: 4.17e-03, grad_scale: 8.0 +2023-03-09 21:08:01,807 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5039, 6.0535, 5.6190, 5.8579, 5.6646, 5.4866, 6.1090, 6.0899], + device='cuda:0'), covar=tensor([0.1221, 0.0779, 0.0459, 0.0703, 0.1378, 0.0710, 0.0566, 0.0680], + device='cuda:0'), in_proj_covar=tensor([0.0634, 0.0559, 0.0403, 0.0580, 0.0780, 0.0582, 0.0790, 0.0604], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 21:08:02,865 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95695.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:08:06,025 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-09 21:08:16,301 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.614e+02 2.547e+02 2.881e+02 3.580e+02 8.977e+02, threshold=5.762e+02, percent-clipped=2.0 +2023-03-09 21:08:50,126 INFO [train.py:898] (0/4) Epoch 27, batch 1250, loss[loss=0.1418, simple_loss=0.2248, pruned_loss=0.02944, over 18452.00 frames. ], tot_loss[loss=0.1566, simple_loss=0.2474, pruned_loss=0.0329, over 3598444.16 frames. ], batch size: 43, lr: 4.17e-03, grad_scale: 8.0 +2023-03-09 21:09:02,203 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3814, 2.7141, 2.5104, 2.7083, 3.4980, 3.3564, 3.0041, 2.7189], + device='cuda:0'), covar=tensor([0.0279, 0.0341, 0.0626, 0.0448, 0.0225, 0.0197, 0.0413, 0.0467], + device='cuda:0'), in_proj_covar=tensor([0.0147, 0.0146, 0.0170, 0.0164, 0.0142, 0.0128, 0.0162, 0.0166], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 21:09:20,645 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95761.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 21:09:49,132 INFO [train.py:898] (0/4) Epoch 27, batch 1300, loss[loss=0.1653, simple_loss=0.2488, pruned_loss=0.0409, over 18556.00 frames. ], tot_loss[loss=0.1572, simple_loss=0.2482, pruned_loss=0.03306, over 3601770.25 frames. ], batch size: 49, lr: 4.17e-03, grad_scale: 8.0 +2023-03-09 21:09:49,615 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8558, 3.6878, 4.9795, 4.4107, 3.4475, 3.2045, 4.5391, 5.2660], + device='cuda:0'), covar=tensor([0.0819, 0.1515, 0.0218, 0.0417, 0.0919, 0.1165, 0.0368, 0.0217], + device='cuda:0'), in_proj_covar=tensor([0.0155, 0.0282, 0.0170, 0.0186, 0.0196, 0.0198, 0.0200, 0.0211], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 21:09:52,359 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0374, 2.2362, 3.1546, 3.1403, 2.1875, 3.3469, 3.1408, 2.3099], + device='cuda:0'), covar=tensor([0.0559, 0.1792, 0.0582, 0.0432, 0.1862, 0.0389, 0.0831, 0.1212], + device='cuda:0'), in_proj_covar=tensor([0.0221, 0.0245, 0.0232, 0.0172, 0.0228, 0.0219, 0.0258, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 21:10:15,109 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.836e+02 2.534e+02 2.954e+02 3.886e+02 9.660e+02, threshold=5.908e+02, percent-clipped=7.0 +2023-03-09 21:10:17,506 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=95809.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 21:10:21,609 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95812.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 21:10:48,198 INFO [train.py:898] (0/4) Epoch 27, batch 1350, loss[loss=0.1366, simple_loss=0.2269, pruned_loss=0.02318, over 18548.00 frames. ], tot_loss[loss=0.1574, simple_loss=0.2484, pruned_loss=0.03321, over 3581274.12 frames. ], batch size: 49, lr: 4.17e-03, grad_scale: 8.0 +2023-03-09 21:11:02,586 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95847.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:11:32,161 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95873.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 21:11:46,333 INFO [train.py:898] (0/4) Epoch 27, batch 1400, loss[loss=0.1647, simple_loss=0.2569, pruned_loss=0.03624, over 13018.00 frames. ], tot_loss[loss=0.1577, simple_loss=0.2485, pruned_loss=0.03348, over 3579611.16 frames. ], batch size: 129, lr: 4.17e-03, grad_scale: 8.0 +2023-03-09 21:12:11,667 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.788e+02 2.502e+02 2.884e+02 3.426e+02 1.018e+03, threshold=5.768e+02, percent-clipped=3.0 +2023-03-09 21:12:13,302 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=95908.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 21:12:22,281 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95916.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:12:31,971 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95924.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:12:45,207 INFO [train.py:898] (0/4) Epoch 27, batch 1450, loss[loss=0.1338, simple_loss=0.2234, pruned_loss=0.02208, over 18566.00 frames. ], tot_loss[loss=0.157, simple_loss=0.2478, pruned_loss=0.03314, over 3580799.57 frames. ], batch size: 49, lr: 4.17e-03, grad_scale: 8.0 +2023-03-09 21:12:53,638 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5212, 3.7127, 4.9619, 4.2917, 2.8012, 2.6504, 4.1204, 5.2234], + device='cuda:0'), covar=tensor([0.0881, 0.1358, 0.0195, 0.0409, 0.1168, 0.1365, 0.0475, 0.0182], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0281, 0.0170, 0.0186, 0.0195, 0.0197, 0.0199, 0.0210], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 21:13:19,370 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=95964.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:13:28,449 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=95972.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:13:30,281 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=95973.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:13:38,265 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=95980.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:13:44,686 INFO [train.py:898] (0/4) Epoch 27, batch 1500, loss[loss=0.1525, simple_loss=0.2522, pruned_loss=0.02639, over 18628.00 frames. ], tot_loss[loss=0.1569, simple_loss=0.248, pruned_loss=0.03294, over 3583895.98 frames. ], batch size: 52, lr: 4.17e-03, grad_scale: 8.0 +2023-03-09 21:13:56,454 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=95995.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:14:02,868 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-96000.pt +2023-03-09 21:14:15,335 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.678e+02 2.536e+02 3.010e+02 3.569e+02 8.325e+02, threshold=6.021e+02, percent-clipped=2.0 +2023-03-09 21:14:23,721 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1136, 2.5855, 2.2801, 2.5677, 3.2273, 3.1142, 2.8570, 2.6248], + device='cuda:0'), covar=tensor([0.0227, 0.0323, 0.0619, 0.0411, 0.0239, 0.0207, 0.0403, 0.0383], + device='cuda:0'), in_proj_covar=tensor([0.0145, 0.0146, 0.0169, 0.0164, 0.0141, 0.0127, 0.0161, 0.0163], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 21:14:48,711 INFO [train.py:898] (0/4) Epoch 27, batch 1550, loss[loss=0.1635, simple_loss=0.2516, pruned_loss=0.03766, over 18639.00 frames. ], tot_loss[loss=0.1567, simple_loss=0.2475, pruned_loss=0.03291, over 3579360.74 frames. ], batch size: 52, lr: 4.16e-03, grad_scale: 8.0 +2023-03-09 21:14:56,508 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96041.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:14:58,678 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=96043.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:15:42,247 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96080.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:15:47,443 INFO [train.py:898] (0/4) Epoch 27, batch 1600, loss[loss=0.15, simple_loss=0.2401, pruned_loss=0.02989, over 18496.00 frames. ], tot_loss[loss=0.1565, simple_loss=0.2473, pruned_loss=0.03289, over 3588600.71 frames. ], batch size: 47, lr: 4.16e-03, grad_scale: 8.0 +2023-03-09 21:15:57,543 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9519, 3.7999, 5.1380, 4.6757, 3.6347, 3.2081, 4.6314, 5.4780], + device='cuda:0'), covar=tensor([0.0775, 0.1427, 0.0198, 0.0340, 0.0842, 0.1135, 0.0356, 0.0175], + device='cuda:0'), in_proj_covar=tensor([0.0155, 0.0284, 0.0171, 0.0187, 0.0196, 0.0199, 0.0201, 0.0212], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 21:16:13,692 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.626e+02 2.554e+02 2.952e+02 3.654e+02 6.643e+02, threshold=5.903e+02, percent-clipped=2.0 +2023-03-09 21:16:31,085 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-09 21:16:45,550 INFO [train.py:898] (0/4) Epoch 27, batch 1650, loss[loss=0.1297, simple_loss=0.2182, pruned_loss=0.02063, over 18255.00 frames. ], tot_loss[loss=0.156, simple_loss=0.2465, pruned_loss=0.03274, over 3597428.21 frames. ], batch size: 45, lr: 4.16e-03, grad_scale: 8.0 +2023-03-09 21:16:53,596 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96141.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:17:21,317 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96165.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:17:24,584 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96168.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 21:17:44,182 INFO [train.py:898] (0/4) Epoch 27, batch 1700, loss[loss=0.1502, simple_loss=0.2338, pruned_loss=0.03336, over 18499.00 frames. ], tot_loss[loss=0.1562, simple_loss=0.2463, pruned_loss=0.03301, over 3587934.30 frames. ], batch size: 44, lr: 4.16e-03, grad_scale: 8.0 +2023-03-09 21:18:06,267 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96203.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 21:18:10,514 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.955e+02 2.460e+02 3.228e+02 3.841e+02 7.050e+02, threshold=6.456e+02, percent-clipped=7.0 +2023-03-09 21:18:32,801 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96226.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:18:42,662 INFO [train.py:898] (0/4) Epoch 27, batch 1750, loss[loss=0.1656, simple_loss=0.2655, pruned_loss=0.03283, over 17978.00 frames. ], tot_loss[loss=0.1564, simple_loss=0.2468, pruned_loss=0.03305, over 3589202.79 frames. ], batch size: 65, lr: 4.16e-03, grad_scale: 8.0 +2023-03-09 21:19:27,941 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96273.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:19:41,558 INFO [train.py:898] (0/4) Epoch 27, batch 1800, loss[loss=0.171, simple_loss=0.2641, pruned_loss=0.0389, over 18358.00 frames. ], tot_loss[loss=0.1561, simple_loss=0.2463, pruned_loss=0.03297, over 3592027.66 frames. ], batch size: 56, lr: 4.16e-03, grad_scale: 8.0 +2023-03-09 21:20:07,994 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.809e+02 2.546e+02 2.957e+02 3.630e+02 5.615e+02, threshold=5.915e+02, percent-clipped=0.0 +2023-03-09 21:20:24,641 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=96321.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:20:28,179 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96324.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:20:40,258 INFO [train.py:898] (0/4) Epoch 27, batch 1850, loss[loss=0.1728, simple_loss=0.2644, pruned_loss=0.04057, over 18302.00 frames. ], tot_loss[loss=0.1564, simple_loss=0.2465, pruned_loss=0.03313, over 3595327.53 frames. ], batch size: 54, lr: 4.16e-03, grad_scale: 8.0 +2023-03-09 21:20:41,786 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96336.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:20:49,303 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-09 21:21:37,996 INFO [train.py:898] (0/4) Epoch 27, batch 1900, loss[loss=0.1536, simple_loss=0.235, pruned_loss=0.03608, over 18269.00 frames. ], tot_loss[loss=0.1567, simple_loss=0.2468, pruned_loss=0.03327, over 3591639.78 frames. ], batch size: 47, lr: 4.16e-03, grad_scale: 8.0 +2023-03-09 21:21:38,408 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96385.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:21:44,122 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96390.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:22:04,759 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.806e+02 2.747e+02 3.313e+02 4.135e+02 7.560e+02, threshold=6.625e+02, percent-clipped=3.0 +2023-03-09 21:22:16,172 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8277, 5.2228, 2.5976, 5.0887, 4.9747, 5.2224, 5.0147, 2.5847], + device='cuda:0'), covar=tensor([0.0250, 0.0074, 0.0857, 0.0078, 0.0079, 0.0102, 0.0102, 0.1049], + device='cuda:0'), in_proj_covar=tensor([0.0092, 0.0083, 0.0098, 0.0098, 0.0090, 0.0080, 0.0086, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 21:22:36,437 INFO [train.py:898] (0/4) Epoch 27, batch 1950, loss[loss=0.1435, simple_loss=0.2359, pruned_loss=0.0255, over 18411.00 frames. ], tot_loss[loss=0.1571, simple_loss=0.2474, pruned_loss=0.03335, over 3601386.35 frames. ], batch size: 48, lr: 4.16e-03, grad_scale: 8.0 +2023-03-09 21:22:37,716 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96436.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:22:55,474 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96451.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:23:15,671 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96468.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 21:23:35,121 INFO [train.py:898] (0/4) Epoch 27, batch 2000, loss[loss=0.1727, simple_loss=0.2659, pruned_loss=0.03969, over 18497.00 frames. ], tot_loss[loss=0.1571, simple_loss=0.2478, pruned_loss=0.03327, over 3597336.16 frames. ], batch size: 59, lr: 4.15e-03, grad_scale: 8.0 +2023-03-09 21:23:56,282 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96503.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:24:00,977 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.734e+02 2.669e+02 3.146e+02 3.749e+02 6.001e+02, threshold=6.292e+02, percent-clipped=0.0 +2023-03-09 21:24:11,842 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=96516.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 21:24:17,451 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96521.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:24:33,570 INFO [train.py:898] (0/4) Epoch 27, batch 2050, loss[loss=0.1583, simple_loss=0.2568, pruned_loss=0.02991, over 18383.00 frames. ], tot_loss[loss=0.1576, simple_loss=0.2485, pruned_loss=0.0334, over 3593277.60 frames. ], batch size: 50, lr: 4.15e-03, grad_scale: 8.0 +2023-03-09 21:24:50,151 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-09 21:24:51,978 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=96551.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:25:07,971 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.58 vs. limit=2.0 +2023-03-09 21:25:15,760 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8224, 4.4333, 4.3611, 3.3219, 3.6156, 3.4645, 2.6357, 2.5494], + device='cuda:0'), covar=tensor([0.0208, 0.0153, 0.0106, 0.0313, 0.0367, 0.0231, 0.0705, 0.0830], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0064, 0.0068, 0.0071, 0.0093, 0.0071, 0.0079, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 21:25:28,733 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4882, 3.9101, 3.7858, 3.1262, 3.3677, 3.2154, 2.4432, 2.3961], + device='cuda:0'), covar=tensor([0.0245, 0.0197, 0.0147, 0.0329, 0.0377, 0.0262, 0.0778, 0.0868], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0064, 0.0068, 0.0071, 0.0093, 0.0071, 0.0079, 0.0085], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 21:25:32,775 INFO [train.py:898] (0/4) Epoch 27, batch 2100, loss[loss=0.1819, simple_loss=0.2821, pruned_loss=0.04088, over 17963.00 frames. ], tot_loss[loss=0.1584, simple_loss=0.2492, pruned_loss=0.03383, over 3582371.81 frames. ], batch size: 65, lr: 4.15e-03, grad_scale: 8.0 +2023-03-09 21:25:50,721 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1855, 3.1239, 2.1164, 3.8281, 2.6969, 3.3346, 2.2333, 3.2093], + device='cuda:0'), covar=tensor([0.0603, 0.0845, 0.1348, 0.0520, 0.0797, 0.0279, 0.1248, 0.0518], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0231, 0.0195, 0.0295, 0.0195, 0.0271, 0.0205, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 21:25:58,864 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.510e+02 2.412e+02 2.976e+02 3.502e+02 7.599e+02, threshold=5.952e+02, percent-clipped=1.0 +2023-03-09 21:26:32,734 INFO [train.py:898] (0/4) Epoch 27, batch 2150, loss[loss=0.1609, simple_loss=0.2483, pruned_loss=0.03673, over 18230.00 frames. ], tot_loss[loss=0.1583, simple_loss=0.2492, pruned_loss=0.03365, over 3581349.64 frames. ], batch size: 60, lr: 4.15e-03, grad_scale: 8.0 +2023-03-09 21:26:34,117 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96636.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:26:39,929 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7404, 5.2109, 5.1547, 5.2608, 4.6626, 5.1086, 4.4895, 5.1057], + device='cuda:0'), covar=tensor([0.0279, 0.0360, 0.0277, 0.0412, 0.0413, 0.0270, 0.1245, 0.0350], + device='cuda:0'), in_proj_covar=tensor([0.0234, 0.0275, 0.0278, 0.0355, 0.0287, 0.0287, 0.0319, 0.0278], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 21:27:26,032 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96680.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:27:30,970 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=96684.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:27:31,867 INFO [train.py:898] (0/4) Epoch 27, batch 2200, loss[loss=0.1427, simple_loss=0.2333, pruned_loss=0.0261, over 18498.00 frames. ], tot_loss[loss=0.1581, simple_loss=0.2491, pruned_loss=0.03358, over 3574058.69 frames. ], batch size: 47, lr: 4.15e-03, grad_scale: 8.0 +2023-03-09 21:27:32,566 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.80 vs. limit=2.0 +2023-03-09 21:27:45,771 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6428, 3.5250, 2.3906, 4.4791, 3.2569, 4.2411, 2.7781, 3.9665], + device='cuda:0'), covar=tensor([0.0645, 0.0745, 0.1315, 0.0410, 0.0707, 0.0278, 0.1044, 0.0441], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0231, 0.0196, 0.0295, 0.0196, 0.0272, 0.0206, 0.0208], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 21:27:56,877 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 2.105e+02 2.631e+02 3.073e+02 3.700e+02 7.959e+02, threshold=6.147e+02, percent-clipped=1.0 +2023-03-09 21:28:29,885 INFO [train.py:898] (0/4) Epoch 27, batch 2250, loss[loss=0.1664, simple_loss=0.2583, pruned_loss=0.03724, over 18318.00 frames. ], tot_loss[loss=0.1583, simple_loss=0.2489, pruned_loss=0.03385, over 3568067.66 frames. ], batch size: 57, lr: 4.15e-03, grad_scale: 8.0 +2023-03-09 21:28:31,863 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96736.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:28:43,178 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=96746.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:28:59,179 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.4304, 3.8162, 2.3426, 3.7160, 4.7352, 2.5907, 3.5704, 3.6041], + device='cuda:0'), covar=tensor([0.0253, 0.1167, 0.1695, 0.0737, 0.0119, 0.1269, 0.0734, 0.0857], + device='cuda:0'), in_proj_covar=tensor([0.0183, 0.0282, 0.0210, 0.0202, 0.0142, 0.0188, 0.0223, 0.0233], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 21:29:18,879 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=96776.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:29:27,594 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=96784.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:29:28,576 INFO [train.py:898] (0/4) Epoch 27, batch 2300, loss[loss=0.1535, simple_loss=0.2508, pruned_loss=0.02806, over 17070.00 frames. ], tot_loss[loss=0.1572, simple_loss=0.2475, pruned_loss=0.03349, over 3561675.18 frames. ], batch size: 78, lr: 4.15e-03, grad_scale: 8.0 +2023-03-09 21:29:53,977 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.778e+02 2.521e+02 2.973e+02 3.966e+02 7.107e+02, threshold=5.946e+02, percent-clipped=4.0 +2023-03-09 21:30:11,233 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96821.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:30:27,082 INFO [train.py:898] (0/4) Epoch 27, batch 2350, loss[loss=0.1434, simple_loss=0.2314, pruned_loss=0.02765, over 18246.00 frames. ], tot_loss[loss=0.1569, simple_loss=0.2473, pruned_loss=0.03324, over 3576847.91 frames. ], batch size: 45, lr: 4.15e-03, grad_scale: 8.0 +2023-03-09 21:30:29,652 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=96837.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:31:07,529 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=96869.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:31:26,987 INFO [train.py:898] (0/4) Epoch 27, batch 2400, loss[loss=0.1616, simple_loss=0.2585, pruned_loss=0.03237, over 18281.00 frames. ], tot_loss[loss=0.1576, simple_loss=0.2485, pruned_loss=0.03337, over 3570078.74 frames. ], batch size: 57, lr: 4.15e-03, grad_scale: 8.0 +2023-03-09 21:31:52,349 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.581e+02 2.380e+02 2.700e+02 3.447e+02 6.459e+02, threshold=5.400e+02, percent-clipped=1.0 +2023-03-09 21:32:24,652 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9861, 4.1433, 2.7616, 4.0030, 5.2703, 2.9181, 4.0251, 4.0964], + device='cuda:0'), covar=tensor([0.0183, 0.1279, 0.1429, 0.0723, 0.0098, 0.1141, 0.0622, 0.0676], + device='cuda:0'), in_proj_covar=tensor([0.0184, 0.0284, 0.0211, 0.0202, 0.0143, 0.0188, 0.0225, 0.0234], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 21:32:25,355 INFO [train.py:898] (0/4) Epoch 27, batch 2450, loss[loss=0.1364, simple_loss=0.2231, pruned_loss=0.02478, over 18519.00 frames. ], tot_loss[loss=0.1567, simple_loss=0.2473, pruned_loss=0.03311, over 3589938.94 frames. ], batch size: 47, lr: 4.14e-03, grad_scale: 8.0 +2023-03-09 21:32:57,926 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.74 vs. limit=2.0 +2023-03-09 21:33:18,783 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=96980.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:33:24,784 INFO [train.py:898] (0/4) Epoch 27, batch 2500, loss[loss=0.1394, simple_loss=0.2234, pruned_loss=0.02769, over 18428.00 frames. ], tot_loss[loss=0.1569, simple_loss=0.2475, pruned_loss=0.0332, over 3596996.16 frames. ], batch size: 43, lr: 4.14e-03, grad_scale: 8.0 +2023-03-09 21:33:42,787 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97000.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:33:50,763 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.816e+02 2.553e+02 2.843e+02 3.348e+02 7.332e+02, threshold=5.685e+02, percent-clipped=3.0 +2023-03-09 21:34:01,177 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3155, 5.2915, 4.9365, 5.1964, 5.2711, 4.6281, 5.1185, 4.8575], + device='cuda:0'), covar=tensor([0.0423, 0.0437, 0.1312, 0.0788, 0.0505, 0.0434, 0.0462, 0.1168], + device='cuda:0'), in_proj_covar=tensor([0.0516, 0.0588, 0.0730, 0.0450, 0.0473, 0.0532, 0.0569, 0.0703], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 21:34:15,101 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=97028.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:34:23,800 INFO [train.py:898] (0/4) Epoch 27, batch 2550, loss[loss=0.1614, simple_loss=0.2539, pruned_loss=0.03445, over 17808.00 frames. ], tot_loss[loss=0.1572, simple_loss=0.2477, pruned_loss=0.0333, over 3599636.49 frames. ], batch size: 70, lr: 4.14e-03, grad_scale: 8.0 +2023-03-09 21:34:36,586 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97046.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:34:54,320 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97061.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:35:22,676 INFO [train.py:898] (0/4) Epoch 27, batch 2600, loss[loss=0.1457, simple_loss=0.2357, pruned_loss=0.02781, over 18273.00 frames. ], tot_loss[loss=0.1578, simple_loss=0.2482, pruned_loss=0.03364, over 3589295.79 frames. ], batch size: 49, lr: 4.14e-03, grad_scale: 8.0 +2023-03-09 21:35:32,736 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=97094.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:35:47,782 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.742e+02 2.535e+02 2.936e+02 3.647e+02 8.103e+02, threshold=5.871e+02, percent-clipped=2.0 +2023-03-09 21:36:16,473 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97132.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:36:19,912 INFO [train.py:898] (0/4) Epoch 27, batch 2650, loss[loss=0.1352, simple_loss=0.2181, pruned_loss=0.02617, over 17689.00 frames. ], tot_loss[loss=0.1565, simple_loss=0.2467, pruned_loss=0.03309, over 3592420.72 frames. ], batch size: 39, lr: 4.14e-03, grad_scale: 8.0 +2023-03-09 21:36:57,443 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.77 vs. limit=2.0 +2023-03-09 21:37:18,539 INFO [train.py:898] (0/4) Epoch 27, batch 2700, loss[loss=0.1412, simple_loss=0.2243, pruned_loss=0.02909, over 18433.00 frames. ], tot_loss[loss=0.1567, simple_loss=0.247, pruned_loss=0.03318, over 3592330.74 frames. ], batch size: 43, lr: 4.14e-03, grad_scale: 8.0 +2023-03-09 21:37:37,397 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6526, 3.5422, 2.3556, 4.4543, 3.1584, 4.2601, 2.6754, 4.1013], + device='cuda:0'), covar=tensor([0.0682, 0.0824, 0.1571, 0.0476, 0.0884, 0.0348, 0.1191, 0.0409], + device='cuda:0'), in_proj_covar=tensor([0.0226, 0.0234, 0.0198, 0.0299, 0.0198, 0.0275, 0.0208, 0.0210], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 21:37:44,984 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.852e+02 2.663e+02 3.178e+02 4.002e+02 6.925e+02, threshold=6.355e+02, percent-clipped=3.0 +2023-03-09 21:38:17,172 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-09 21:38:17,482 INFO [train.py:898] (0/4) Epoch 27, batch 2750, loss[loss=0.138, simple_loss=0.2264, pruned_loss=0.02477, over 18484.00 frames. ], tot_loss[loss=0.157, simple_loss=0.2477, pruned_loss=0.03315, over 3575376.68 frames. ], batch size: 47, lr: 4.14e-03, grad_scale: 8.0 +2023-03-09 21:38:47,026 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97260.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:39:00,150 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.95 vs. limit=2.0 +2023-03-09 21:39:16,421 INFO [train.py:898] (0/4) Epoch 27, batch 2800, loss[loss=0.1823, simple_loss=0.2702, pruned_loss=0.04724, over 18111.00 frames. ], tot_loss[loss=0.1576, simple_loss=0.2483, pruned_loss=0.03343, over 3570388.63 frames. ], batch size: 62, lr: 4.14e-03, grad_scale: 8.0 +2023-03-09 21:39:21,942 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5698, 6.0799, 5.5824, 5.9173, 5.7518, 5.5704, 6.1828, 6.1489], + device='cuda:0'), covar=tensor([0.1329, 0.0939, 0.0490, 0.0723, 0.1481, 0.0700, 0.0649, 0.0740], + device='cuda:0'), in_proj_covar=tensor([0.0644, 0.0568, 0.0409, 0.0588, 0.0789, 0.0584, 0.0802, 0.0613], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 21:39:44,376 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.673e+02 2.484e+02 2.831e+02 3.589e+02 1.148e+03, threshold=5.662e+02, percent-clipped=3.0 +2023-03-09 21:40:00,195 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97321.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:40:15,698 INFO [train.py:898] (0/4) Epoch 27, batch 2850, loss[loss=0.1778, simple_loss=0.2721, pruned_loss=0.0418, over 17757.00 frames. ], tot_loss[loss=0.1577, simple_loss=0.2484, pruned_loss=0.03348, over 3560606.83 frames. ], batch size: 70, lr: 4.14e-03, grad_scale: 8.0 +2023-03-09 21:40:24,553 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97342.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:40:30,782 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97347.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:40:41,764 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97356.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:41:02,964 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97374.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:41:11,293 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8111, 4.0061, 2.3263, 3.9823, 5.1611, 2.7339, 3.7803, 3.9377], + device='cuda:0'), covar=tensor([0.0226, 0.1302, 0.1792, 0.0703, 0.0112, 0.1228, 0.0713, 0.0744], + device='cuda:0'), in_proj_covar=tensor([0.0183, 0.0283, 0.0211, 0.0201, 0.0142, 0.0187, 0.0223, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 21:41:15,307 INFO [train.py:898] (0/4) Epoch 27, batch 2900, loss[loss=0.1528, simple_loss=0.237, pruned_loss=0.03431, over 18478.00 frames. ], tot_loss[loss=0.1576, simple_loss=0.2486, pruned_loss=0.03334, over 3575096.06 frames. ], batch size: 44, lr: 4.14e-03, grad_scale: 8.0 +2023-03-09 21:41:38,510 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97403.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 21:41:43,554 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.636e+02 2.502e+02 2.896e+02 3.589e+02 7.058e+02, threshold=5.793e+02, percent-clipped=2.0 +2023-03-09 21:41:43,880 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97408.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:41:58,068 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.89 vs. limit=5.0 +2023-03-09 21:42:11,812 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97432.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:42:15,016 INFO [train.py:898] (0/4) Epoch 27, batch 2950, loss[loss=0.1554, simple_loss=0.2357, pruned_loss=0.03759, over 18406.00 frames. ], tot_loss[loss=0.157, simple_loss=0.2477, pruned_loss=0.03318, over 3570550.32 frames. ], batch size: 48, lr: 4.13e-03, grad_scale: 4.0 +2023-03-09 21:42:15,388 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97435.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:42:57,605 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2164, 4.1403, 4.0101, 4.1358, 4.1682, 3.7016, 4.1337, 3.9712], + device='cuda:0'), covar=tensor([0.0473, 0.0710, 0.1115, 0.0725, 0.0615, 0.0498, 0.0510, 0.1079], + device='cuda:0'), in_proj_covar=tensor([0.0516, 0.0591, 0.0728, 0.0457, 0.0477, 0.0537, 0.0572, 0.0708], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0005, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 21:43:08,292 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=97480.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:43:13,710 INFO [train.py:898] (0/4) Epoch 27, batch 3000, loss[loss=0.1733, simple_loss=0.2616, pruned_loss=0.04247, over 18502.00 frames. ], tot_loss[loss=0.1558, simple_loss=0.2464, pruned_loss=0.03259, over 3582023.27 frames. ], batch size: 53, lr: 4.13e-03, grad_scale: 4.0 +2023-03-09 21:43:13,713 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 21:43:26,377 INFO [train.py:932] (0/4) Epoch 27, validation: loss=0.1498, simple_loss=0.2479, pruned_loss=0.02584, over 944034.00 frames. +2023-03-09 21:43:26,378 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 21:43:54,643 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.734e+02 2.624e+02 3.078e+02 3.801e+02 7.902e+02, threshold=6.156e+02, percent-clipped=2.0 +2023-03-09 21:44:25,245 INFO [train.py:898] (0/4) Epoch 27, batch 3050, loss[loss=0.1683, simple_loss=0.2654, pruned_loss=0.03563, over 18477.00 frames. ], tot_loss[loss=0.1557, simple_loss=0.2464, pruned_loss=0.03251, over 3581220.02 frames. ], batch size: 53, lr: 4.13e-03, grad_scale: 4.0 +2023-03-09 21:44:49,573 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97555.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:45:24,077 INFO [train.py:898] (0/4) Epoch 27, batch 3100, loss[loss=0.1388, simple_loss=0.2155, pruned_loss=0.03107, over 18442.00 frames. ], tot_loss[loss=0.1563, simple_loss=0.247, pruned_loss=0.03281, over 3595876.48 frames. ], batch size: 43, lr: 4.13e-03, grad_scale: 4.0 +2023-03-09 21:45:52,560 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.712e+02 2.595e+02 3.045e+02 3.637e+02 1.953e+03, threshold=6.090e+02, percent-clipped=1.0 +2023-03-09 21:46:00,678 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97616.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:46:00,804 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97616.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:46:22,624 INFO [train.py:898] (0/4) Epoch 27, batch 3150, loss[loss=0.1623, simple_loss=0.2629, pruned_loss=0.03086, over 16954.00 frames. ], tot_loss[loss=0.1561, simple_loss=0.2469, pruned_loss=0.0327, over 3592836.85 frames. ], batch size: 78, lr: 4.13e-03, grad_scale: 4.0 +2023-03-09 21:46:47,676 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97656.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:47:09,202 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9023, 4.5383, 4.5509, 3.4776, 3.8015, 3.5766, 2.6677, 2.5426], + device='cuda:0'), covar=tensor([0.0205, 0.0127, 0.0090, 0.0304, 0.0339, 0.0226, 0.0708, 0.0850], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0063, 0.0069, 0.0071, 0.0094, 0.0071, 0.0079, 0.0087], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0007, 0.0005, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 21:47:09,226 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97674.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 21:47:21,020 INFO [train.py:898] (0/4) Epoch 27, batch 3200, loss[loss=0.1803, simple_loss=0.2721, pruned_loss=0.04424, over 18355.00 frames. ], tot_loss[loss=0.1568, simple_loss=0.2476, pruned_loss=0.03298, over 3595397.56 frames. ], batch size: 55, lr: 4.13e-03, grad_scale: 8.0 +2023-03-09 21:47:36,650 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97698.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 21:47:43,539 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97703.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:47:44,759 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=97704.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:47:50,085 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.671e+02 2.606e+02 3.062e+02 3.639e+02 1.226e+03, threshold=6.124e+02, percent-clipped=5.0 +2023-03-09 21:47:50,760 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-09 21:48:14,880 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97730.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:48:16,161 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8801, 4.6041, 4.6589, 3.5925, 3.8221, 3.6319, 2.7927, 2.7873], + device='cuda:0'), covar=tensor([0.0230, 0.0152, 0.0081, 0.0267, 0.0317, 0.0226, 0.0672, 0.0731], + device='cuda:0'), in_proj_covar=tensor([0.0075, 0.0063, 0.0068, 0.0071, 0.0094, 0.0071, 0.0079, 0.0086], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0007, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 21:48:20,235 INFO [train.py:898] (0/4) Epoch 27, batch 3250, loss[loss=0.1285, simple_loss=0.2189, pruned_loss=0.01909, over 18276.00 frames. ], tot_loss[loss=0.1563, simple_loss=0.2471, pruned_loss=0.03277, over 3584893.85 frames. ], batch size: 47, lr: 4.13e-03, grad_scale: 8.0 +2023-03-09 21:48:20,602 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97735.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 21:48:43,498 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6084, 2.7459, 4.2058, 3.5061, 2.5736, 4.4361, 3.8597, 2.7554], + device='cuda:0'), covar=tensor([0.0532, 0.1626, 0.0366, 0.0543, 0.1630, 0.0271, 0.0609, 0.1044], + device='cuda:0'), in_proj_covar=tensor([0.0218, 0.0241, 0.0230, 0.0171, 0.0225, 0.0218, 0.0255, 0.0196], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 21:49:19,395 INFO [train.py:898] (0/4) Epoch 27, batch 3300, loss[loss=0.1366, simple_loss=0.2195, pruned_loss=0.02683, over 18463.00 frames. ], tot_loss[loss=0.1561, simple_loss=0.2469, pruned_loss=0.03263, over 3593231.54 frames. ], batch size: 43, lr: 4.13e-03, grad_scale: 8.0 +2023-03-09 21:49:48,580 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.710e+02 2.346e+02 2.851e+02 3.327e+02 7.225e+02, threshold=5.702e+02, percent-clipped=2.0 +2023-03-09 21:50:09,471 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6559, 3.4766, 4.6638, 4.1322, 3.1723, 2.8556, 4.1871, 4.9615], + device='cuda:0'), covar=tensor([0.0872, 0.1482, 0.0294, 0.0446, 0.1012, 0.1321, 0.0444, 0.0209], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0285, 0.0173, 0.0188, 0.0198, 0.0199, 0.0201, 0.0213], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 21:50:18,959 INFO [train.py:898] (0/4) Epoch 27, batch 3350, loss[loss=0.1545, simple_loss=0.248, pruned_loss=0.03051, over 18502.00 frames. ], tot_loss[loss=0.1559, simple_loss=0.2468, pruned_loss=0.0325, over 3591707.95 frames. ], batch size: 53, lr: 4.13e-03, grad_scale: 8.0 +2023-03-09 21:50:21,481 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6576, 3.6103, 3.4759, 3.0745, 3.3721, 2.7402, 2.7885, 3.6639], + device='cuda:0'), covar=tensor([0.0084, 0.0109, 0.0105, 0.0174, 0.0118, 0.0224, 0.0221, 0.0081], + device='cuda:0'), in_proj_covar=tensor([0.0158, 0.0178, 0.0148, 0.0198, 0.0156, 0.0190, 0.0193, 0.0135], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 21:50:31,940 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-09 21:51:17,701 INFO [train.py:898] (0/4) Epoch 27, batch 3400, loss[loss=0.1761, simple_loss=0.2683, pruned_loss=0.04199, over 18307.00 frames. ], tot_loss[loss=0.1558, simple_loss=0.2465, pruned_loss=0.03256, over 3580481.70 frames. ], batch size: 54, lr: 4.12e-03, grad_scale: 8.0 +2023-03-09 21:51:21,342 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=97888.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:51:45,800 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.740e+02 2.527e+02 3.006e+02 3.880e+02 7.423e+02, threshold=6.013e+02, percent-clipped=1.0 +2023-03-09 21:51:48,321 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=97911.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:51:54,466 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97916.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:52:16,400 INFO [train.py:898] (0/4) Epoch 27, batch 3450, loss[loss=0.138, simple_loss=0.2244, pruned_loss=0.02574, over 18544.00 frames. ], tot_loss[loss=0.1554, simple_loss=0.2457, pruned_loss=0.03252, over 3582618.63 frames. ], batch size: 49, lr: 4.12e-03, grad_scale: 8.0 +2023-03-09 21:52:33,391 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=97949.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:52:51,313 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=97964.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:53:03,403 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2489, 3.6923, 2.4470, 3.5234, 4.5495, 2.3513, 3.5031, 3.7483], + device='cuda:0'), covar=tensor([0.0259, 0.1058, 0.1677, 0.0748, 0.0131, 0.1486, 0.0806, 0.0641], + device='cuda:0'), in_proj_covar=tensor([0.0184, 0.0282, 0.0211, 0.0202, 0.0143, 0.0188, 0.0224, 0.0233], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 21:53:15,532 INFO [train.py:898] (0/4) Epoch 27, batch 3500, loss[loss=0.172, simple_loss=0.2646, pruned_loss=0.03968, over 18476.00 frames. ], tot_loss[loss=0.1558, simple_loss=0.2462, pruned_loss=0.03273, over 3573139.24 frames. ], batch size: 59, lr: 4.12e-03, grad_scale: 8.0 +2023-03-09 21:53:31,182 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=97998.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:53:33,505 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-98000.pt +2023-03-09 21:53:41,429 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98003.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:53:48,595 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.729e+02 2.463e+02 3.009e+02 3.694e+02 8.089e+02, threshold=6.019e+02, percent-clipped=2.0 +2023-03-09 21:53:52,283 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1186, 3.4100, 3.3377, 2.9060, 2.9876, 2.9009, 2.5759, 2.3995], + device='cuda:0'), covar=tensor([0.0270, 0.0188, 0.0160, 0.0299, 0.0365, 0.0239, 0.0571, 0.0715], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0063, 0.0068, 0.0071, 0.0093, 0.0070, 0.0078, 0.0086], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 21:54:12,089 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98030.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 21:54:12,116 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98030.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:54:17,273 INFO [train.py:898] (0/4) Epoch 27, batch 3550, loss[loss=0.1608, simple_loss=0.2494, pruned_loss=0.03616, over 18366.00 frames. ], tot_loss[loss=0.1559, simple_loss=0.2465, pruned_loss=0.03265, over 3578384.32 frames. ], batch size: 56, lr: 4.12e-03, grad_scale: 8.0 +2023-03-09 21:54:25,796 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8285, 4.1460, 2.4371, 3.9567, 5.2600, 2.6417, 3.7559, 3.9368], + device='cuda:0'), covar=tensor([0.0224, 0.1132, 0.1681, 0.0691, 0.0088, 0.1235, 0.0745, 0.0774], + device='cuda:0'), in_proj_covar=tensor([0.0184, 0.0280, 0.0210, 0.0201, 0.0142, 0.0187, 0.0223, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 21:54:29,872 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=98046.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:54:35,247 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=98051.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:54:40,857 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7715, 4.7398, 2.8281, 4.5737, 4.5031, 4.7123, 4.5423, 2.5591], + device='cuda:0'), covar=tensor([0.0240, 0.0066, 0.0697, 0.0097, 0.0083, 0.0077, 0.0095, 0.1045], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0084, 0.0099, 0.0100, 0.0091, 0.0081, 0.0088, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 21:55:02,927 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9509, 4.5218, 4.6073, 3.5030, 3.7354, 3.6439, 2.8357, 2.8890], + device='cuda:0'), covar=tensor([0.0200, 0.0168, 0.0081, 0.0292, 0.0328, 0.0225, 0.0666, 0.0698], + device='cuda:0'), in_proj_covar=tensor([0.0074, 0.0063, 0.0068, 0.0071, 0.0093, 0.0070, 0.0078, 0.0086], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0006, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 21:55:04,916 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=98078.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:55:12,367 INFO [train.py:898] (0/4) Epoch 27, batch 3600, loss[loss=0.1525, simple_loss=0.2461, pruned_loss=0.02949, over 18493.00 frames. ], tot_loss[loss=0.1557, simple_loss=0.2466, pruned_loss=0.03242, over 3575656.70 frames. ], batch size: 51, lr: 4.12e-03, grad_scale: 8.0 +2023-03-09 21:55:35,502 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98106.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:55:38,146 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.847e+02 2.404e+02 2.873e+02 3.414e+02 5.036e+02, threshold=5.745e+02, percent-clipped=0.0 +2023-03-09 21:55:48,207 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-27.pt +2023-03-09 21:56:16,270 INFO [train.py:898] (0/4) Epoch 28, batch 0, loss[loss=0.1618, simple_loss=0.2506, pruned_loss=0.03644, over 18411.00 frames. ], tot_loss[loss=0.1618, simple_loss=0.2506, pruned_loss=0.03644, over 18411.00 frames. ], batch size: 48, lr: 4.04e-03, grad_scale: 8.0 +2023-03-09 21:56:16,272 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 21:56:28,181 INFO [train.py:932] (0/4) Epoch 28, validation: loss=0.1499, simple_loss=0.2483, pruned_loss=0.02581, over 944034.00 frames. +2023-03-09 21:56:28,182 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 21:57:24,366 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98167.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:57:26,203 INFO [train.py:898] (0/4) Epoch 28, batch 50, loss[loss=0.153, simple_loss=0.2474, pruned_loss=0.02924, over 18312.00 frames. ], tot_loss[loss=0.1601, simple_loss=0.2527, pruned_loss=0.03373, over 803962.20 frames. ], batch size: 54, lr: 4.04e-03, grad_scale: 8.0 +2023-03-09 21:58:15,012 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.839e+02 2.500e+02 2.819e+02 3.475e+02 6.212e+02, threshold=5.638e+02, percent-clipped=1.0 +2023-03-09 21:58:17,518 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98211.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:58:26,486 INFO [train.py:898] (0/4) Epoch 28, batch 100, loss[loss=0.1678, simple_loss=0.2627, pruned_loss=0.03647, over 16335.00 frames. ], tot_loss[loss=0.1596, simple_loss=0.2516, pruned_loss=0.03384, over 1431367.73 frames. ], batch size: 95, lr: 4.04e-03, grad_scale: 8.0 +2023-03-09 21:58:34,793 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4413, 5.9643, 5.5849, 5.6933, 5.5282, 5.3174, 6.0165, 5.9818], + device='cuda:0'), covar=tensor([0.1180, 0.0780, 0.0459, 0.0742, 0.1435, 0.0773, 0.0549, 0.0707], + device='cuda:0'), in_proj_covar=tensor([0.0635, 0.0559, 0.0399, 0.0581, 0.0778, 0.0575, 0.0788, 0.0607], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 21:58:41,245 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7832, 2.4763, 2.7796, 2.8094, 3.3905, 5.0440, 4.9289, 3.3316], + device='cuda:0'), covar=tensor([0.2146, 0.2686, 0.3256, 0.2070, 0.2466, 0.0351, 0.0367, 0.1134], + device='cuda:0'), in_proj_covar=tensor([0.0329, 0.0364, 0.0410, 0.0292, 0.0397, 0.0268, 0.0302, 0.0272], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 21:58:53,000 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7662, 4.0404, 2.3032, 3.9526, 5.1463, 2.6963, 3.6417, 3.8570], + device='cuda:0'), covar=tensor([0.0259, 0.1329, 0.1861, 0.0732, 0.0109, 0.1266, 0.0845, 0.0830], + device='cuda:0'), in_proj_covar=tensor([0.0185, 0.0282, 0.0210, 0.0202, 0.0143, 0.0187, 0.0223, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 21:58:55,978 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98244.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:59:14,191 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=98259.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 21:59:25,572 INFO [train.py:898] (0/4) Epoch 28, batch 150, loss[loss=0.1556, simple_loss=0.2449, pruned_loss=0.03312, over 18475.00 frames. ], tot_loss[loss=0.1577, simple_loss=0.2491, pruned_loss=0.03314, over 1904584.89 frames. ], batch size: 51, lr: 4.04e-03, grad_scale: 8.0 +2023-03-09 21:59:37,580 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5745, 3.5113, 4.7521, 4.0722, 3.1653, 2.8282, 4.1394, 4.9977], + device='cuda:0'), covar=tensor([0.0848, 0.1333, 0.0226, 0.0487, 0.1017, 0.1344, 0.0463, 0.0273], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0284, 0.0172, 0.0188, 0.0198, 0.0198, 0.0201, 0.0211], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 22:00:11,994 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7709, 4.7310, 4.4403, 4.6694, 4.7016, 4.1276, 4.6071, 4.4301], + device='cuda:0'), covar=tensor([0.0461, 0.0505, 0.1283, 0.0811, 0.0555, 0.0494, 0.0520, 0.1046], + device='cuda:0'), in_proj_covar=tensor([0.0516, 0.0589, 0.0728, 0.0455, 0.0474, 0.0538, 0.0572, 0.0701], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0005, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 22:00:13,405 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6328, 2.5029, 2.5817, 2.6898, 3.1898, 4.9584, 4.9816, 3.4734], + device='cuda:0'), covar=tensor([0.2237, 0.2754, 0.3381, 0.2158, 0.2787, 0.0303, 0.0322, 0.1045], + device='cuda:0'), in_proj_covar=tensor([0.0331, 0.0365, 0.0411, 0.0293, 0.0398, 0.0268, 0.0302, 0.0273], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 22:00:14,047 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.664e+02 2.460e+02 2.957e+02 3.466e+02 5.415e+02, threshold=5.915e+02, percent-clipped=0.0 +2023-03-09 22:00:25,432 INFO [train.py:898] (0/4) Epoch 28, batch 200, loss[loss=0.1615, simple_loss=0.2562, pruned_loss=0.03342, over 18351.00 frames. ], tot_loss[loss=0.1563, simple_loss=0.2477, pruned_loss=0.03244, over 2283227.29 frames. ], batch size: 56, lr: 4.04e-03, grad_scale: 8.0 +2023-03-09 22:00:38,070 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98330.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 22:00:52,420 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6497, 3.6245, 4.9840, 4.2829, 3.4111, 3.0078, 4.3644, 5.2254], + device='cuda:0'), covar=tensor([0.0834, 0.1394, 0.0219, 0.0452, 0.0967, 0.1274, 0.0414, 0.0191], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0284, 0.0172, 0.0187, 0.0197, 0.0198, 0.0201, 0.0211], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 22:01:24,246 INFO [train.py:898] (0/4) Epoch 28, batch 250, loss[loss=0.1417, simple_loss=0.2309, pruned_loss=0.02628, over 18375.00 frames. ], tot_loss[loss=0.1557, simple_loss=0.2472, pruned_loss=0.0321, over 2573498.08 frames. ], batch size: 46, lr: 4.04e-03, grad_scale: 8.0 +2023-03-09 22:01:30,444 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8970, 4.1299, 2.3476, 4.0582, 5.2366, 2.7482, 3.9100, 4.0599], + device='cuda:0'), covar=tensor([0.0225, 0.1249, 0.1748, 0.0637, 0.0100, 0.1190, 0.0650, 0.0716], + device='cuda:0'), in_proj_covar=tensor([0.0183, 0.0280, 0.0209, 0.0201, 0.0143, 0.0186, 0.0222, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 22:01:34,703 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=98378.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 22:01:43,818 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8374, 3.2854, 4.5939, 3.7837, 2.8778, 4.8405, 4.0122, 3.1974], + device='cuda:0'), covar=tensor([0.0521, 0.1301, 0.0251, 0.0488, 0.1542, 0.0206, 0.0610, 0.0878], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0246, 0.0233, 0.0174, 0.0229, 0.0221, 0.0260, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 22:02:11,178 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.633e+02 2.515e+02 2.972e+02 3.446e+02 6.764e+02, threshold=5.944e+02, percent-clipped=3.0 +2023-03-09 22:02:12,593 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98410.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:02:23,113 INFO [train.py:898] (0/4) Epoch 28, batch 300, loss[loss=0.2017, simple_loss=0.2817, pruned_loss=0.06089, over 12338.00 frames. ], tot_loss[loss=0.1555, simple_loss=0.247, pruned_loss=0.03198, over 2805170.68 frames. ], batch size: 131, lr: 4.04e-03, grad_scale: 8.0 +2023-03-09 22:02:32,686 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0138, 5.0149, 5.0957, 4.7795, 4.8752, 4.8404, 5.1603, 5.2193], + device='cuda:0'), covar=tensor([0.0074, 0.0076, 0.0065, 0.0134, 0.0062, 0.0174, 0.0090, 0.0099], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0075, 0.0080, 0.0100, 0.0080, 0.0109, 0.0092, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 22:02:56,609 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5492, 2.3785, 2.5168, 2.5982, 3.0997, 4.3630, 4.3076, 3.1041], + device='cuda:0'), covar=tensor([0.2191, 0.2690, 0.3265, 0.2218, 0.2554, 0.0373, 0.0453, 0.1165], + device='cuda:0'), in_proj_covar=tensor([0.0331, 0.0365, 0.0412, 0.0293, 0.0398, 0.0268, 0.0302, 0.0273], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 22:03:14,381 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98462.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:03:22,727 INFO [train.py:898] (0/4) Epoch 28, batch 350, loss[loss=0.1468, simple_loss=0.2294, pruned_loss=0.03215, over 18163.00 frames. ], tot_loss[loss=0.1555, simple_loss=0.2467, pruned_loss=0.03211, over 2984662.58 frames. ], batch size: 44, lr: 4.04e-03, grad_scale: 8.0 +2023-03-09 22:03:25,385 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98471.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:04:09,223 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.910e+02 2.547e+02 2.961e+02 3.445e+02 6.973e+02, threshold=5.923e+02, percent-clipped=1.0 +2023-03-09 22:04:21,295 INFO [train.py:898] (0/4) Epoch 28, batch 400, loss[loss=0.1407, simple_loss=0.2346, pruned_loss=0.02344, over 18384.00 frames. ], tot_loss[loss=0.1557, simple_loss=0.247, pruned_loss=0.03223, over 3126961.28 frames. ], batch size: 50, lr: 4.04e-03, grad_scale: 8.0 +2023-03-09 22:04:37,393 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-09 22:04:43,815 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5703, 2.8994, 2.4917, 2.8219, 3.6269, 3.4804, 3.1001, 2.8233], + device='cuda:0'), covar=tensor([0.0162, 0.0292, 0.0587, 0.0409, 0.0200, 0.0175, 0.0428, 0.0422], + device='cuda:0'), in_proj_covar=tensor([0.0147, 0.0147, 0.0168, 0.0168, 0.0144, 0.0130, 0.0163, 0.0166], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 22:04:49,516 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98543.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:04:50,561 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98544.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:05:20,088 INFO [train.py:898] (0/4) Epoch 28, batch 450, loss[loss=0.1357, simple_loss=0.2231, pruned_loss=0.02418, over 18374.00 frames. ], tot_loss[loss=0.1557, simple_loss=0.2469, pruned_loss=0.03227, over 3230268.39 frames. ], batch size: 46, lr: 4.04e-03, grad_scale: 8.0 +2023-03-09 22:05:46,790 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=98592.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:06:01,919 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98604.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:06:07,231 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.606e+02 2.466e+02 2.902e+02 3.493e+02 6.837e+02, threshold=5.804e+02, percent-clipped=4.0 +2023-03-09 22:06:13,855 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5425, 2.4166, 4.2238, 3.6803, 2.2072, 4.3637, 3.6409, 2.6579], + device='cuda:0'), covar=tensor([0.0513, 0.2231, 0.0313, 0.0448, 0.2183, 0.0323, 0.0680, 0.1284], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0246, 0.0233, 0.0174, 0.0228, 0.0221, 0.0260, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 22:06:19,753 INFO [train.py:898] (0/4) Epoch 28, batch 500, loss[loss=0.1467, simple_loss=0.247, pruned_loss=0.02322, over 18359.00 frames. ], tot_loss[loss=0.1562, simple_loss=0.2473, pruned_loss=0.03256, over 3315795.59 frames. ], batch size: 55, lr: 4.03e-03, grad_scale: 8.0 +2023-03-09 22:06:24,703 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5071, 2.8427, 3.9210, 3.3202, 2.5786, 4.1550, 3.6506, 2.7846], + device='cuda:0'), covar=tensor([0.0519, 0.1487, 0.0372, 0.0587, 0.1569, 0.0277, 0.0671, 0.0974], + device='cuda:0'), in_proj_covar=tensor([0.0222, 0.0245, 0.0233, 0.0174, 0.0228, 0.0221, 0.0260, 0.0198], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 22:07:14,966 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.07 vs. limit=5.0 +2023-03-09 22:07:18,533 INFO [train.py:898] (0/4) Epoch 28, batch 550, loss[loss=0.159, simple_loss=0.253, pruned_loss=0.03246, over 18148.00 frames. ], tot_loss[loss=0.1563, simple_loss=0.2473, pruned_loss=0.03265, over 3372067.38 frames. ], batch size: 62, lr: 4.03e-03, grad_scale: 8.0 +2023-03-09 22:07:38,570 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.70 vs. limit=2.0 +2023-03-09 22:08:05,609 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.504e+02 2.472e+02 2.832e+02 3.276e+02 6.900e+02, threshold=5.664e+02, percent-clipped=1.0 +2023-03-09 22:08:17,664 INFO [train.py:898] (0/4) Epoch 28, batch 600, loss[loss=0.1592, simple_loss=0.2524, pruned_loss=0.033, over 17334.00 frames. ], tot_loss[loss=0.1557, simple_loss=0.2468, pruned_loss=0.03233, over 3427266.81 frames. ], batch size: 78, lr: 4.03e-03, grad_scale: 8.0 +2023-03-09 22:08:38,848 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.3766, 3.3678, 2.0960, 4.1887, 2.8477, 3.9157, 2.3620, 3.6321], + device='cuda:0'), covar=tensor([0.0668, 0.0865, 0.1588, 0.0537, 0.0910, 0.0309, 0.1294, 0.0497], + device='cuda:0'), in_proj_covar=tensor([0.0225, 0.0235, 0.0196, 0.0299, 0.0198, 0.0274, 0.0209, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 22:09:08,712 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=98762.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:09:13,002 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98766.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:09:16,298 INFO [train.py:898] (0/4) Epoch 28, batch 650, loss[loss=0.1747, simple_loss=0.264, pruned_loss=0.0427, over 18374.00 frames. ], tot_loss[loss=0.1565, simple_loss=0.2477, pruned_loss=0.03265, over 3456404.26 frames. ], batch size: 56, lr: 4.03e-03, grad_scale: 8.0 +2023-03-09 22:09:42,723 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-09 22:10:03,543 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.904e+02 2.586e+02 3.038e+02 3.660e+02 8.440e+02, threshold=6.077e+02, percent-clipped=4.0 +2023-03-09 22:10:04,948 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=98810.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:10:15,452 INFO [train.py:898] (0/4) Epoch 28, batch 700, loss[loss=0.161, simple_loss=0.246, pruned_loss=0.03801, over 12399.00 frames. ], tot_loss[loss=0.156, simple_loss=0.2471, pruned_loss=0.03251, over 3486328.01 frames. ], batch size: 130, lr: 4.03e-03, grad_scale: 8.0 +2023-03-09 22:10:44,945 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2720, 5.8328, 5.4509, 5.6323, 5.4377, 5.2957, 5.8791, 5.8630], + device='cuda:0'), covar=tensor([0.1189, 0.0774, 0.0623, 0.0624, 0.1297, 0.0668, 0.0565, 0.0673], + device='cuda:0'), in_proj_covar=tensor([0.0643, 0.0566, 0.0406, 0.0586, 0.0789, 0.0583, 0.0803, 0.0613], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 22:11:14,541 INFO [train.py:898] (0/4) Epoch 28, batch 750, loss[loss=0.1601, simple_loss=0.2468, pruned_loss=0.03674, over 17972.00 frames. ], tot_loss[loss=0.1567, simple_loss=0.2476, pruned_loss=0.03286, over 3500585.40 frames. ], batch size: 65, lr: 4.03e-03, grad_scale: 8.0 +2023-03-09 22:11:39,942 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98890.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:11:46,344 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98895.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:11:50,670 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=98899.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:12:02,003 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.555e+02 2.589e+02 3.008e+02 3.946e+02 6.579e+02, threshold=6.017e+02, percent-clipped=1.0 +2023-03-09 22:12:13,966 INFO [train.py:898] (0/4) Epoch 28, batch 800, loss[loss=0.1767, simple_loss=0.2658, pruned_loss=0.04379, over 17981.00 frames. ], tot_loss[loss=0.1565, simple_loss=0.2476, pruned_loss=0.03273, over 3518554.11 frames. ], batch size: 65, lr: 4.03e-03, grad_scale: 8.0 +2023-03-09 22:12:52,248 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98951.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:12:58,115 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=98956.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:12:59,642 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-03-09 22:13:12,395 INFO [train.py:898] (0/4) Epoch 28, batch 850, loss[loss=0.1361, simple_loss=0.2264, pruned_loss=0.02292, over 18243.00 frames. ], tot_loss[loss=0.1563, simple_loss=0.2474, pruned_loss=0.03257, over 3540569.07 frames. ], batch size: 45, lr: 4.03e-03, grad_scale: 8.0 +2023-03-09 22:13:15,076 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=98971.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:14:00,115 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.700e+02 2.503e+02 2.990e+02 3.802e+02 7.789e+02, threshold=5.979e+02, percent-clipped=2.0 +2023-03-09 22:14:11,500 INFO [train.py:898] (0/4) Epoch 28, batch 900, loss[loss=0.1734, simple_loss=0.2659, pruned_loss=0.0405, over 18078.00 frames. ], tot_loss[loss=0.1561, simple_loss=0.2472, pruned_loss=0.0325, over 3548955.49 frames. ], batch size: 62, lr: 4.03e-03, grad_scale: 8.0 +2023-03-09 22:14:14,841 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.89 vs. limit=2.0 +2023-03-09 22:14:27,836 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99032.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:14:49,953 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99051.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:15:07,197 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99066.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:15:10,378 INFO [train.py:898] (0/4) Epoch 28, batch 950, loss[loss=0.1525, simple_loss=0.249, pruned_loss=0.02797, over 18391.00 frames. ], tot_loss[loss=0.1563, simple_loss=0.2475, pruned_loss=0.03248, over 3555914.93 frames. ], batch size: 50, lr: 4.02e-03, grad_scale: 8.0 +2023-03-09 22:15:39,238 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.2075, 3.6677, 2.4134, 3.5461, 4.4411, 2.4806, 3.4134, 3.6062], + device='cuda:0'), covar=tensor([0.0302, 0.1208, 0.1832, 0.0786, 0.0180, 0.1419, 0.0854, 0.0746], + device='cuda:0'), in_proj_covar=tensor([0.0184, 0.0280, 0.0209, 0.0201, 0.0143, 0.0185, 0.0222, 0.0231], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 22:15:57,908 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.971e+02 2.497e+02 3.033e+02 3.631e+02 7.579e+02, threshold=6.066e+02, percent-clipped=1.0 +2023-03-09 22:16:01,734 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99112.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:16:03,847 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=99114.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:16:09,567 INFO [train.py:898] (0/4) Epoch 28, batch 1000, loss[loss=0.1379, simple_loss=0.22, pruned_loss=0.02786, over 18152.00 frames. ], tot_loss[loss=0.1568, simple_loss=0.2482, pruned_loss=0.03276, over 3569902.17 frames. ], batch size: 44, lr: 4.02e-03, grad_scale: 8.0 +2023-03-09 22:17:08,462 INFO [train.py:898] (0/4) Epoch 28, batch 1050, loss[loss=0.1343, simple_loss=0.2186, pruned_loss=0.02497, over 18431.00 frames. ], tot_loss[loss=0.1573, simple_loss=0.2484, pruned_loss=0.0331, over 3568823.32 frames. ], batch size: 43, lr: 4.02e-03, grad_scale: 8.0 +2023-03-09 22:17:43,528 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99199.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:17:55,237 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.556e+02 2.410e+02 2.895e+02 3.644e+02 7.761e+02, threshold=5.789e+02, percent-clipped=3.0 +2023-03-09 22:18:06,880 INFO [train.py:898] (0/4) Epoch 28, batch 1100, loss[loss=0.1863, simple_loss=0.2747, pruned_loss=0.04894, over 18301.00 frames. ], tot_loss[loss=0.1576, simple_loss=0.2486, pruned_loss=0.03329, over 3581591.59 frames. ], batch size: 57, lr: 4.02e-03, grad_scale: 8.0 +2023-03-09 22:18:18,465 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99229.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:18:39,004 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99246.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:18:40,203 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=99247.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:18:44,753 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99251.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:19:05,439 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99268.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:19:06,231 INFO [train.py:898] (0/4) Epoch 28, batch 1150, loss[loss=0.1406, simple_loss=0.231, pruned_loss=0.02505, over 18511.00 frames. ], tot_loss[loss=0.157, simple_loss=0.2478, pruned_loss=0.03309, over 3582539.38 frames. ], batch size: 47, lr: 4.02e-03, grad_scale: 8.0 +2023-03-09 22:19:31,251 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99290.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:19:53,825 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.536e+02 2.497e+02 2.883e+02 3.542e+02 9.328e+02, threshold=5.765e+02, percent-clipped=1.0 +2023-03-09 22:20:00,929 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8017, 4.9502, 2.6347, 4.7582, 4.7271, 4.9382, 4.7044, 2.5486], + device='cuda:0'), covar=tensor([0.0263, 0.0065, 0.0794, 0.0105, 0.0074, 0.0075, 0.0098, 0.1050], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0084, 0.0098, 0.0100, 0.0091, 0.0080, 0.0087, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 22:20:05,842 INFO [train.py:898] (0/4) Epoch 28, batch 1200, loss[loss=0.1389, simple_loss=0.2203, pruned_loss=0.02873, over 18365.00 frames. ], tot_loss[loss=0.157, simple_loss=0.2478, pruned_loss=0.03311, over 3592269.55 frames. ], batch size: 42, lr: 4.02e-03, grad_scale: 8.0 +2023-03-09 22:20:08,766 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.20 vs. limit=2.0 +2023-03-09 22:20:11,985 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99324.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:20:15,128 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99327.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:20:17,753 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99329.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:20:54,483 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0705, 5.2038, 5.3463, 5.3622, 4.9578, 5.8862, 5.5084, 5.1642], + device='cuda:0'), covar=tensor([0.1179, 0.0697, 0.0859, 0.0858, 0.1383, 0.0752, 0.0647, 0.1680], + device='cuda:0'), in_proj_covar=tensor([0.0373, 0.0304, 0.0330, 0.0333, 0.0340, 0.0442, 0.0296, 0.0433], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 22:21:04,022 INFO [train.py:898] (0/4) Epoch 28, batch 1250, loss[loss=0.1465, simple_loss=0.2478, pruned_loss=0.02261, over 18490.00 frames. ], tot_loss[loss=0.1559, simple_loss=0.2469, pruned_loss=0.03251, over 3603233.22 frames. ], batch size: 53, lr: 4.02e-03, grad_scale: 8.0 +2023-03-09 22:21:23,340 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99385.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:21:43,181 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.27 vs. limit=5.0 +2023-03-09 22:21:49,417 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99407.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:21:51,422 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.740e+02 2.716e+02 3.087e+02 3.833e+02 7.403e+02, threshold=6.174e+02, percent-clipped=2.0 +2023-03-09 22:22:03,393 INFO [train.py:898] (0/4) Epoch 28, batch 1300, loss[loss=0.1443, simple_loss=0.2251, pruned_loss=0.03178, over 18498.00 frames. ], tot_loss[loss=0.1558, simple_loss=0.2467, pruned_loss=0.03241, over 3606092.20 frames. ], batch size: 44, lr: 4.02e-03, grad_scale: 16.0 +2023-03-09 22:22:04,888 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99420.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:22:12,637 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-09 22:23:00,180 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9314, 4.1928, 2.6576, 3.9834, 5.2673, 2.6605, 3.9077, 4.1481], + device='cuda:0'), covar=tensor([0.0197, 0.1158, 0.1490, 0.0724, 0.0090, 0.1221, 0.0651, 0.0673], + device='cuda:0'), in_proj_covar=tensor([0.0185, 0.0281, 0.0209, 0.0201, 0.0143, 0.0186, 0.0223, 0.0232], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0003, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 22:23:01,803 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99468.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 22:23:02,623 INFO [train.py:898] (0/4) Epoch 28, batch 1350, loss[loss=0.1719, simple_loss=0.2663, pruned_loss=0.03874, over 18481.00 frames. ], tot_loss[loss=0.1561, simple_loss=0.2468, pruned_loss=0.03268, over 3594860.78 frames. ], batch size: 59, lr: 4.02e-03, grad_scale: 16.0 +2023-03-09 22:23:03,059 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8435, 2.4330, 2.7429, 2.9044, 3.3172, 4.9430, 5.0111, 3.5641], + device='cuda:0'), covar=tensor([0.1937, 0.2615, 0.2967, 0.1894, 0.2487, 0.0264, 0.0296, 0.0990], + device='cuda:0'), in_proj_covar=tensor([0.0330, 0.0364, 0.0411, 0.0291, 0.0396, 0.0266, 0.0302, 0.0272], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 22:23:17,569 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99481.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:23:49,957 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.936e+02 2.551e+02 3.034e+02 3.736e+02 7.669e+02, threshold=6.068e+02, percent-clipped=1.0 +2023-03-09 22:24:01,498 INFO [train.py:898] (0/4) Epoch 28, batch 1400, loss[loss=0.1524, simple_loss=0.2529, pruned_loss=0.02596, over 18353.00 frames. ], tot_loss[loss=0.1559, simple_loss=0.2466, pruned_loss=0.03257, over 3595858.10 frames. ], batch size: 55, lr: 4.02e-03, grad_scale: 16.0 +2023-03-09 22:24:13,800 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99529.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 22:24:20,986 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-09 22:24:32,962 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99546.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:24:39,337 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99551.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:24:59,069 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99568.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:24:59,945 INFO [train.py:898] (0/4) Epoch 28, batch 1450, loss[loss=0.1322, simple_loss=0.2203, pruned_loss=0.02204, over 18347.00 frames. ], tot_loss[loss=0.1556, simple_loss=0.2467, pruned_loss=0.03232, over 3600971.18 frames. ], batch size: 46, lr: 4.01e-03, grad_scale: 16.0 +2023-03-09 22:25:19,568 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99585.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:25:24,191 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99589.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:25:29,571 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=99594.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:25:32,083 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8809, 4.1921, 2.4321, 4.0118, 5.2086, 2.5762, 3.8295, 4.1290], + device='cuda:0'), covar=tensor([0.0223, 0.1218, 0.1695, 0.0671, 0.0104, 0.1253, 0.0684, 0.0661], + device='cuda:0'), in_proj_covar=tensor([0.0187, 0.0284, 0.0211, 0.0203, 0.0145, 0.0188, 0.0225, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 22:25:35,187 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=99599.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:25:39,570 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1533, 5.2483, 5.4600, 5.4742, 5.0500, 5.9847, 5.6348, 5.2747], + device='cuda:0'), covar=tensor([0.1280, 0.0643, 0.0729, 0.0725, 0.1618, 0.0721, 0.0742, 0.1696], + device='cuda:0'), in_proj_covar=tensor([0.0375, 0.0307, 0.0332, 0.0335, 0.0342, 0.0446, 0.0300, 0.0437], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 22:25:48,002 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.800e+02 2.454e+02 2.982e+02 3.510e+02 7.061e+02, threshold=5.964e+02, percent-clipped=2.0 +2023-03-09 22:25:53,956 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.6023, 6.0724, 5.6253, 5.9184, 5.7358, 5.5747, 6.1742, 6.1048], + device='cuda:0'), covar=tensor([0.1080, 0.0770, 0.0462, 0.0692, 0.1356, 0.0700, 0.0531, 0.0698], + device='cuda:0'), in_proj_covar=tensor([0.0639, 0.0558, 0.0403, 0.0582, 0.0783, 0.0580, 0.0795, 0.0611], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 22:25:59,277 INFO [train.py:898] (0/4) Epoch 28, batch 1500, loss[loss=0.1465, simple_loss=0.2404, pruned_loss=0.02627, over 18491.00 frames. ], tot_loss[loss=0.1551, simple_loss=0.2462, pruned_loss=0.03196, over 3596778.40 frames. ], batch size: 47, lr: 4.01e-03, grad_scale: 16.0 +2023-03-09 22:25:59,537 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0494, 5.1448, 5.3036, 5.3765, 4.9653, 5.8490, 5.4981, 5.1199], + device='cuda:0'), covar=tensor([0.1215, 0.0690, 0.0792, 0.0997, 0.1562, 0.0799, 0.0742, 0.1845], + device='cuda:0'), in_proj_covar=tensor([0.0374, 0.0307, 0.0333, 0.0335, 0.0342, 0.0446, 0.0300, 0.0437], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 22:26:05,300 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99624.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:26:09,213 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99627.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:26:11,619 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99629.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:26:11,925 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.14 vs. limit=2.0 +2023-03-09 22:26:15,397 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.30 vs. limit=2.0 +2023-03-09 22:26:35,854 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99650.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:26:58,087 INFO [train.py:898] (0/4) Epoch 28, batch 1550, loss[loss=0.1552, simple_loss=0.2488, pruned_loss=0.03083, over 18537.00 frames. ], tot_loss[loss=0.1551, simple_loss=0.2463, pruned_loss=0.03192, over 3606714.40 frames. ], batch size: 49, lr: 4.01e-03, grad_scale: 16.0 +2023-03-09 22:27:05,107 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=99675.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:27:11,291 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99680.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:27:11,489 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=99680.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:27:43,190 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99707.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:27:45,104 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.790e+02 2.548e+02 3.050e+02 3.629e+02 6.290e+02, threshold=6.100e+02, percent-clipped=1.0 +2023-03-09 22:27:57,334 INFO [train.py:898] (0/4) Epoch 28, batch 1600, loss[loss=0.1614, simple_loss=0.2597, pruned_loss=0.0316, over 18339.00 frames. ], tot_loss[loss=0.155, simple_loss=0.246, pruned_loss=0.03205, over 3608208.54 frames. ], batch size: 56, lr: 4.01e-03, grad_scale: 16.0 +2023-03-09 22:28:22,786 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=99741.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:28:39,265 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=99755.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:28:56,207 INFO [train.py:898] (0/4) Epoch 28, batch 1650, loss[loss=0.1542, simple_loss=0.2444, pruned_loss=0.03204, over 18284.00 frames. ], tot_loss[loss=0.155, simple_loss=0.2459, pruned_loss=0.03202, over 3599281.14 frames. ], batch size: 49, lr: 4.01e-03, grad_scale: 16.0 +2023-03-09 22:28:56,606 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6634, 2.9364, 4.4992, 3.8001, 2.9540, 4.7059, 4.0164, 3.1145], + device='cuda:0'), covar=tensor([0.0532, 0.1506, 0.0278, 0.0448, 0.1463, 0.0213, 0.0527, 0.0917], + device='cuda:0'), in_proj_covar=tensor([0.0222, 0.0245, 0.0233, 0.0173, 0.0228, 0.0220, 0.0256, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 22:29:04,448 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99776.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:29:34,798 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.88 vs. limit=2.0 +2023-03-09 22:29:36,445 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5472, 3.4896, 2.2637, 4.4317, 3.0854, 4.1487, 2.5883, 3.8119], + device='cuda:0'), covar=tensor([0.0722, 0.0879, 0.1516, 0.0477, 0.0903, 0.0433, 0.1220, 0.0516], + device='cuda:0'), in_proj_covar=tensor([0.0226, 0.0234, 0.0196, 0.0299, 0.0199, 0.0274, 0.0208, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 22:29:42,828 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.627e+02 2.331e+02 2.829e+02 3.453e+02 7.033e+02, threshold=5.658e+02, percent-clipped=1.0 +2023-03-09 22:29:55,638 INFO [train.py:898] (0/4) Epoch 28, batch 1700, loss[loss=0.1827, simple_loss=0.2696, pruned_loss=0.04787, over 13416.00 frames. ], tot_loss[loss=0.1554, simple_loss=0.2462, pruned_loss=0.03225, over 3594005.34 frames. ], batch size: 129, lr: 4.01e-03, grad_scale: 16.0 +2023-03-09 22:30:01,616 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99824.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 22:30:03,982 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7927, 3.1789, 4.6027, 3.9261, 3.1091, 4.8201, 4.1546, 3.1181], + device='cuda:0'), covar=tensor([0.0547, 0.1430, 0.0322, 0.0462, 0.1408, 0.0238, 0.0568, 0.1021], + device='cuda:0'), in_proj_covar=tensor([0.0222, 0.0244, 0.0233, 0.0173, 0.0228, 0.0220, 0.0256, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 22:30:09,812 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5517, 3.4391, 2.2845, 4.4191, 3.0900, 4.1406, 2.5149, 3.8546], + device='cuda:0'), covar=tensor([0.0723, 0.0906, 0.1535, 0.0467, 0.0878, 0.0369, 0.1192, 0.0469], + device='cuda:0'), in_proj_covar=tensor([0.0226, 0.0234, 0.0196, 0.0299, 0.0200, 0.0275, 0.0208, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 22:30:14,402 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7760, 5.0469, 2.5621, 4.8679, 4.8015, 5.0449, 4.8428, 2.5277], + device='cuda:0'), covar=tensor([0.0264, 0.0071, 0.0888, 0.0097, 0.0072, 0.0069, 0.0099, 0.1108], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0084, 0.0099, 0.0100, 0.0091, 0.0081, 0.0088, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 22:30:54,424 INFO [train.py:898] (0/4) Epoch 28, batch 1750, loss[loss=0.1696, simple_loss=0.263, pruned_loss=0.03809, over 18343.00 frames. ], tot_loss[loss=0.1558, simple_loss=0.2468, pruned_loss=0.03238, over 3587658.25 frames. ], batch size: 56, lr: 4.01e-03, grad_scale: 16.0 +2023-03-09 22:31:12,607 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99885.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:31:17,847 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5782, 3.1508, 4.3999, 3.5526, 2.7790, 4.6044, 3.9267, 2.9649], + device='cuda:0'), covar=tensor([0.0551, 0.1248, 0.0253, 0.0524, 0.1440, 0.0206, 0.0563, 0.0949], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0246, 0.0235, 0.0174, 0.0229, 0.0221, 0.0257, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 22:31:31,546 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.78 vs. limit=2.0 +2023-03-09 22:31:39,960 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.785e+02 2.621e+02 3.029e+02 3.741e+02 5.430e+02, threshold=6.058e+02, percent-clipped=0.0 +2023-03-09 22:31:51,913 INFO [train.py:898] (0/4) Epoch 28, batch 1800, loss[loss=0.1586, simple_loss=0.2542, pruned_loss=0.03153, over 18087.00 frames. ], tot_loss[loss=0.1553, simple_loss=0.2462, pruned_loss=0.03219, over 3590460.16 frames. ], batch size: 62, lr: 4.01e-03, grad_scale: 16.0 +2023-03-09 22:31:58,528 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99924.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:31:58,586 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99924.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:32:08,641 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=99933.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:32:22,799 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=99945.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:32:33,202 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8402, 3.3189, 4.5361, 3.8327, 3.1864, 4.8281, 4.1347, 3.2031], + device='cuda:0'), covar=tensor([0.0487, 0.1274, 0.0301, 0.0445, 0.1302, 0.0227, 0.0506, 0.0888], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0247, 0.0235, 0.0174, 0.0230, 0.0222, 0.0257, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 22:32:51,448 INFO [train.py:898] (0/4) Epoch 28, batch 1850, loss[loss=0.1576, simple_loss=0.2555, pruned_loss=0.02981, over 18504.00 frames. ], tot_loss[loss=0.1549, simple_loss=0.2458, pruned_loss=0.03202, over 3584138.04 frames. ], batch size: 53, lr: 4.01e-03, grad_scale: 16.0 +2023-03-09 22:32:55,880 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=99972.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:33:05,201 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=99980.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:33:28,669 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-100000.pt +2023-03-09 22:33:41,507 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100007.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 22:33:43,361 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.762e+02 2.401e+02 2.680e+02 3.285e+02 5.327e+02, threshold=5.360e+02, percent-clipped=0.0 +2023-03-09 22:33:55,965 INFO [train.py:898] (0/4) Epoch 28, batch 1900, loss[loss=0.1403, simple_loss=0.2282, pruned_loss=0.02624, over 18500.00 frames. ], tot_loss[loss=0.1541, simple_loss=0.2449, pruned_loss=0.03167, over 3579421.21 frames. ], batch size: 47, lr: 4.01e-03, grad_scale: 16.0 +2023-03-09 22:34:06,788 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=100028.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:34:07,075 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7642, 4.1893, 2.3893, 4.0031, 5.1524, 2.5186, 3.7044, 4.0422], + device='cuda:0'), covar=tensor([0.0217, 0.1081, 0.1682, 0.0695, 0.0110, 0.1287, 0.0738, 0.0697], + device='cuda:0'), in_proj_covar=tensor([0.0187, 0.0284, 0.0212, 0.0202, 0.0145, 0.0187, 0.0225, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 22:34:09,413 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8360, 3.5879, 5.1247, 2.8621, 4.4199, 2.6052, 3.1151, 1.7886], + device='cuda:0'), covar=tensor([0.1302, 0.1078, 0.0146, 0.1078, 0.0517, 0.2700, 0.2738, 0.2422], + device='cuda:0'), in_proj_covar=tensor([0.0231, 0.0253, 0.0228, 0.0208, 0.0265, 0.0279, 0.0336, 0.0246], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 22:34:15,697 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100036.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:34:53,683 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100068.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 22:34:54,384 INFO [train.py:898] (0/4) Epoch 28, batch 1950, loss[loss=0.1688, simple_loss=0.2621, pruned_loss=0.03777, over 15999.00 frames. ], tot_loss[loss=0.1546, simple_loss=0.2456, pruned_loss=0.0318, over 3585402.24 frames. ], batch size: 94, lr: 4.00e-03, grad_scale: 16.0 +2023-03-09 22:35:03,452 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100076.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:35:41,392 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.807e+02 2.428e+02 2.748e+02 3.405e+02 9.642e+02, threshold=5.496e+02, percent-clipped=3.0 +2023-03-09 22:35:53,097 INFO [train.py:898] (0/4) Epoch 28, batch 2000, loss[loss=0.2079, simple_loss=0.2845, pruned_loss=0.06569, over 12158.00 frames. ], tot_loss[loss=0.155, simple_loss=0.2462, pruned_loss=0.03188, over 3581055.72 frames. ], batch size: 129, lr: 4.00e-03, grad_scale: 16.0 +2023-03-09 22:35:59,732 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=100124.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:35:59,912 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100124.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 22:36:52,990 INFO [train.py:898] (0/4) Epoch 28, batch 2050, loss[loss=0.1619, simple_loss=0.2618, pruned_loss=0.03098, over 16118.00 frames. ], tot_loss[loss=0.1554, simple_loss=0.2464, pruned_loss=0.03218, over 3569731.08 frames. ], batch size: 94, lr: 4.00e-03, grad_scale: 16.0 +2023-03-09 22:36:56,635 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=100172.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 22:36:58,905 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8488, 5.2874, 5.2893, 5.3492, 4.7598, 5.1800, 4.5792, 5.1661], + device='cuda:0'), covar=tensor([0.0247, 0.0304, 0.0219, 0.0398, 0.0406, 0.0255, 0.1119, 0.0353], + device='cuda:0'), in_proj_covar=tensor([0.0236, 0.0278, 0.0280, 0.0356, 0.0289, 0.0287, 0.0320, 0.0281], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 22:37:39,778 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.832e+02 2.490e+02 2.932e+02 3.409e+02 5.416e+02, threshold=5.865e+02, percent-clipped=0.0 +2023-03-09 22:37:46,958 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9994, 5.0631, 5.1239, 4.7589, 4.8209, 4.8389, 5.1898, 5.1573], + device='cuda:0'), covar=tensor([0.0072, 0.0059, 0.0052, 0.0107, 0.0053, 0.0159, 0.0057, 0.0074], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0075, 0.0080, 0.0100, 0.0080, 0.0109, 0.0092, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 22:37:51,129 INFO [train.py:898] (0/4) Epoch 28, batch 2100, loss[loss=0.1587, simple_loss=0.2558, pruned_loss=0.03083, over 17990.00 frames. ], tot_loss[loss=0.1558, simple_loss=0.2468, pruned_loss=0.03244, over 3570509.68 frames. ], batch size: 65, lr: 4.00e-03, grad_scale: 16.0 +2023-03-09 22:37:51,993 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-09 22:37:57,846 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100224.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:38:14,624 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100238.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:38:22,617 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100245.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:38:50,275 INFO [train.py:898] (0/4) Epoch 28, batch 2150, loss[loss=0.1509, simple_loss=0.2398, pruned_loss=0.03096, over 18294.00 frames. ], tot_loss[loss=0.1566, simple_loss=0.2478, pruned_loss=0.03267, over 3579220.21 frames. ], batch size: 49, lr: 4.00e-03, grad_scale: 16.0 +2023-03-09 22:38:54,621 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=100272.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:39:00,471 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3075, 5.2838, 5.6452, 5.6400, 5.2184, 6.1061, 5.7991, 5.3759], + device='cuda:0'), covar=tensor([0.1064, 0.0568, 0.0744, 0.0770, 0.1307, 0.0644, 0.0646, 0.1599], + device='cuda:0'), in_proj_covar=tensor([0.0375, 0.0305, 0.0333, 0.0334, 0.0341, 0.0448, 0.0299, 0.0441], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 22:39:10,604 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7188, 3.5395, 4.8659, 4.2224, 3.3097, 3.0217, 4.2429, 5.1018], + device='cuda:0'), covar=tensor([0.0806, 0.1491, 0.0205, 0.0454, 0.0992, 0.1186, 0.0432, 0.0261], + device='cuda:0'), in_proj_covar=tensor([0.0155, 0.0286, 0.0174, 0.0188, 0.0199, 0.0199, 0.0204, 0.0216], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 22:39:19,963 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=100293.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:39:27,050 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100299.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:39:38,058 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.888e+02 2.494e+02 2.925e+02 3.420e+02 7.055e+02, threshold=5.850e+02, percent-clipped=2.0 +2023-03-09 22:39:50,187 INFO [train.py:898] (0/4) Epoch 28, batch 2200, loss[loss=0.1411, simple_loss=0.2271, pruned_loss=0.02752, over 18241.00 frames. ], tot_loss[loss=0.1563, simple_loss=0.2476, pruned_loss=0.03256, over 3582921.29 frames. ], batch size: 45, lr: 4.00e-03, grad_scale: 16.0 +2023-03-09 22:39:59,992 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2940, 5.2861, 5.5698, 5.7232, 5.2912, 6.1731, 5.8420, 5.4198], + device='cuda:0'), covar=tensor([0.1135, 0.0593, 0.0870, 0.0652, 0.1305, 0.0680, 0.0607, 0.1814], + device='cuda:0'), in_proj_covar=tensor([0.0375, 0.0306, 0.0333, 0.0334, 0.0341, 0.0447, 0.0300, 0.0442], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 22:40:11,030 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100336.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:40:43,181 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100363.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 22:40:49,837 INFO [train.py:898] (0/4) Epoch 28, batch 2250, loss[loss=0.1638, simple_loss=0.2606, pruned_loss=0.03353, over 16114.00 frames. ], tot_loss[loss=0.1564, simple_loss=0.2478, pruned_loss=0.03249, over 3582870.49 frames. ], batch size: 94, lr: 4.00e-03, grad_scale: 16.0 +2023-03-09 22:41:07,382 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=100384.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:41:29,661 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100402.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:41:37,424 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.960e+02 2.460e+02 2.795e+02 3.389e+02 7.041e+02, threshold=5.590e+02, percent-clipped=1.0 +2023-03-09 22:41:49,311 INFO [train.py:898] (0/4) Epoch 28, batch 2300, loss[loss=0.1587, simple_loss=0.2604, pruned_loss=0.02847, over 18376.00 frames. ], tot_loss[loss=0.1564, simple_loss=0.2478, pruned_loss=0.03251, over 3590740.93 frames. ], batch size: 50, lr: 4.00e-03, grad_scale: 16.0 +2023-03-09 22:42:41,785 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100463.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 22:42:48,945 INFO [train.py:898] (0/4) Epoch 28, batch 2350, loss[loss=0.1504, simple_loss=0.2526, pruned_loss=0.02404, over 18378.00 frames. ], tot_loss[loss=0.1565, simple_loss=0.248, pruned_loss=0.03246, over 3588709.96 frames. ], batch size: 55, lr: 4.00e-03, grad_scale: 16.0 +2023-03-09 22:43:34,972 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7138, 3.7206, 3.5883, 3.1890, 3.5355, 2.9212, 2.9741, 3.7479], + device='cuda:0'), covar=tensor([0.0081, 0.0106, 0.0096, 0.0144, 0.0100, 0.0211, 0.0211, 0.0074], + device='cuda:0'), in_proj_covar=tensor([0.0159, 0.0178, 0.0148, 0.0198, 0.0157, 0.0190, 0.0193, 0.0136], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 22:43:36,869 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.732e+02 2.362e+02 2.730e+02 3.461e+02 8.407e+02, threshold=5.459e+02, percent-clipped=2.0 +2023-03-09 22:43:48,278 INFO [train.py:898] (0/4) Epoch 28, batch 2400, loss[loss=0.168, simple_loss=0.2644, pruned_loss=0.03578, over 17132.00 frames. ], tot_loss[loss=0.1563, simple_loss=0.2475, pruned_loss=0.03255, over 3578131.14 frames. ], batch size: 78, lr: 4.00e-03, grad_scale: 16.0 +2023-03-09 22:44:21,817 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6557, 2.9479, 2.5599, 2.9465, 3.6808, 3.6549, 3.2531, 2.9141], + device='cuda:0'), covar=tensor([0.0177, 0.0288, 0.0584, 0.0390, 0.0199, 0.0163, 0.0367, 0.0362], + device='cuda:0'), in_proj_covar=tensor([0.0147, 0.0147, 0.0168, 0.0167, 0.0143, 0.0129, 0.0161, 0.0164], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 22:44:46,656 INFO [train.py:898] (0/4) Epoch 28, batch 2450, loss[loss=0.1937, simple_loss=0.2769, pruned_loss=0.05531, over 18312.00 frames. ], tot_loss[loss=0.1559, simple_loss=0.2471, pruned_loss=0.0324, over 3593994.89 frames. ], batch size: 57, lr: 3.99e-03, grad_scale: 16.0 +2023-03-09 22:45:16,711 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100594.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:45:34,580 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.748e+02 2.510e+02 2.947e+02 3.432e+02 5.400e+02, threshold=5.894e+02, percent-clipped=0.0 +2023-03-09 22:45:45,731 INFO [train.py:898] (0/4) Epoch 28, batch 2500, loss[loss=0.1471, simple_loss=0.2339, pruned_loss=0.0301, over 18345.00 frames. ], tot_loss[loss=0.1558, simple_loss=0.247, pruned_loss=0.03235, over 3599249.83 frames. ], batch size: 46, lr: 3.99e-03, grad_scale: 16.0 +2023-03-09 22:45:46,137 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6049, 2.9363, 4.3464, 3.6278, 2.5977, 4.5868, 3.9006, 2.9467], + device='cuda:0'), covar=tensor([0.0529, 0.1408, 0.0316, 0.0501, 0.1579, 0.0207, 0.0590, 0.0896], + device='cuda:0'), in_proj_covar=tensor([0.0225, 0.0247, 0.0236, 0.0175, 0.0231, 0.0221, 0.0258, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 22:46:39,057 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100663.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 22:46:45,733 INFO [train.py:898] (0/4) Epoch 28, batch 2550, loss[loss=0.1583, simple_loss=0.2504, pruned_loss=0.03305, over 18070.00 frames. ], tot_loss[loss=0.1559, simple_loss=0.2471, pruned_loss=0.03233, over 3601591.74 frames. ], batch size: 62, lr: 3.99e-03, grad_scale: 16.0 +2023-03-09 22:47:23,720 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6017, 3.5214, 2.2590, 4.4005, 3.1346, 4.1201, 2.5689, 4.0272], + device='cuda:0'), covar=tensor([0.0683, 0.0846, 0.1558, 0.0547, 0.0825, 0.0396, 0.1183, 0.0384], + device='cuda:0'), in_proj_covar=tensor([0.0224, 0.0233, 0.0195, 0.0297, 0.0198, 0.0273, 0.0207, 0.0208], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 22:47:33,007 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.831e+02 2.422e+02 2.744e+02 3.332e+02 5.019e+02, threshold=5.488e+02, percent-clipped=0.0 +2023-03-09 22:47:36,123 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=100711.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 22:47:44,788 INFO [train.py:898] (0/4) Epoch 28, batch 2600, loss[loss=0.1573, simple_loss=0.2514, pruned_loss=0.03166, over 18500.00 frames. ], tot_loss[loss=0.1555, simple_loss=0.2467, pruned_loss=0.03216, over 3602308.61 frames. ], batch size: 51, lr: 3.99e-03, grad_scale: 16.0 +2023-03-09 22:47:58,478 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=100730.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 22:48:31,052 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=100758.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 22:48:43,895 INFO [train.py:898] (0/4) Epoch 28, batch 2650, loss[loss=0.1468, simple_loss=0.2481, pruned_loss=0.02271, over 18394.00 frames. ], tot_loss[loss=0.1558, simple_loss=0.2473, pruned_loss=0.03216, over 3601336.13 frames. ], batch size: 52, lr: 3.99e-03, grad_scale: 16.0 +2023-03-09 22:49:09,489 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=100791.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 22:49:31,295 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.698e+02 2.383e+02 2.847e+02 3.406e+02 8.018e+02, threshold=5.694e+02, percent-clipped=5.0 +2023-03-09 22:49:43,084 INFO [train.py:898] (0/4) Epoch 28, batch 2700, loss[loss=0.1576, simple_loss=0.2539, pruned_loss=0.03065, over 18355.00 frames. ], tot_loss[loss=0.1554, simple_loss=0.2468, pruned_loss=0.03203, over 3599593.12 frames. ], batch size: 55, lr: 3.99e-03, grad_scale: 16.0 +2023-03-09 22:50:41,488 INFO [train.py:898] (0/4) Epoch 28, batch 2750, loss[loss=0.1656, simple_loss=0.265, pruned_loss=0.03313, over 18114.00 frames. ], tot_loss[loss=0.1554, simple_loss=0.2466, pruned_loss=0.03209, over 3603859.54 frames. ], batch size: 62, lr: 3.99e-03, grad_scale: 16.0 +2023-03-09 22:51:05,214 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6983, 3.6165, 2.3288, 4.5025, 3.2595, 4.3405, 2.6872, 4.1376], + device='cuda:0'), covar=tensor([0.0604, 0.0797, 0.1418, 0.0577, 0.0755, 0.0396, 0.1112, 0.0388], + device='cuda:0'), in_proj_covar=tensor([0.0224, 0.0234, 0.0195, 0.0299, 0.0199, 0.0273, 0.0207, 0.0209], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 22:51:10,561 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=100894.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:51:16,805 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.31 vs. limit=2.0 +2023-03-09 22:51:21,996 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8411, 5.3748, 5.3531, 5.3769, 4.8382, 5.2666, 4.7122, 5.2259], + device='cuda:0'), covar=tensor([0.0273, 0.0258, 0.0193, 0.0434, 0.0387, 0.0212, 0.1075, 0.0361], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0276, 0.0278, 0.0355, 0.0288, 0.0285, 0.0319, 0.0279], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 22:51:27,174 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.680e+02 2.345e+02 2.707e+02 3.433e+02 5.623e+02, threshold=5.414e+02, percent-clipped=0.0 +2023-03-09 22:51:40,783 INFO [train.py:898] (0/4) Epoch 28, batch 2800, loss[loss=0.1542, simple_loss=0.2521, pruned_loss=0.02808, over 18476.00 frames. ], tot_loss[loss=0.1551, simple_loss=0.2462, pruned_loss=0.03201, over 3601675.16 frames. ], batch size: 53, lr: 3.99e-03, grad_scale: 16.0 +2023-03-09 22:52:03,010 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5878, 2.3234, 2.5402, 2.5224, 2.9921, 4.5336, 4.3763, 3.3490], + device='cuda:0'), covar=tensor([0.2030, 0.2653, 0.3077, 0.2144, 0.2699, 0.0345, 0.0443, 0.0994], + device='cuda:0'), in_proj_covar=tensor([0.0333, 0.0367, 0.0417, 0.0293, 0.0400, 0.0269, 0.0305, 0.0275], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0002], + device='cuda:0') +2023-03-09 22:52:07,432 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=100942.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:52:38,800 INFO [train.py:898] (0/4) Epoch 28, batch 2850, loss[loss=0.1388, simple_loss=0.2182, pruned_loss=0.02973, over 18408.00 frames. ], tot_loss[loss=0.155, simple_loss=0.246, pruned_loss=0.03203, over 3606080.78 frames. ], batch size: 42, lr: 3.99e-03, grad_scale: 16.0 +2023-03-09 22:53:08,959 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7370, 3.6863, 5.1532, 2.9023, 4.4578, 2.6438, 2.9982, 1.8434], + device='cuda:0'), covar=tensor([0.1358, 0.0979, 0.0166, 0.1013, 0.0536, 0.2700, 0.2997, 0.2401], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0256, 0.0232, 0.0211, 0.0269, 0.0283, 0.0341, 0.0250], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 22:53:14,102 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5694, 2.2911, 2.5157, 2.6117, 2.9862, 4.5031, 4.4387, 3.1518], + device='cuda:0'), covar=tensor([0.2156, 0.2659, 0.3179, 0.2027, 0.2703, 0.0359, 0.0418, 0.1115], + device='cuda:0'), in_proj_covar=tensor([0.0333, 0.0366, 0.0417, 0.0292, 0.0398, 0.0269, 0.0304, 0.0275], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 22:53:26,396 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.467e+02 2.435e+02 2.818e+02 3.350e+02 5.539e+02, threshold=5.635e+02, percent-clipped=3.0 +2023-03-09 22:53:37,851 INFO [train.py:898] (0/4) Epoch 28, batch 2900, loss[loss=0.1404, simple_loss=0.2303, pruned_loss=0.02522, over 18506.00 frames. ], tot_loss[loss=0.1554, simple_loss=0.2466, pruned_loss=0.03208, over 3593812.35 frames. ], batch size: 47, lr: 3.99e-03, grad_scale: 16.0 +2023-03-09 22:54:25,163 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=101058.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:54:37,493 INFO [train.py:898] (0/4) Epoch 28, batch 2950, loss[loss=0.1434, simple_loss=0.2314, pruned_loss=0.02772, over 18275.00 frames. ], tot_loss[loss=0.1546, simple_loss=0.2458, pruned_loss=0.03168, over 3603022.60 frames. ], batch size: 47, lr: 3.98e-03, grad_scale: 16.0 +2023-03-09 22:54:58,144 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=101086.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 22:55:21,289 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=101106.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:55:22,643 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101107.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:55:24,471 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.540e+02 2.503e+02 3.089e+02 3.620e+02 7.481e+02, threshold=6.178e+02, percent-clipped=1.0 +2023-03-09 22:55:35,814 INFO [train.py:898] (0/4) Epoch 28, batch 3000, loss[loss=0.1505, simple_loss=0.2419, pruned_loss=0.02957, over 16137.00 frames. ], tot_loss[loss=0.1552, simple_loss=0.2466, pruned_loss=0.0319, over 3596258.18 frames. ], batch size: 94, lr: 3.98e-03, grad_scale: 16.0 +2023-03-09 22:55:35,816 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 22:55:47,893 INFO [train.py:932] (0/4) Epoch 28, validation: loss=0.1496, simple_loss=0.2475, pruned_loss=0.02587, over 944034.00 frames. +2023-03-09 22:55:47,894 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 22:56:45,976 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=101168.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 22:56:46,679 INFO [train.py:898] (0/4) Epoch 28, batch 3050, loss[loss=0.1618, simple_loss=0.251, pruned_loss=0.03635, over 18395.00 frames. ], tot_loss[loss=0.1553, simple_loss=0.2463, pruned_loss=0.03216, over 3592434.35 frames. ], batch size: 50, lr: 3.98e-03, grad_scale: 16.0 +2023-03-09 22:57:32,632 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.576e+02 2.344e+02 2.759e+02 3.609e+02 9.420e+02, threshold=5.517e+02, percent-clipped=1.0 +2023-03-09 22:57:45,999 INFO [train.py:898] (0/4) Epoch 28, batch 3100, loss[loss=0.1598, simple_loss=0.2549, pruned_loss=0.03234, over 16976.00 frames. ], tot_loss[loss=0.1562, simple_loss=0.2472, pruned_loss=0.03259, over 3577607.55 frames. ], batch size: 78, lr: 3.98e-03, grad_scale: 16.0 +2023-03-09 22:58:32,509 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8247, 3.6562, 4.8363, 4.2793, 3.1655, 2.9226, 4.2908, 5.1739], + device='cuda:0'), covar=tensor([0.0802, 0.1386, 0.0266, 0.0443, 0.1115, 0.1296, 0.0452, 0.0227], + device='cuda:0'), in_proj_covar=tensor([0.0156, 0.0288, 0.0177, 0.0190, 0.0200, 0.0200, 0.0204, 0.0216], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 22:58:45,529 INFO [train.py:898] (0/4) Epoch 28, batch 3150, loss[loss=0.1942, simple_loss=0.2748, pruned_loss=0.05677, over 12136.00 frames. ], tot_loss[loss=0.1558, simple_loss=0.2466, pruned_loss=0.03247, over 3558349.95 frames. ], batch size: 130, lr: 3.98e-03, grad_scale: 16.0 +2023-03-09 22:58:51,656 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4972, 6.0566, 5.6543, 5.8257, 5.6301, 5.4114, 6.1064, 6.0424], + device='cuda:0'), covar=tensor([0.1171, 0.0781, 0.0411, 0.0672, 0.1482, 0.0725, 0.0554, 0.0696], + device='cuda:0'), in_proj_covar=tensor([0.0644, 0.0569, 0.0408, 0.0590, 0.0788, 0.0589, 0.0801, 0.0618], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 22:59:31,374 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.645e+02 2.594e+02 3.004e+02 3.371e+02 7.669e+02, threshold=6.008e+02, percent-clipped=3.0 +2023-03-09 22:59:43,364 INFO [train.py:898] (0/4) Epoch 28, batch 3200, loss[loss=0.1396, simple_loss=0.2144, pruned_loss=0.03241, over 17775.00 frames. ], tot_loss[loss=0.1553, simple_loss=0.2459, pruned_loss=0.03235, over 3559590.49 frames. ], batch size: 39, lr: 3.98e-03, grad_scale: 16.0 +2023-03-09 23:00:41,853 INFO [train.py:898] (0/4) Epoch 28, batch 3250, loss[loss=0.1491, simple_loss=0.2317, pruned_loss=0.03322, over 18408.00 frames. ], tot_loss[loss=0.1554, simple_loss=0.246, pruned_loss=0.03242, over 3573623.07 frames. ], batch size: 42, lr: 3.98e-03, grad_scale: 16.0 +2023-03-09 23:00:59,541 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8166, 3.6217, 5.0891, 3.0181, 4.4163, 2.5258, 3.0737, 1.7552], + device='cuda:0'), covar=tensor([0.1294, 0.0993, 0.0178, 0.0955, 0.0493, 0.2741, 0.2725, 0.2389], + device='cuda:0'), in_proj_covar=tensor([0.0232, 0.0254, 0.0230, 0.0209, 0.0266, 0.0282, 0.0338, 0.0248], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 23:01:02,694 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=101386.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 23:01:29,437 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.724e+02 2.697e+02 3.326e+02 3.955e+02 8.507e+02, threshold=6.653e+02, percent-clipped=3.0 +2023-03-09 23:01:40,850 INFO [train.py:898] (0/4) Epoch 28, batch 3300, loss[loss=0.14, simple_loss=0.2265, pruned_loss=0.02674, over 18500.00 frames. ], tot_loss[loss=0.1557, simple_loss=0.2462, pruned_loss=0.03258, over 3561999.43 frames. ], batch size: 47, lr: 3.98e-03, grad_scale: 32.0 +2023-03-09 23:02:00,153 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=101434.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 23:02:02,535 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8005, 2.8939, 2.0036, 3.3160, 2.3829, 2.5509, 2.2424, 2.7445], + device='cuda:0'), covar=tensor([0.0605, 0.0776, 0.1367, 0.0616, 0.0879, 0.0286, 0.1180, 0.0602], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0232, 0.0193, 0.0295, 0.0197, 0.0269, 0.0206, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 23:02:17,656 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7195, 3.6175, 4.9185, 4.3061, 3.2909, 2.9453, 4.2984, 5.1951], + device='cuda:0'), covar=tensor([0.0883, 0.1546, 0.0261, 0.0475, 0.1029, 0.1320, 0.0472, 0.0238], + device='cuda:0'), in_proj_covar=tensor([0.0154, 0.0285, 0.0174, 0.0188, 0.0198, 0.0198, 0.0201, 0.0214], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 23:02:33,270 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=101463.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:02:39,848 INFO [train.py:898] (0/4) Epoch 28, batch 3350, loss[loss=0.1927, simple_loss=0.2741, pruned_loss=0.05566, over 12675.00 frames. ], tot_loss[loss=0.1561, simple_loss=0.2469, pruned_loss=0.03261, over 3560536.66 frames. ], batch size: 129, lr: 3.98e-03, grad_scale: 16.0 +2023-03-09 23:02:57,782 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5034, 3.3767, 2.1467, 4.2595, 2.9985, 3.9556, 2.4190, 3.6709], + device='cuda:0'), covar=tensor([0.0650, 0.0844, 0.1538, 0.0551, 0.0863, 0.0315, 0.1227, 0.0484], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0232, 0.0194, 0.0296, 0.0197, 0.0270, 0.0207, 0.0207], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 23:03:25,419 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8374, 3.6280, 5.2303, 3.2718, 4.5103, 2.5812, 3.1065, 1.7424], + device='cuda:0'), covar=tensor([0.1257, 0.0918, 0.0131, 0.0828, 0.0503, 0.2628, 0.2664, 0.2324], + device='cuda:0'), in_proj_covar=tensor([0.0232, 0.0254, 0.0230, 0.0210, 0.0267, 0.0282, 0.0339, 0.0249], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 23:03:28,358 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.802e+02 2.584e+02 3.127e+02 3.784e+02 7.730e+02, threshold=6.254e+02, percent-clipped=1.0 +2023-03-09 23:03:31,232 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.18 vs. limit=2.0 +2023-03-09 23:03:38,772 INFO [train.py:898] (0/4) Epoch 28, batch 3400, loss[loss=0.1721, simple_loss=0.2591, pruned_loss=0.04259, over 18354.00 frames. ], tot_loss[loss=0.1559, simple_loss=0.2466, pruned_loss=0.0326, over 3564686.74 frames. ], batch size: 56, lr: 3.98e-03, grad_scale: 16.0 +2023-03-09 23:04:38,397 INFO [train.py:898] (0/4) Epoch 28, batch 3450, loss[loss=0.1561, simple_loss=0.2484, pruned_loss=0.03186, over 18151.00 frames. ], tot_loss[loss=0.1557, simple_loss=0.2463, pruned_loss=0.03252, over 3562188.54 frames. ], batch size: 62, lr: 3.98e-03, grad_scale: 16.0 +2023-03-09 23:04:42,146 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6061, 2.3786, 2.6027, 2.5845, 3.1352, 4.7798, 4.6414, 3.3534], + device='cuda:0'), covar=tensor([0.2240, 0.2786, 0.3002, 0.2145, 0.2590, 0.0310, 0.0415, 0.1091], + device='cuda:0'), in_proj_covar=tensor([0.0332, 0.0365, 0.0417, 0.0292, 0.0398, 0.0269, 0.0303, 0.0274], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 23:05:28,087 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.546e+02 2.589e+02 3.041e+02 3.646e+02 6.515e+02, threshold=6.081e+02, percent-clipped=1.0 +2023-03-09 23:05:37,347 INFO [train.py:898] (0/4) Epoch 28, batch 3500, loss[loss=0.166, simple_loss=0.2643, pruned_loss=0.03389, over 18502.00 frames. ], tot_loss[loss=0.1555, simple_loss=0.2461, pruned_loss=0.03243, over 3576226.75 frames. ], batch size: 53, lr: 3.97e-03, grad_scale: 8.0 +2023-03-09 23:06:34,325 INFO [train.py:898] (0/4) Epoch 28, batch 3550, loss[loss=0.2095, simple_loss=0.2878, pruned_loss=0.06556, over 12405.00 frames. ], tot_loss[loss=0.1554, simple_loss=0.2458, pruned_loss=0.03252, over 3568383.27 frames. ], batch size: 130, lr: 3.97e-03, grad_scale: 8.0 +2023-03-09 23:06:50,646 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101684.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:07:08,793 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3801, 5.8902, 5.5022, 5.6227, 5.4782, 5.2973, 5.9604, 5.8823], + device='cuda:0'), covar=tensor([0.1225, 0.0870, 0.0527, 0.0762, 0.1473, 0.0845, 0.0607, 0.0773], + device='cuda:0'), in_proj_covar=tensor([0.0645, 0.0569, 0.0408, 0.0594, 0.0788, 0.0590, 0.0805, 0.0618], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 23:07:19,608 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.673e+02 2.328e+02 2.616e+02 3.053e+02 4.840e+02, threshold=5.233e+02, percent-clipped=0.0 +2023-03-09 23:07:28,385 INFO [train.py:898] (0/4) Epoch 28, batch 3600, loss[loss=0.1647, simple_loss=0.2596, pruned_loss=0.0349, over 17735.00 frames. ], tot_loss[loss=0.1558, simple_loss=0.2466, pruned_loss=0.03253, over 3578959.75 frames. ], batch size: 70, lr: 3.97e-03, grad_scale: 8.0 +2023-03-09 23:07:51,902 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4866, 2.9222, 4.2764, 3.6616, 2.7063, 4.5330, 3.9013, 2.9232], + device='cuda:0'), covar=tensor([0.0582, 0.1532, 0.0337, 0.0483, 0.1516, 0.0238, 0.0598, 0.0960], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0246, 0.0235, 0.0174, 0.0228, 0.0219, 0.0258, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 23:07:57,171 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=101745.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:08:02,781 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.1611, 4.1342, 3.8489, 4.0725, 4.1158, 3.6351, 4.0852, 3.7984], + device='cuda:0'), covar=tensor([0.0627, 0.0808, 0.1454, 0.0958, 0.0762, 0.0668, 0.0637, 0.1386], + device='cuda:0'), in_proj_covar=tensor([0.0522, 0.0590, 0.0730, 0.0458, 0.0483, 0.0540, 0.0573, 0.0711], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0005, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 23:08:05,109 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-28.pt +2023-03-09 23:08:29,829 INFO [train.py:898] (0/4) Epoch 29, batch 0, loss[loss=0.1844, simple_loss=0.2803, pruned_loss=0.04426, over 17730.00 frames. ], tot_loss[loss=0.1844, simple_loss=0.2803, pruned_loss=0.04426, over 17730.00 frames. ], batch size: 70, lr: 3.90e-03, grad_scale: 8.0 +2023-03-09 23:08:29,831 INFO [train.py:923] (0/4) Computing validation loss +2023-03-09 23:08:41,778 INFO [train.py:932] (0/4) Epoch 29, validation: loss=0.1494, simple_loss=0.2476, pruned_loss=0.02556, over 944034.00 frames. +2023-03-09 23:08:41,779 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-09 23:08:54,470 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=101763.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:09:40,295 INFO [train.py:898] (0/4) Epoch 29, batch 50, loss[loss=0.1351, simple_loss=0.2247, pruned_loss=0.02276, over 18507.00 frames. ], tot_loss[loss=0.1526, simple_loss=0.2435, pruned_loss=0.03083, over 821969.89 frames. ], batch size: 47, lr: 3.90e-03, grad_scale: 4.0 +2023-03-09 23:09:50,575 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=101811.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:09:51,449 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.892e+02 2.454e+02 2.836e+02 3.650e+02 1.455e+03, threshold=5.673e+02, percent-clipped=9.0 +2023-03-09 23:10:10,830 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5341, 2.8537, 2.3927, 2.8174, 3.5933, 3.5369, 3.0895, 2.8014], + device='cuda:0'), covar=tensor([0.0170, 0.0296, 0.0612, 0.0406, 0.0191, 0.0176, 0.0414, 0.0408], + device='cuda:0'), in_proj_covar=tensor([0.0148, 0.0146, 0.0168, 0.0167, 0.0143, 0.0131, 0.0162, 0.0164], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 23:10:38,661 INFO [train.py:898] (0/4) Epoch 29, batch 100, loss[loss=0.1683, simple_loss=0.269, pruned_loss=0.03378, over 18304.00 frames. ], tot_loss[loss=0.1563, simple_loss=0.2474, pruned_loss=0.03262, over 1439984.88 frames. ], batch size: 57, lr: 3.90e-03, grad_scale: 4.0 +2023-03-09 23:11:38,507 INFO [train.py:898] (0/4) Epoch 29, batch 150, loss[loss=0.1701, simple_loss=0.2545, pruned_loss=0.04287, over 17009.00 frames. ], tot_loss[loss=0.1567, simple_loss=0.2478, pruned_loss=0.03276, over 1915840.14 frames. ], batch size: 78, lr: 3.90e-03, grad_scale: 4.0 +2023-03-09 23:11:49,336 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.704e+02 2.518e+02 2.968e+02 3.653e+02 8.556e+02, threshold=5.936e+02, percent-clipped=5.0 +2023-03-09 23:12:00,466 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101920.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 23:12:10,749 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8576, 5.4187, 2.8744, 5.2601, 5.2055, 5.4482, 5.2276, 3.0467], + device='cuda:0'), covar=tensor([0.0271, 0.0085, 0.0774, 0.0084, 0.0070, 0.0078, 0.0101, 0.0842], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0084, 0.0098, 0.0099, 0.0092, 0.0080, 0.0087, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-09 23:12:14,256 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9043, 4.2163, 4.1290, 4.2393, 3.8315, 4.1272, 3.8354, 4.1507], + device='cuda:0'), covar=tensor([0.0283, 0.0346, 0.0266, 0.0488, 0.0321, 0.0249, 0.0804, 0.0322], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0279, 0.0279, 0.0358, 0.0289, 0.0287, 0.0318, 0.0280], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 23:12:38,430 INFO [train.py:898] (0/4) Epoch 29, batch 200, loss[loss=0.1401, simple_loss=0.2201, pruned_loss=0.03007, over 18450.00 frames. ], tot_loss[loss=0.1566, simple_loss=0.2475, pruned_loss=0.0328, over 2277470.42 frames. ], batch size: 43, lr: 3.90e-03, grad_scale: 4.0 +2023-03-09 23:13:10,353 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.44 vs. limit=2.0 +2023-03-09 23:13:12,299 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=101981.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 23:13:18,841 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=101987.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:13:34,491 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-102000.pt +2023-03-09 23:13:42,398 INFO [train.py:898] (0/4) Epoch 29, batch 250, loss[loss=0.1521, simple_loss=0.238, pruned_loss=0.03306, over 18365.00 frames. ], tot_loss[loss=0.155, simple_loss=0.2456, pruned_loss=0.03217, over 2580863.78 frames. ], batch size: 46, lr: 3.90e-03, grad_scale: 4.0 +2023-03-09 23:13:52,741 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.574e+02 2.573e+02 2.918e+02 3.581e+02 7.098e+02, threshold=5.836e+02, percent-clipped=3.0 +2023-03-09 23:14:26,062 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=102040.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:14:35,287 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=102048.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:14:41,077 INFO [train.py:898] (0/4) Epoch 29, batch 300, loss[loss=0.137, simple_loss=0.2238, pruned_loss=0.02509, over 18261.00 frames. ], tot_loss[loss=0.1543, simple_loss=0.2446, pruned_loss=0.03195, over 2808031.94 frames. ], batch size: 47, lr: 3.90e-03, grad_scale: 4.0 +2023-03-09 23:14:44,877 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1100, 5.5662, 5.5391, 5.5673, 5.0419, 5.4826, 4.9445, 5.4666], + device='cuda:0'), covar=tensor([0.0237, 0.0238, 0.0166, 0.0366, 0.0376, 0.0211, 0.0950, 0.0257], + device='cuda:0'), in_proj_covar=tensor([0.0234, 0.0277, 0.0277, 0.0354, 0.0287, 0.0286, 0.0316, 0.0278], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 23:14:47,225 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7204, 3.2061, 4.4082, 3.8625, 3.0070, 4.6637, 4.0463, 2.9550], + device='cuda:0'), covar=tensor([0.0532, 0.1355, 0.0350, 0.0455, 0.1454, 0.0249, 0.0550, 0.1037], + device='cuda:0'), in_proj_covar=tensor([0.0225, 0.0248, 0.0236, 0.0175, 0.0230, 0.0222, 0.0260, 0.0200], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 23:14:54,259 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7036, 2.9665, 2.6927, 2.9737, 3.7154, 3.7155, 3.2193, 2.9985], + device='cuda:0'), covar=tensor([0.0212, 0.0299, 0.0534, 0.0375, 0.0186, 0.0164, 0.0426, 0.0437], + device='cuda:0'), in_proj_covar=tensor([0.0149, 0.0148, 0.0168, 0.0167, 0.0144, 0.0132, 0.0163, 0.0167], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 23:15:02,760 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1159, 5.2152, 5.2383, 4.9668, 4.9658, 4.9301, 5.2708, 5.3277], + device='cuda:0'), covar=tensor([0.0077, 0.0085, 0.0069, 0.0136, 0.0061, 0.0167, 0.0123, 0.0120], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0075, 0.0081, 0.0101, 0.0079, 0.0109, 0.0092, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 23:15:40,909 INFO [train.py:898] (0/4) Epoch 29, batch 350, loss[loss=0.1522, simple_loss=0.2396, pruned_loss=0.03241, over 18543.00 frames. ], tot_loss[loss=0.1535, simple_loss=0.2439, pruned_loss=0.03151, over 2995687.41 frames. ], batch size: 49, lr: 3.89e-03, grad_scale: 4.0 +2023-03-09 23:15:51,452 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.890e+02 2.462e+02 2.956e+02 3.341e+02 6.335e+02, threshold=5.913e+02, percent-clipped=2.0 +2023-03-09 23:16:13,370 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4137, 5.3826, 5.0359, 5.2893, 5.3167, 4.6145, 5.1970, 4.9574], + device='cuda:0'), covar=tensor([0.0377, 0.0417, 0.1128, 0.0753, 0.0585, 0.0453, 0.0433, 0.1129], + device='cuda:0'), in_proj_covar=tensor([0.0526, 0.0596, 0.0738, 0.0464, 0.0489, 0.0546, 0.0577, 0.0719], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0005, 0.0005, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 23:16:40,563 INFO [train.py:898] (0/4) Epoch 29, batch 400, loss[loss=0.1484, simple_loss=0.2459, pruned_loss=0.0255, over 18494.00 frames. ], tot_loss[loss=0.1532, simple_loss=0.2438, pruned_loss=0.03128, over 3127141.80 frames. ], batch size: 51, lr: 3.89e-03, grad_scale: 8.0 +2023-03-09 23:17:40,601 INFO [train.py:898] (0/4) Epoch 29, batch 450, loss[loss=0.1708, simple_loss=0.2651, pruned_loss=0.03825, over 18341.00 frames. ], tot_loss[loss=0.154, simple_loss=0.2449, pruned_loss=0.03151, over 3231704.34 frames. ], batch size: 56, lr: 3.89e-03, grad_scale: 8.0 +2023-03-09 23:17:50,993 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.718e+02 2.619e+02 3.042e+02 3.635e+02 5.576e+02, threshold=6.084e+02, percent-clipped=0.0 +2023-03-09 23:18:13,721 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5871, 3.5646, 3.4389, 3.0646, 3.3129, 2.8230, 2.7070, 3.6161], + device='cuda:0'), covar=tensor([0.0075, 0.0119, 0.0096, 0.0151, 0.0114, 0.0192, 0.0228, 0.0074], + device='cuda:0'), in_proj_covar=tensor([0.0159, 0.0179, 0.0149, 0.0197, 0.0158, 0.0190, 0.0194, 0.0137], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 23:18:40,653 INFO [train.py:898] (0/4) Epoch 29, batch 500, loss[loss=0.1429, simple_loss=0.2325, pruned_loss=0.02662, over 18507.00 frames. ], tot_loss[loss=0.1534, simple_loss=0.2444, pruned_loss=0.03119, over 3321097.70 frames. ], batch size: 47, lr: 3.89e-03, grad_scale: 8.0 +2023-03-09 23:19:07,693 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=102276.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 23:19:40,228 INFO [train.py:898] (0/4) Epoch 29, batch 550, loss[loss=0.1406, simple_loss=0.2282, pruned_loss=0.0265, over 18501.00 frames. ], tot_loss[loss=0.1538, simple_loss=0.2447, pruned_loss=0.03143, over 3372046.61 frames. ], batch size: 47, lr: 3.89e-03, grad_scale: 8.0 +2023-03-09 23:19:50,536 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.948e+02 2.553e+02 3.006e+02 3.452e+02 5.617e+02, threshold=6.011e+02, percent-clipped=0.0 +2023-03-09 23:20:07,468 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8184, 3.3922, 4.6694, 3.9629, 3.2617, 4.8766, 4.2299, 3.2816], + device='cuda:0'), covar=tensor([0.0543, 0.1249, 0.0270, 0.0476, 0.1292, 0.0250, 0.0494, 0.0869], + device='cuda:0'), in_proj_covar=tensor([0.0226, 0.0249, 0.0237, 0.0175, 0.0231, 0.0222, 0.0261, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 23:20:24,176 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=102340.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:20:28,125 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=102343.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:20:37,293 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.59 vs. limit=2.0 +2023-03-09 23:20:39,915 INFO [train.py:898] (0/4) Epoch 29, batch 600, loss[loss=0.1742, simple_loss=0.2693, pruned_loss=0.0395, over 17974.00 frames. ], tot_loss[loss=0.154, simple_loss=0.245, pruned_loss=0.03154, over 3423858.81 frames. ], batch size: 65, lr: 3.89e-03, grad_scale: 8.0 +2023-03-09 23:21:17,830 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5690, 2.3505, 2.5586, 2.5327, 3.1013, 4.7808, 4.7424, 3.2284], + device='cuda:0'), covar=tensor([0.2234, 0.2673, 0.3128, 0.2214, 0.2698, 0.0289, 0.0366, 0.1122], + device='cuda:0'), in_proj_covar=tensor([0.0332, 0.0365, 0.0416, 0.0292, 0.0398, 0.0270, 0.0303, 0.0274], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0001], + device='cuda:0') +2023-03-09 23:21:20,767 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=102388.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:21:39,517 INFO [train.py:898] (0/4) Epoch 29, batch 650, loss[loss=0.1559, simple_loss=0.2589, pruned_loss=0.02645, over 18500.00 frames. ], tot_loss[loss=0.1546, simple_loss=0.2455, pruned_loss=0.03183, over 3453446.17 frames. ], batch size: 53, lr: 3.89e-03, grad_scale: 8.0 +2023-03-09 23:21:49,329 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.777e+02 2.518e+02 2.994e+02 3.652e+02 5.297e+02, threshold=5.988e+02, percent-clipped=0.0 +2023-03-09 23:22:06,446 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0671, 5.1865, 5.3563, 5.4202, 5.0359, 5.9248, 5.5636, 5.1652], + device='cuda:0'), covar=tensor([0.1205, 0.0758, 0.0773, 0.0915, 0.1403, 0.0766, 0.0777, 0.1941], + device='cuda:0'), in_proj_covar=tensor([0.0381, 0.0311, 0.0338, 0.0339, 0.0347, 0.0452, 0.0306, 0.0448], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 23:22:38,250 INFO [train.py:898] (0/4) Epoch 29, batch 700, loss[loss=0.19, simple_loss=0.2781, pruned_loss=0.0509, over 18360.00 frames. ], tot_loss[loss=0.1553, simple_loss=0.2464, pruned_loss=0.03212, over 3485930.14 frames. ], batch size: 56, lr: 3.89e-03, grad_scale: 8.0 +2023-03-09 23:23:37,713 INFO [train.py:898] (0/4) Epoch 29, batch 750, loss[loss=0.163, simple_loss=0.2573, pruned_loss=0.03436, over 18470.00 frames. ], tot_loss[loss=0.1553, simple_loss=0.2462, pruned_loss=0.03219, over 3510619.45 frames. ], batch size: 59, lr: 3.89e-03, grad_scale: 8.0 +2023-03-09 23:23:48,499 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.674e+02 2.550e+02 3.053e+02 3.427e+02 7.248e+02, threshold=6.105e+02, percent-clipped=3.0 +2023-03-09 23:23:58,979 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5904, 4.6252, 4.6773, 4.4048, 4.4394, 4.4123, 4.7451, 4.7395], + device='cuda:0'), covar=tensor([0.0083, 0.0079, 0.0073, 0.0128, 0.0075, 0.0174, 0.0077, 0.0108], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0074, 0.0080, 0.0100, 0.0079, 0.0108, 0.0091, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 23:24:18,286 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8739, 5.3933, 5.3745, 5.3740, 4.8901, 5.3297, 4.7696, 5.2503], + device='cuda:0'), covar=tensor([0.0251, 0.0275, 0.0181, 0.0513, 0.0394, 0.0217, 0.1022, 0.0330], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0281, 0.0280, 0.0360, 0.0290, 0.0289, 0.0321, 0.0281], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 23:24:35,932 INFO [train.py:898] (0/4) Epoch 29, batch 800, loss[loss=0.1336, simple_loss=0.2254, pruned_loss=0.02084, over 18268.00 frames. ], tot_loss[loss=0.1557, simple_loss=0.2467, pruned_loss=0.03234, over 3534695.29 frames. ], batch size: 47, lr: 3.89e-03, grad_scale: 8.0 +2023-03-09 23:24:57,307 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.25 vs. limit=2.0 +2023-03-09 23:25:03,656 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=102576.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 23:25:35,087 INFO [train.py:898] (0/4) Epoch 29, batch 850, loss[loss=0.1718, simple_loss=0.2582, pruned_loss=0.04276, over 16080.00 frames. ], tot_loss[loss=0.1555, simple_loss=0.2464, pruned_loss=0.03234, over 3537758.87 frames. ], batch size: 94, lr: 3.88e-03, grad_scale: 8.0 +2023-03-09 23:25:44,477 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.87 vs. limit=5.0 +2023-03-09 23:25:45,055 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8331, 4.0720, 2.3544, 3.9195, 5.1992, 2.5780, 3.7993, 4.0295], + device='cuda:0'), covar=tensor([0.0211, 0.1158, 0.1708, 0.0672, 0.0108, 0.1243, 0.0718, 0.0682], + device='cuda:0'), in_proj_covar=tensor([0.0189, 0.0288, 0.0213, 0.0204, 0.0147, 0.0188, 0.0226, 0.0236], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 23:25:45,741 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.923e+02 2.608e+02 3.070e+02 3.838e+02 1.078e+03, threshold=6.141e+02, percent-clipped=3.0 +2023-03-09 23:25:47,192 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6838, 3.2932, 4.5434, 3.8708, 2.9120, 4.7827, 3.9947, 3.0927], + device='cuda:0'), covar=tensor([0.0575, 0.1240, 0.0283, 0.0446, 0.1488, 0.0208, 0.0583, 0.0941], + device='cuda:0'), in_proj_covar=tensor([0.0227, 0.0249, 0.0237, 0.0175, 0.0231, 0.0222, 0.0262, 0.0202], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-09 23:26:00,158 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=102624.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 23:26:10,175 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8680, 5.3560, 5.2880, 5.3252, 4.8051, 5.2487, 4.6914, 5.2297], + device='cuda:0'), covar=tensor([0.0261, 0.0261, 0.0208, 0.0430, 0.0434, 0.0231, 0.1093, 0.0309], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0282, 0.0282, 0.0362, 0.0291, 0.0291, 0.0322, 0.0282], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 23:26:22,217 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=102643.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:26:34,016 INFO [train.py:898] (0/4) Epoch 29, batch 900, loss[loss=0.1572, simple_loss=0.2545, pruned_loss=0.02997, over 18318.00 frames. ], tot_loss[loss=0.1549, simple_loss=0.2458, pruned_loss=0.03199, over 3554591.79 frames. ], batch size: 57, lr: 3.88e-03, grad_scale: 8.0 +2023-03-09 23:26:39,896 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5471, 6.0544, 5.6494, 5.8497, 5.6683, 5.4779, 6.1237, 6.0522], + device='cuda:0'), covar=tensor([0.1103, 0.0782, 0.0393, 0.0682, 0.1388, 0.0694, 0.0568, 0.0724], + device='cuda:0'), in_proj_covar=tensor([0.0650, 0.0572, 0.0409, 0.0597, 0.0793, 0.0594, 0.0814, 0.0624], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 23:27:18,860 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=102691.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:27:21,280 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7554, 4.7778, 4.8640, 4.5554, 4.5870, 4.6373, 4.8926, 4.9202], + device='cuda:0'), covar=tensor([0.0087, 0.0091, 0.0071, 0.0144, 0.0080, 0.0155, 0.0091, 0.0102], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0075, 0.0081, 0.0101, 0.0080, 0.0109, 0.0092, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 23:27:32,893 INFO [train.py:898] (0/4) Epoch 29, batch 950, loss[loss=0.1263, simple_loss=0.2077, pruned_loss=0.02239, over 18404.00 frames. ], tot_loss[loss=0.1549, simple_loss=0.2461, pruned_loss=0.03186, over 3557796.71 frames. ], batch size: 42, lr: 3.88e-03, grad_scale: 8.0 +2023-03-09 23:27:43,049 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.728e+02 2.495e+02 2.903e+02 3.483e+02 7.298e+02, threshold=5.805e+02, percent-clipped=1.0 +2023-03-09 23:28:24,439 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.87 vs. limit=5.0 +2023-03-09 23:28:31,549 INFO [train.py:898] (0/4) Epoch 29, batch 1000, loss[loss=0.1641, simple_loss=0.2624, pruned_loss=0.03291, over 18459.00 frames. ], tot_loss[loss=0.1547, simple_loss=0.2458, pruned_loss=0.03177, over 3563083.49 frames. ], batch size: 59, lr: 3.88e-03, grad_scale: 8.0 +2023-03-09 23:28:48,386 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9061, 4.2145, 4.1311, 4.2245, 3.7967, 4.1201, 3.8270, 4.1262], + device='cuda:0'), covar=tensor([0.0279, 0.0313, 0.0263, 0.0517, 0.0360, 0.0262, 0.0815, 0.0366], + device='cuda:0'), in_proj_covar=tensor([0.0239, 0.0283, 0.0282, 0.0363, 0.0292, 0.0292, 0.0322, 0.0284], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 23:28:56,597 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-09 23:29:00,179 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5007, 2.3470, 2.5088, 2.5639, 3.0322, 4.7253, 4.6412, 3.5299], + device='cuda:0'), covar=tensor([0.2237, 0.2826, 0.3293, 0.2166, 0.2764, 0.0304, 0.0391, 0.0946], + device='cuda:0'), in_proj_covar=tensor([0.0333, 0.0366, 0.0417, 0.0294, 0.0398, 0.0271, 0.0305, 0.0275], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0002], + device='cuda:0') +2023-03-09 23:29:04,566 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6141, 3.9904, 5.1811, 4.3563, 3.0692, 2.6661, 4.5960, 5.4347], + device='cuda:0'), covar=tensor([0.0850, 0.1287, 0.0182, 0.0439, 0.1176, 0.1441, 0.0370, 0.0177], + device='cuda:0'), in_proj_covar=tensor([0.0156, 0.0287, 0.0178, 0.0189, 0.0201, 0.0200, 0.0204, 0.0218], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 23:29:11,648 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-09 23:29:31,336 INFO [train.py:898] (0/4) Epoch 29, batch 1050, loss[loss=0.1453, simple_loss=0.2411, pruned_loss=0.02473, over 18320.00 frames. ], tot_loss[loss=0.155, simple_loss=0.2463, pruned_loss=0.03184, over 3568016.85 frames. ], batch size: 54, lr: 3.88e-03, grad_scale: 8.0 +2023-03-09 23:29:34,132 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0272, 3.9066, 5.3878, 3.1233, 4.7737, 2.7702, 3.2229, 1.9347], + device='cuda:0'), covar=tensor([0.1153, 0.0900, 0.0108, 0.0921, 0.0412, 0.2597, 0.2659, 0.2217], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0256, 0.0233, 0.0211, 0.0267, 0.0283, 0.0341, 0.0250], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 23:29:42,516 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.563e+02 2.351e+02 2.635e+02 3.103e+02 8.908e+02, threshold=5.271e+02, percent-clipped=2.0 +2023-03-09 23:30:27,400 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=102851.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:30:29,919 INFO [train.py:898] (0/4) Epoch 29, batch 1100, loss[loss=0.1627, simple_loss=0.2538, pruned_loss=0.03583, over 18360.00 frames. ], tot_loss[loss=0.1549, simple_loss=0.2462, pruned_loss=0.03185, over 3570775.45 frames. ], batch size: 56, lr: 3.88e-03, grad_scale: 8.0 +2023-03-09 23:31:10,249 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=102887.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:31:28,099 INFO [train.py:898] (0/4) Epoch 29, batch 1150, loss[loss=0.1686, simple_loss=0.2555, pruned_loss=0.04088, over 17981.00 frames. ], tot_loss[loss=0.1552, simple_loss=0.2465, pruned_loss=0.0319, over 3578131.85 frames. ], batch size: 65, lr: 3.88e-03, grad_scale: 8.0 +2023-03-09 23:31:32,383 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0251, 4.9968, 5.0806, 4.8529, 4.8260, 4.9479, 5.1669, 5.1768], + device='cuda:0'), covar=tensor([0.0078, 0.0067, 0.0067, 0.0114, 0.0061, 0.0137, 0.0072, 0.0087], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0074, 0.0080, 0.0099, 0.0079, 0.0108, 0.0091, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 23:31:38,790 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.637e+02 2.444e+02 3.032e+02 3.732e+02 6.439e+02, threshold=6.064e+02, percent-clipped=4.0 +2023-03-09 23:31:39,210 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=102912.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:32:22,234 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=102948.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:32:27,615 INFO [train.py:898] (0/4) Epoch 29, batch 1200, loss[loss=0.1457, simple_loss=0.2366, pruned_loss=0.02743, over 18621.00 frames. ], tot_loss[loss=0.1553, simple_loss=0.2465, pruned_loss=0.03204, over 3576004.78 frames. ], batch size: 52, lr: 3.88e-03, grad_scale: 8.0 +2023-03-09 23:32:57,417 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7545, 3.6881, 4.9916, 4.4150, 3.3797, 3.0264, 4.5292, 5.1973], + device='cuda:0'), covar=tensor([0.0811, 0.1428, 0.0234, 0.0392, 0.0933, 0.1241, 0.0395, 0.0324], + device='cuda:0'), in_proj_covar=tensor([0.0153, 0.0282, 0.0175, 0.0186, 0.0197, 0.0196, 0.0201, 0.0214], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 23:33:17,627 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8287, 4.9151, 4.8850, 4.6074, 4.6870, 4.6742, 4.9856, 4.9962], + device='cuda:0'), covar=tensor([0.0080, 0.0056, 0.0057, 0.0121, 0.0058, 0.0166, 0.0063, 0.0073], + device='cuda:0'), in_proj_covar=tensor([0.0100, 0.0074, 0.0080, 0.0100, 0.0079, 0.0108, 0.0091, 0.0091], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-09 23:33:26,584 INFO [train.py:898] (0/4) Epoch 29, batch 1250, loss[loss=0.1786, simple_loss=0.2734, pruned_loss=0.04192, over 18085.00 frames. ], tot_loss[loss=0.1554, simple_loss=0.2463, pruned_loss=0.0323, over 3565180.78 frames. ], batch size: 62, lr: 3.88e-03, grad_scale: 8.0 +2023-03-09 23:33:37,003 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.897e+02 2.667e+02 2.982e+02 3.378e+02 7.201e+02, threshold=5.965e+02, percent-clipped=1.0 +2023-03-09 23:33:43,272 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0953, 5.5442, 5.5272, 5.5363, 4.9706, 5.4410, 4.9240, 5.4469], + device='cuda:0'), covar=tensor([0.0213, 0.0241, 0.0166, 0.0370, 0.0399, 0.0205, 0.0970, 0.0311], + device='cuda:0'), in_proj_covar=tensor([0.0236, 0.0281, 0.0280, 0.0360, 0.0290, 0.0291, 0.0319, 0.0281], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 23:34:22,118 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5214, 5.4860, 5.1572, 5.4165, 5.4826, 4.8365, 5.3475, 5.0698], + device='cuda:0'), covar=tensor([0.0415, 0.0415, 0.1175, 0.0798, 0.0516, 0.0398, 0.0402, 0.1057], + device='cuda:0'), in_proj_covar=tensor([0.0524, 0.0593, 0.0733, 0.0458, 0.0482, 0.0538, 0.0571, 0.0713], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-09 23:34:25,434 INFO [train.py:898] (0/4) Epoch 29, batch 1300, loss[loss=0.1411, simple_loss=0.2354, pruned_loss=0.02336, over 18370.00 frames. ], tot_loss[loss=0.155, simple_loss=0.2459, pruned_loss=0.03209, over 3563842.51 frames. ], batch size: 55, lr: 3.88e-03, grad_scale: 8.0 +2023-03-09 23:34:35,171 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5762, 3.5675, 3.3858, 3.0515, 3.3225, 2.8056, 2.7487, 3.5255], + device='cuda:0'), covar=tensor([0.0077, 0.0106, 0.0095, 0.0167, 0.0130, 0.0217, 0.0228, 0.0091], + device='cuda:0'), in_proj_covar=tensor([0.0160, 0.0180, 0.0151, 0.0201, 0.0160, 0.0192, 0.0196, 0.0140], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 23:35:19,351 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103098.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:35:24,724 INFO [train.py:898] (0/4) Epoch 29, batch 1350, loss[loss=0.1579, simple_loss=0.2502, pruned_loss=0.03275, over 18206.00 frames. ], tot_loss[loss=0.1551, simple_loss=0.246, pruned_loss=0.03215, over 3568532.77 frames. ], batch size: 60, lr: 3.88e-03, grad_scale: 8.0 +2023-03-09 23:35:35,180 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.567e+02 2.524e+02 2.935e+02 3.557e+02 7.291e+02, threshold=5.869e+02, percent-clipped=2.0 +2023-03-09 23:35:58,986 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103132.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:36:11,518 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0976, 5.1908, 5.3876, 5.3979, 5.0993, 5.9371, 5.6220, 5.2475], + device='cuda:0'), covar=tensor([0.1155, 0.0682, 0.0796, 0.0821, 0.1399, 0.0733, 0.0662, 0.1658], + device='cuda:0'), in_proj_covar=tensor([0.0381, 0.0310, 0.0337, 0.0339, 0.0347, 0.0450, 0.0305, 0.0447], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 23:36:23,953 INFO [train.py:898] (0/4) Epoch 29, batch 1400, loss[loss=0.1533, simple_loss=0.2446, pruned_loss=0.03102, over 18279.00 frames. ], tot_loss[loss=0.1557, simple_loss=0.2468, pruned_loss=0.03236, over 3576678.89 frames. ], batch size: 49, lr: 3.87e-03, grad_scale: 8.0 +2023-03-09 23:36:31,061 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103159.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:37:11,171 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103193.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:37:22,824 INFO [train.py:898] (0/4) Epoch 29, batch 1450, loss[loss=0.1611, simple_loss=0.2551, pruned_loss=0.03357, over 18236.00 frames. ], tot_loss[loss=0.1562, simple_loss=0.2472, pruned_loss=0.03262, over 3566715.59 frames. ], batch size: 60, lr: 3.87e-03, grad_scale: 8.0 +2023-03-09 23:37:27,604 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103207.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:37:33,042 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.757e+02 2.357e+02 2.757e+02 3.131e+02 6.312e+02, threshold=5.514e+02, percent-clipped=1.0 +2023-03-09 23:37:39,472 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.66 vs. limit=2.0 +2023-03-09 23:38:09,451 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103243.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:38:21,767 INFO [train.py:898] (0/4) Epoch 29, batch 1500, loss[loss=0.1256, simple_loss=0.2071, pruned_loss=0.02204, over 18413.00 frames. ], tot_loss[loss=0.1567, simple_loss=0.2479, pruned_loss=0.03275, over 3570693.33 frames. ], batch size: 42, lr: 3.87e-03, grad_scale: 8.0 +2023-03-09 23:38:34,431 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2149, 5.7283, 5.3577, 5.5032, 5.3366, 5.1503, 5.8050, 5.7338], + device='cuda:0'), covar=tensor([0.1223, 0.0776, 0.0597, 0.0727, 0.1290, 0.0737, 0.0557, 0.0740], + device='cuda:0'), in_proj_covar=tensor([0.0645, 0.0569, 0.0407, 0.0594, 0.0791, 0.0592, 0.0809, 0.0622], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 23:38:34,643 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6228, 3.4966, 4.7494, 4.1409, 3.1459, 2.8601, 4.2349, 4.9238], + device='cuda:0'), covar=tensor([0.0836, 0.1468, 0.0228, 0.0455, 0.1023, 0.1334, 0.0436, 0.0291], + device='cuda:0'), in_proj_covar=tensor([0.0155, 0.0286, 0.0176, 0.0189, 0.0199, 0.0199, 0.0203, 0.0216], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 23:39:11,294 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0497, 4.5191, 4.2662, 4.2514, 4.0765, 4.7889, 4.5039, 4.1662], + device='cuda:0'), covar=tensor([0.1449, 0.1235, 0.1062, 0.1119, 0.1505, 0.1128, 0.0860, 0.1643], + device='cuda:0'), in_proj_covar=tensor([0.0380, 0.0308, 0.0336, 0.0338, 0.0346, 0.0449, 0.0305, 0.0445], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 23:39:20,021 INFO [train.py:898] (0/4) Epoch 29, batch 1550, loss[loss=0.1632, simple_loss=0.2518, pruned_loss=0.03731, over 18477.00 frames. ], tot_loss[loss=0.1561, simple_loss=0.2473, pruned_loss=0.03244, over 3564706.63 frames. ], batch size: 59, lr: 3.87e-03, grad_scale: 8.0 +2023-03-09 23:39:29,920 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.405e+02 2.624e+02 2.952e+02 3.606e+02 8.116e+02, threshold=5.905e+02, percent-clipped=6.0 +2023-03-09 23:39:36,206 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.17 vs. limit=2.0 +2023-03-09 23:40:18,838 INFO [train.py:898] (0/4) Epoch 29, batch 1600, loss[loss=0.1357, simple_loss=0.2193, pruned_loss=0.02607, over 18240.00 frames. ], tot_loss[loss=0.1559, simple_loss=0.2469, pruned_loss=0.03239, over 3566228.76 frames. ], batch size: 45, lr: 3.87e-03, grad_scale: 8.0 +2023-03-09 23:40:34,702 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103366.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 23:41:18,050 INFO [train.py:898] (0/4) Epoch 29, batch 1650, loss[loss=0.1495, simple_loss=0.2451, pruned_loss=0.02695, over 18417.00 frames. ], tot_loss[loss=0.1547, simple_loss=0.2457, pruned_loss=0.03183, over 3579846.69 frames. ], batch size: 48, lr: 3.87e-03, grad_scale: 8.0 +2023-03-09 23:41:29,744 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.750e+02 2.322e+02 2.729e+02 3.382e+02 5.618e+02, threshold=5.459e+02, percent-clipped=0.0 +2023-03-09 23:41:47,130 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103427.0, num_to_drop=1, layers_to_drop={3} +2023-03-09 23:42:17,236 INFO [train.py:898] (0/4) Epoch 29, batch 1700, loss[loss=0.1533, simple_loss=0.2465, pruned_loss=0.02999, over 18132.00 frames. ], tot_loss[loss=0.1549, simple_loss=0.2458, pruned_loss=0.03205, over 3570273.86 frames. ], batch size: 62, lr: 3.87e-03, grad_scale: 8.0 +2023-03-09 23:42:18,599 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103454.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:42:58,834 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103488.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:43:16,095 INFO [train.py:898] (0/4) Epoch 29, batch 1750, loss[loss=0.151, simple_loss=0.2399, pruned_loss=0.03103, over 18364.00 frames. ], tot_loss[loss=0.155, simple_loss=0.2461, pruned_loss=0.03198, over 3585570.35 frames. ], batch size: 46, lr: 3.87e-03, grad_scale: 8.0 +2023-03-09 23:43:20,983 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103507.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:43:27,006 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.672e+02 2.484e+02 2.930e+02 3.685e+02 5.975e+02, threshold=5.860e+02, percent-clipped=1.0 +2023-03-09 23:43:27,619 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8332, 2.4322, 2.7055, 2.8456, 3.1802, 4.9189, 4.9901, 3.3355], + device='cuda:0'), covar=tensor([0.2066, 0.2705, 0.3289, 0.1956, 0.2746, 0.0273, 0.0304, 0.1149], + device='cuda:0'), in_proj_covar=tensor([0.0335, 0.0368, 0.0420, 0.0294, 0.0400, 0.0270, 0.0305, 0.0276], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0002], + device='cuda:0') +2023-03-09 23:44:04,081 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103543.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:44:16,109 INFO [train.py:898] (0/4) Epoch 29, batch 1800, loss[loss=0.1696, simple_loss=0.2695, pruned_loss=0.0349, over 16183.00 frames. ], tot_loss[loss=0.1551, simple_loss=0.2461, pruned_loss=0.03202, over 3584951.43 frames. ], batch size: 95, lr: 3.87e-03, grad_scale: 8.0 +2023-03-09 23:44:18,471 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=103555.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:44:33,212 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8625, 3.5691, 5.0428, 2.9283, 4.3495, 2.5863, 3.1285, 1.7386], + device='cuda:0'), covar=tensor([0.1247, 0.1054, 0.0169, 0.1055, 0.0544, 0.2564, 0.2657, 0.2411], + device='cuda:0'), in_proj_covar=tensor([0.0232, 0.0253, 0.0232, 0.0208, 0.0265, 0.0280, 0.0336, 0.0246], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-09 23:45:00,579 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=103591.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:45:15,441 INFO [train.py:898] (0/4) Epoch 29, batch 1850, loss[loss=0.1488, simple_loss=0.239, pruned_loss=0.02926, over 18249.00 frames. ], tot_loss[loss=0.1545, simple_loss=0.2454, pruned_loss=0.03176, over 3591546.08 frames. ], batch size: 45, lr: 3.87e-03, grad_scale: 8.0 +2023-03-09 23:45:25,551 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.856e+02 2.569e+02 2.972e+02 3.705e+02 6.711e+02, threshold=5.945e+02, percent-clipped=3.0 +2023-03-09 23:46:14,390 INFO [train.py:898] (0/4) Epoch 29, batch 1900, loss[loss=0.159, simple_loss=0.2551, pruned_loss=0.03149, over 18619.00 frames. ], tot_loss[loss=0.1542, simple_loss=0.2453, pruned_loss=0.03154, over 3595129.21 frames. ], batch size: 52, lr: 3.87e-03, grad_scale: 8.0 +2023-03-09 23:47:14,032 INFO [train.py:898] (0/4) Epoch 29, batch 1950, loss[loss=0.1659, simple_loss=0.2586, pruned_loss=0.03656, over 18291.00 frames. ], tot_loss[loss=0.1554, simple_loss=0.2463, pruned_loss=0.03218, over 3580063.32 frames. ], batch size: 49, lr: 3.86e-03, grad_scale: 8.0 +2023-03-09 23:47:24,463 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.577e+02 2.406e+02 2.960e+02 3.499e+02 1.191e+03, threshold=5.920e+02, percent-clipped=3.0 +2023-03-09 23:47:36,603 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=103722.0, num_to_drop=1, layers_to_drop={0} +2023-03-09 23:47:57,813 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103740.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:48:13,383 INFO [train.py:898] (0/4) Epoch 29, batch 2000, loss[loss=0.1433, simple_loss=0.2255, pruned_loss=0.03057, over 18404.00 frames. ], tot_loss[loss=0.1548, simple_loss=0.2458, pruned_loss=0.03191, over 3576335.14 frames. ], batch size: 42, lr: 3.86e-03, grad_scale: 8.0 +2023-03-09 23:48:14,871 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103754.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:48:27,193 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103765.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:48:40,258 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.5962, 3.9172, 2.5738, 3.8531, 4.8864, 2.5837, 3.7253, 3.8262], + device='cuda:0'), covar=tensor([0.0254, 0.1110, 0.1560, 0.0676, 0.0120, 0.1213, 0.0687, 0.0740], + device='cuda:0'), in_proj_covar=tensor([0.0188, 0.0286, 0.0212, 0.0203, 0.0147, 0.0188, 0.0225, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 23:48:54,495 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=103788.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:49:09,624 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103801.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:49:11,183 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=103802.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:49:12,645 INFO [train.py:898] (0/4) Epoch 29, batch 2050, loss[loss=0.1441, simple_loss=0.2285, pruned_loss=0.02989, over 18246.00 frames. ], tot_loss[loss=0.155, simple_loss=0.2457, pruned_loss=0.03216, over 3582022.80 frames. ], batch size: 45, lr: 3.86e-03, grad_scale: 4.0 +2023-03-09 23:49:23,960 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.828e+02 2.583e+02 3.001e+02 3.513e+02 6.272e+02, threshold=6.003e+02, percent-clipped=1.0 +2023-03-09 23:49:39,232 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103826.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:49:40,248 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7269, 5.2611, 5.2451, 5.3067, 4.6132, 5.1660, 4.5513, 5.1209], + device='cuda:0'), covar=tensor([0.0299, 0.0347, 0.0238, 0.0417, 0.0447, 0.0269, 0.1225, 0.0400], + device='cuda:0'), in_proj_covar=tensor([0.0236, 0.0282, 0.0281, 0.0361, 0.0290, 0.0290, 0.0320, 0.0281], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 23:49:49,988 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.3334, 4.7358, 4.6986, 4.7472, 4.2194, 4.6523, 4.2054, 4.6203], + device='cuda:0'), covar=tensor([0.0270, 0.0309, 0.0229, 0.0500, 0.0393, 0.0238, 0.0953, 0.0381], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0282, 0.0281, 0.0360, 0.0289, 0.0289, 0.0319, 0.0281], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 23:49:50,928 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=103836.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:50:10,329 INFO [train.py:898] (0/4) Epoch 29, batch 2100, loss[loss=0.1453, simple_loss=0.2376, pruned_loss=0.02647, over 18278.00 frames. ], tot_loss[loss=0.1549, simple_loss=0.2456, pruned_loss=0.03212, over 3581565.39 frames. ], batch size: 49, lr: 3.86e-03, grad_scale: 4.0 +2023-03-09 23:50:27,314 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7990, 3.5915, 4.9122, 4.3125, 3.5036, 2.9164, 4.4281, 5.1379], + device='cuda:0'), covar=tensor([0.0780, 0.1554, 0.0264, 0.0421, 0.0871, 0.1259, 0.0387, 0.0218], + device='cuda:0'), in_proj_covar=tensor([0.0156, 0.0287, 0.0178, 0.0190, 0.0200, 0.0199, 0.0204, 0.0218], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-09 23:50:38,490 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103876.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:51:09,168 INFO [train.py:898] (0/4) Epoch 29, batch 2150, loss[loss=0.1871, simple_loss=0.2711, pruned_loss=0.05153, over 12684.00 frames. ], tot_loss[loss=0.1561, simple_loss=0.2471, pruned_loss=0.03253, over 3577430.69 frames. ], batch size: 129, lr: 3.86e-03, grad_scale: 4.0 +2023-03-09 23:51:21,586 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.841e+02 2.540e+02 2.937e+02 3.473e+02 7.203e+02, threshold=5.874e+02, percent-clipped=1.0 +2023-03-09 23:51:50,024 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=103937.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 23:52:07,793 INFO [train.py:898] (0/4) Epoch 29, batch 2200, loss[loss=0.1503, simple_loss=0.2411, pruned_loss=0.02977, over 17972.00 frames. ], tot_loss[loss=0.1563, simple_loss=0.2475, pruned_loss=0.03257, over 3572265.41 frames. ], batch size: 65, lr: 3.86e-03, grad_scale: 4.0 +2023-03-09 23:52:52,491 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103990.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:53:00,501 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=103997.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:53:04,068 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-104000.pt +2023-03-09 23:53:11,844 INFO [train.py:898] (0/4) Epoch 29, batch 2250, loss[loss=0.1647, simple_loss=0.2524, pruned_loss=0.03848, over 18270.00 frames. ], tot_loss[loss=0.1561, simple_loss=0.2469, pruned_loss=0.03264, over 3575376.34 frames. ], batch size: 49, lr: 3.86e-03, grad_scale: 4.0 +2023-03-09 23:53:19,158 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6052, 2.8896, 2.6409, 2.9145, 3.7021, 3.6446, 3.1568, 2.9786], + device='cuda:0'), covar=tensor([0.0216, 0.0329, 0.0593, 0.0398, 0.0200, 0.0174, 0.0444, 0.0417], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0150, 0.0169, 0.0168, 0.0145, 0.0132, 0.0164, 0.0168], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 23:53:23,833 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.696e+02 2.605e+02 3.021e+02 3.898e+02 6.777e+02, threshold=6.041e+02, percent-clipped=7.0 +2023-03-09 23:53:35,019 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104022.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 23:53:47,724 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.34 vs. limit=2.0 +2023-03-09 23:53:58,957 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.2425, 2.6964, 2.4504, 2.7059, 3.4264, 3.2798, 2.9500, 2.7639], + device='cuda:0'), covar=tensor([0.0214, 0.0317, 0.0560, 0.0407, 0.0214, 0.0230, 0.0439, 0.0410], + device='cuda:0'), in_proj_covar=tensor([0.0149, 0.0149, 0.0168, 0.0168, 0.0145, 0.0132, 0.0164, 0.0167], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 23:54:06,847 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=104049.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:54:09,194 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=104051.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:54:11,015 INFO [train.py:898] (0/4) Epoch 29, batch 2300, loss[loss=0.1404, simple_loss=0.2324, pruned_loss=0.02426, over 18415.00 frames. ], tot_loss[loss=0.1554, simple_loss=0.2461, pruned_loss=0.03234, over 3583988.84 frames. ], batch size: 48, lr: 3.86e-03, grad_scale: 4.0 +2023-03-09 23:54:17,157 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=104058.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:54:31,614 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=104070.0, num_to_drop=1, layers_to_drop={1} +2023-03-09 23:54:53,731 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3728, 5.9344, 5.4781, 5.7028, 5.5440, 5.3194, 5.9618, 5.9201], + device='cuda:0'), covar=tensor([0.1237, 0.0789, 0.0578, 0.0755, 0.1352, 0.0771, 0.0649, 0.0713], + device='cuda:0'), in_proj_covar=tensor([0.0650, 0.0571, 0.0411, 0.0595, 0.0795, 0.0589, 0.0812, 0.0622], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-09 23:55:02,149 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104096.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:55:09,808 INFO [train.py:898] (0/4) Epoch 29, batch 2350, loss[loss=0.1403, simple_loss=0.2282, pruned_loss=0.02619, over 18487.00 frames. ], tot_loss[loss=0.1549, simple_loss=0.2454, pruned_loss=0.03218, over 3576683.44 frames. ], batch size: 47, lr: 3.86e-03, grad_scale: 4.0 +2023-03-09 23:55:17,810 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=104110.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:55:20,810 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.751e+02 2.516e+02 3.034e+02 3.575e+02 8.102e+02, threshold=6.068e+02, percent-clipped=2.0 +2023-03-09 23:55:28,615 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5751, 4.1069, 2.5669, 3.9242, 3.9195, 4.0772, 3.9635, 2.6433], + device='cuda:0'), covar=tensor([0.0289, 0.0122, 0.0890, 0.0238, 0.0122, 0.0132, 0.0145, 0.1088], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0084, 0.0098, 0.0100, 0.0091, 0.0080, 0.0087, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-09 23:55:30,716 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104121.0, num_to_drop=0, layers_to_drop=set() +2023-03-09 23:56:08,415 INFO [train.py:898] (0/4) Epoch 29, batch 2400, loss[loss=0.1609, simple_loss=0.2567, pruned_loss=0.03252, over 18625.00 frames. ], tot_loss[loss=0.1544, simple_loss=0.2449, pruned_loss=0.03196, over 3587245.91 frames. ], batch size: 52, lr: 3.86e-03, grad_scale: 8.0 +2023-03-09 23:56:47,138 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.4705, 4.9498, 4.9260, 4.9390, 4.4139, 4.8492, 4.4003, 4.8411], + device='cuda:0'), covar=tensor([0.0297, 0.0266, 0.0208, 0.0475, 0.0389, 0.0247, 0.0919, 0.0320], + device='cuda:0'), in_proj_covar=tensor([0.0233, 0.0277, 0.0279, 0.0357, 0.0286, 0.0287, 0.0317, 0.0277], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 23:57:07,528 INFO [train.py:898] (0/4) Epoch 29, batch 2450, loss[loss=0.1471, simple_loss=0.2426, pruned_loss=0.02583, over 18372.00 frames. ], tot_loss[loss=0.1542, simple_loss=0.2447, pruned_loss=0.03187, over 3594299.49 frames. ], batch size: 50, lr: 3.86e-03, grad_scale: 8.0 +2023-03-09 23:57:18,506 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.747e+02 2.425e+02 2.919e+02 3.520e+02 8.607e+02, threshold=5.838e+02, percent-clipped=3.0 +2023-03-09 23:57:41,185 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104232.0, num_to_drop=1, layers_to_drop={2} +2023-03-09 23:58:05,929 INFO [train.py:898] (0/4) Epoch 29, batch 2500, loss[loss=0.1475, simple_loss=0.2386, pruned_loss=0.02821, over 18282.00 frames. ], tot_loss[loss=0.1549, simple_loss=0.2458, pruned_loss=0.03197, over 3590181.46 frames. ], batch size: 49, lr: 3.85e-03, grad_scale: 8.0 +2023-03-09 23:58:06,254 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9781, 5.4639, 5.4712, 5.4861, 4.8429, 5.3905, 4.8279, 5.3637], + device='cuda:0'), covar=tensor([0.0232, 0.0270, 0.0184, 0.0426, 0.0412, 0.0227, 0.1019, 0.0318], + device='cuda:0'), in_proj_covar=tensor([0.0234, 0.0278, 0.0280, 0.0358, 0.0287, 0.0288, 0.0318, 0.0278], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-09 23:59:04,430 INFO [train.py:898] (0/4) Epoch 29, batch 2550, loss[loss=0.1693, simple_loss=0.2641, pruned_loss=0.03728, over 18147.00 frames. ], tot_loss[loss=0.1546, simple_loss=0.2455, pruned_loss=0.03187, over 3583442.52 frames. ], batch size: 62, lr: 3.85e-03, grad_scale: 8.0 +2023-03-09 23:59:16,569 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.724e+02 2.479e+02 2.941e+02 3.574e+02 6.228e+02, threshold=5.882e+02, percent-clipped=1.0 +2023-03-09 23:59:38,113 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7537, 2.9582, 2.8438, 2.9834, 3.7582, 3.7708, 3.2911, 2.9788], + device='cuda:0'), covar=tensor([0.0166, 0.0313, 0.0495, 0.0432, 0.0184, 0.0149, 0.0363, 0.0429], + device='cuda:0'), in_proj_covar=tensor([0.0149, 0.0149, 0.0168, 0.0168, 0.0145, 0.0132, 0.0164, 0.0167], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0003, 0.0004], + device='cuda:0') +2023-03-09 23:59:49,518 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.54 vs. limit=5.0 +2023-03-09 23:59:55,466 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104346.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:00:03,156 INFO [train.py:898] (0/4) Epoch 29, batch 2600, loss[loss=0.1741, simple_loss=0.2638, pruned_loss=0.0422, over 18383.00 frames. ], tot_loss[loss=0.1553, simple_loss=0.2461, pruned_loss=0.03222, over 3568517.49 frames. ], batch size: 50, lr: 3.85e-03, grad_scale: 8.0 +2023-03-10 00:00:03,376 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104353.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:00:11,610 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.29 vs. limit=2.0 +2023-03-10 00:00:21,709 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-10 00:00:54,627 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104396.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:01:03,255 INFO [train.py:898] (0/4) Epoch 29, batch 2650, loss[loss=0.1879, simple_loss=0.2722, pruned_loss=0.05181, over 13022.00 frames. ], tot_loss[loss=0.1549, simple_loss=0.2458, pruned_loss=0.03199, over 3566510.42 frames. ], batch size: 130, lr: 3.85e-03, grad_scale: 8.0 +2023-03-10 00:01:05,749 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=104405.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:01:15,456 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.642e+02 2.433e+02 2.852e+02 3.546e+02 8.268e+02, threshold=5.705e+02, percent-clipped=2.0 +2023-03-10 00:01:24,957 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104421.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:01:25,200 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6857, 3.6774, 4.9819, 4.3710, 3.2478, 2.9789, 4.4625, 5.2755], + device='cuda:0'), covar=tensor([0.0868, 0.1421, 0.0200, 0.0405, 0.1030, 0.1271, 0.0402, 0.0157], + device='cuda:0'), in_proj_covar=tensor([0.0155, 0.0284, 0.0177, 0.0188, 0.0198, 0.0197, 0.0202, 0.0215], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-10 00:01:45,945 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2918, 5.7876, 5.4192, 5.6137, 5.4417, 5.2914, 5.8342, 5.7956], + device='cuda:0'), covar=tensor([0.1209, 0.0743, 0.0532, 0.0707, 0.1235, 0.0665, 0.0588, 0.0669], + device='cuda:0'), in_proj_covar=tensor([0.0646, 0.0570, 0.0410, 0.0596, 0.0790, 0.0591, 0.0813, 0.0624], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-10 00:01:50,698 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.1766, 3.8346, 5.3790, 3.0000, 4.7378, 2.8154, 3.1982, 1.8991], + device='cuda:0'), covar=tensor([0.1094, 0.0961, 0.0160, 0.0998, 0.0426, 0.2580, 0.2930, 0.2294], + device='cuda:0'), in_proj_covar=tensor([0.0234, 0.0256, 0.0233, 0.0210, 0.0267, 0.0284, 0.0339, 0.0248], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-10 00:01:52,716 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=104444.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:02:03,367 INFO [train.py:898] (0/4) Epoch 29, batch 2700, loss[loss=0.1353, simple_loss=0.2258, pruned_loss=0.02237, over 18576.00 frames. ], tot_loss[loss=0.1546, simple_loss=0.2456, pruned_loss=0.03185, over 3576823.91 frames. ], batch size: 49, lr: 3.85e-03, grad_scale: 8.0 +2023-03-10 00:02:22,276 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=104469.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:02:31,157 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6792, 3.6694, 3.5423, 3.1305, 3.4027, 2.9079, 2.7302, 3.8018], + device='cuda:0'), covar=tensor([0.0076, 0.0090, 0.0089, 0.0153, 0.0118, 0.0210, 0.0240, 0.0063], + device='cuda:0'), in_proj_covar=tensor([0.0158, 0.0177, 0.0147, 0.0196, 0.0157, 0.0187, 0.0192, 0.0137], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-10 00:03:01,346 INFO [train.py:898] (0/4) Epoch 29, batch 2750, loss[loss=0.1635, simple_loss=0.2598, pruned_loss=0.03364, over 17750.00 frames. ], tot_loss[loss=0.1548, simple_loss=0.2458, pruned_loss=0.03192, over 3572694.62 frames. ], batch size: 70, lr: 3.85e-03, grad_scale: 8.0 +2023-03-10 00:03:12,786 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.672e+02 2.565e+02 3.025e+02 3.690e+02 7.436e+02, threshold=6.050e+02, percent-clipped=2.0 +2023-03-10 00:03:35,766 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104532.0, num_to_drop=1, layers_to_drop={2} +2023-03-10 00:03:59,994 INFO [train.py:898] (0/4) Epoch 29, batch 2800, loss[loss=0.145, simple_loss=0.2413, pruned_loss=0.02433, over 18406.00 frames. ], tot_loss[loss=0.1549, simple_loss=0.246, pruned_loss=0.03186, over 3571491.71 frames. ], batch size: 52, lr: 3.85e-03, grad_scale: 8.0 +2023-03-10 00:04:31,633 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=104580.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:04:58,831 INFO [train.py:898] (0/4) Epoch 29, batch 2850, loss[loss=0.1543, simple_loss=0.2551, pruned_loss=0.02675, over 18400.00 frames. ], tot_loss[loss=0.1553, simple_loss=0.2465, pruned_loss=0.03207, over 3561259.67 frames. ], batch size: 52, lr: 3.85e-03, grad_scale: 8.0 +2023-03-10 00:05:10,927 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.658e+02 2.500e+02 2.877e+02 3.433e+02 5.313e+02, threshold=5.755e+02, percent-clipped=0.0 +2023-03-10 00:05:50,642 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104646.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:05:58,418 INFO [train.py:898] (0/4) Epoch 29, batch 2900, loss[loss=0.1389, simple_loss=0.2223, pruned_loss=0.02774, over 17305.00 frames. ], tot_loss[loss=0.1552, simple_loss=0.2463, pruned_loss=0.03205, over 3565456.69 frames. ], batch size: 38, lr: 3.85e-03, grad_scale: 8.0 +2023-03-10 00:05:58,782 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104653.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:06:14,909 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.91 vs. limit=5.0 +2023-03-10 00:06:41,736 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.11 vs. limit=5.0 +2023-03-10 00:06:47,694 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=104694.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:06:56,053 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=104701.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:06:58,047 INFO [train.py:898] (0/4) Epoch 29, batch 2950, loss[loss=0.1504, simple_loss=0.2402, pruned_loss=0.03028, over 18483.00 frames. ], tot_loss[loss=0.1552, simple_loss=0.2467, pruned_loss=0.03186, over 3574428.98 frames. ], batch size: 51, lr: 3.85e-03, grad_scale: 8.0 +2023-03-10 00:07:00,737 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=104705.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:07:09,441 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.771e+02 2.535e+02 2.972e+02 3.582e+02 1.059e+03, threshold=5.945e+02, percent-clipped=4.0 +2023-03-10 00:07:57,844 INFO [train.py:898] (0/4) Epoch 29, batch 3000, loss[loss=0.1706, simple_loss=0.2603, pruned_loss=0.04051, over 12947.00 frames. ], tot_loss[loss=0.1549, simple_loss=0.2462, pruned_loss=0.03183, over 3567663.91 frames. ], batch size: 129, lr: 3.85e-03, grad_scale: 8.0 +2023-03-10 00:07:57,846 INFO [train.py:923] (0/4) Computing validation loss +2023-03-10 00:08:10,054 INFO [train.py:932] (0/4) Epoch 29, validation: loss=0.1493, simple_loss=0.2471, pruned_loss=0.02574, over 944034.00 frames. +2023-03-10 00:08:10,055 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-10 00:08:10,335 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=104753.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:09:09,598 INFO [train.py:898] (0/4) Epoch 29, batch 3050, loss[loss=0.1843, simple_loss=0.2877, pruned_loss=0.04051, over 18291.00 frames. ], tot_loss[loss=0.1553, simple_loss=0.2463, pruned_loss=0.03216, over 3573654.04 frames. ], batch size: 57, lr: 3.84e-03, grad_scale: 8.0 +2023-03-10 00:09:21,239 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.749e+02 2.540e+02 3.003e+02 3.449e+02 8.629e+02, threshold=6.007e+02, percent-clipped=3.0 +2023-03-10 00:09:22,186 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.84 vs. limit=5.0 +2023-03-10 00:10:08,509 INFO [train.py:898] (0/4) Epoch 29, batch 3100, loss[loss=0.1347, simple_loss=0.2152, pruned_loss=0.02713, over 18438.00 frames. ], tot_loss[loss=0.1549, simple_loss=0.2459, pruned_loss=0.03193, over 3579884.00 frames. ], batch size: 42, lr: 3.84e-03, grad_scale: 8.0 +2023-03-10 00:11:07,593 INFO [train.py:898] (0/4) Epoch 29, batch 3150, loss[loss=0.1411, simple_loss=0.2339, pruned_loss=0.02417, over 18298.00 frames. ], tot_loss[loss=0.1549, simple_loss=0.2458, pruned_loss=0.03194, over 3573278.37 frames. ], batch size: 47, lr: 3.84e-03, grad_scale: 8.0 +2023-03-10 00:11:19,633 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.665e+02 2.384e+02 2.834e+02 3.322e+02 6.284e+02, threshold=5.669e+02, percent-clipped=1.0 +2023-03-10 00:12:06,345 INFO [train.py:898] (0/4) Epoch 29, batch 3200, loss[loss=0.1393, simple_loss=0.2233, pruned_loss=0.02763, over 18252.00 frames. ], tot_loss[loss=0.1559, simple_loss=0.2469, pruned_loss=0.03249, over 3560482.80 frames. ], batch size: 45, lr: 3.84e-03, grad_scale: 8.0 +2023-03-10 00:12:13,888 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.50 vs. limit=2.0 +2023-03-10 00:12:58,449 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0121, 5.5171, 5.5004, 5.5008, 4.9295, 5.4288, 4.9187, 5.4247], + device='cuda:0'), covar=tensor([0.0272, 0.0255, 0.0184, 0.0378, 0.0416, 0.0227, 0.0955, 0.0284], + device='cuda:0'), in_proj_covar=tensor([0.0236, 0.0280, 0.0281, 0.0361, 0.0290, 0.0290, 0.0321, 0.0281], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-10 00:13:01,936 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8905, 3.6863, 5.0818, 4.4452, 3.5027, 3.0823, 4.5823, 5.3563], + device='cuda:0'), covar=tensor([0.0771, 0.1572, 0.0210, 0.0397, 0.0886, 0.1203, 0.0369, 0.0169], + device='cuda:0'), in_proj_covar=tensor([0.0156, 0.0285, 0.0177, 0.0188, 0.0199, 0.0198, 0.0203, 0.0216], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-10 00:13:06,477 INFO [train.py:898] (0/4) Epoch 29, batch 3250, loss[loss=0.1539, simple_loss=0.2529, pruned_loss=0.02744, over 18315.00 frames. ], tot_loss[loss=0.1555, simple_loss=0.2466, pruned_loss=0.03225, over 3565627.64 frames. ], batch size: 54, lr: 3.84e-03, grad_scale: 8.0 +2023-03-10 00:13:17,605 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.817e+02 2.551e+02 3.060e+02 3.706e+02 7.132e+02, threshold=6.121e+02, percent-clipped=2.0 +2023-03-10 00:13:42,182 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.32 vs. limit=2.0 +2023-03-10 00:14:05,586 INFO [train.py:898] (0/4) Epoch 29, batch 3300, loss[loss=0.1545, simple_loss=0.2497, pruned_loss=0.02969, over 18412.00 frames. ], tot_loss[loss=0.1543, simple_loss=0.245, pruned_loss=0.03176, over 3574028.51 frames. ], batch size: 52, lr: 3.84e-03, grad_scale: 8.0 +2023-03-10 00:14:23,223 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.75 vs. limit=5.0 +2023-03-10 00:14:27,162 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.63 vs. limit=5.0 +2023-03-10 00:14:53,172 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.16 vs. limit=2.0 +2023-03-10 00:15:04,539 INFO [train.py:898] (0/4) Epoch 29, batch 3350, loss[loss=0.1611, simple_loss=0.2522, pruned_loss=0.03494, over 18353.00 frames. ], tot_loss[loss=0.1544, simple_loss=0.2454, pruned_loss=0.03172, over 3585289.39 frames. ], batch size: 56, lr: 3.84e-03, grad_scale: 8.0 +2023-03-10 00:15:15,600 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.933e+02 2.604e+02 2.932e+02 3.657e+02 1.346e+03, threshold=5.864e+02, percent-clipped=3.0 +2023-03-10 00:15:31,560 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.59 vs. limit=5.0 +2023-03-10 00:16:03,643 INFO [train.py:898] (0/4) Epoch 29, batch 3400, loss[loss=0.1765, simple_loss=0.2786, pruned_loss=0.03721, over 17731.00 frames. ], tot_loss[loss=0.1542, simple_loss=0.2451, pruned_loss=0.03169, over 3597229.32 frames. ], batch size: 70, lr: 3.84e-03, grad_scale: 8.0 +2023-03-10 00:16:05,429 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.47 vs. limit=2.0 +2023-03-10 00:17:01,947 INFO [train.py:898] (0/4) Epoch 29, batch 3450, loss[loss=0.1523, simple_loss=0.2487, pruned_loss=0.02798, over 18308.00 frames. ], tot_loss[loss=0.1549, simple_loss=0.2459, pruned_loss=0.03195, over 3591368.84 frames. ], batch size: 54, lr: 3.84e-03, grad_scale: 8.0 +2023-03-10 00:17:13,785 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.580e+02 2.383e+02 2.808e+02 3.318e+02 5.892e+02, threshold=5.615e+02, percent-clipped=1.0 +2023-03-10 00:17:41,266 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.79 vs. limit=5.0 +2023-03-10 00:18:00,564 INFO [train.py:898] (0/4) Epoch 29, batch 3500, loss[loss=0.177, simple_loss=0.2637, pruned_loss=0.04518, over 18473.00 frames. ], tot_loss[loss=0.1556, simple_loss=0.2464, pruned_loss=0.03241, over 3586183.32 frames. ], batch size: 59, lr: 3.84e-03, grad_scale: 8.0 +2023-03-10 00:18:55,798 INFO [train.py:898] (0/4) Epoch 29, batch 3550, loss[loss=0.1449, simple_loss=0.2361, pruned_loss=0.02686, over 18367.00 frames. ], tot_loss[loss=0.156, simple_loss=0.2464, pruned_loss=0.03276, over 3579733.61 frames. ], batch size: 46, lr: 3.83e-03, grad_scale: 8.0 +2023-03-10 00:18:58,108 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1320, 5.1131, 4.7743, 5.0552, 5.0282, 4.4741, 4.9234, 4.6754], + device='cuda:0'), covar=tensor([0.0431, 0.0540, 0.1290, 0.0799, 0.0669, 0.0465, 0.0488, 0.1290], + device='cuda:0'), in_proj_covar=tensor([0.0520, 0.0589, 0.0726, 0.0457, 0.0489, 0.0538, 0.0566, 0.0708], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-10 00:19:06,459 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.798e+02 2.434e+02 2.865e+02 3.612e+02 9.086e+02, threshold=5.730e+02, percent-clipped=2.0 +2023-03-10 00:19:47,828 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.84 vs. limit=2.0 +2023-03-10 00:19:50,758 INFO [train.py:898] (0/4) Epoch 29, batch 3600, loss[loss=0.149, simple_loss=0.2413, pruned_loss=0.02831, over 18383.00 frames. ], tot_loss[loss=0.1559, simple_loss=0.2466, pruned_loss=0.03259, over 3579741.02 frames. ], batch size: 50, lr: 3.83e-03, grad_scale: 8.0 +2023-03-10 00:20:18,598 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.3540, 2.1726, 2.0134, 2.1506, 2.4560, 2.3834, 2.3741, 2.1213], + device='cuda:0'), covar=tensor([0.0303, 0.0300, 0.0555, 0.0436, 0.0279, 0.0267, 0.0422, 0.0369], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0151, 0.0170, 0.0170, 0.0147, 0.0133, 0.0166, 0.0169], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0003, 0.0004], + device='cuda:0') +2023-03-10 00:20:26,691 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-29.pt +2023-03-10 00:20:53,445 INFO [train.py:898] (0/4) Epoch 30, batch 0, loss[loss=0.152, simple_loss=0.2436, pruned_loss=0.03018, over 18384.00 frames. ], tot_loss[loss=0.152, simple_loss=0.2436, pruned_loss=0.03018, over 18384.00 frames. ], batch size: 50, lr: 3.77e-03, grad_scale: 8.0 +2023-03-10 00:20:53,447 INFO [train.py:923] (0/4) Computing validation loss +2023-03-10 00:21:02,451 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7622, 4.7176, 4.7859, 4.3850, 4.5105, 4.4746, 4.7978, 4.8394], + device='cuda:0'), covar=tensor([0.0067, 0.0051, 0.0053, 0.0119, 0.0066, 0.0147, 0.0063, 0.0069], + device='cuda:0'), in_proj_covar=tensor([0.0101, 0.0075, 0.0080, 0.0101, 0.0080, 0.0108, 0.0092, 0.0092], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-10 00:21:05,361 INFO [train.py:932] (0/4) Epoch 30, validation: loss=0.1503, simple_loss=0.2477, pruned_loss=0.02643, over 944034.00 frames. +2023-03-10 00:21:05,361 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-10 00:21:06,906 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105388.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:21:11,612 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7245, 3.6478, 5.0223, 2.9953, 4.4464, 2.5398, 3.0202, 1.6991], + device='cuda:0'), covar=tensor([0.1388, 0.1003, 0.0231, 0.1028, 0.0511, 0.2912, 0.2944, 0.2608], + device='cuda:0'), in_proj_covar=tensor([0.0234, 0.0256, 0.0235, 0.0209, 0.0268, 0.0283, 0.0340, 0.0250], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-10 00:21:36,269 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.589e+02 2.529e+02 2.877e+02 3.490e+02 8.878e+02, threshold=5.754e+02, percent-clipped=2.0 +2023-03-10 00:22:04,731 INFO [train.py:898] (0/4) Epoch 30, batch 50, loss[loss=0.1614, simple_loss=0.2561, pruned_loss=0.03336, over 18499.00 frames. ], tot_loss[loss=0.1554, simple_loss=0.2467, pruned_loss=0.03203, over 809910.43 frames. ], batch size: 53, lr: 3.77e-03, grad_scale: 4.0 +2023-03-10 00:22:18,478 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105449.0, num_to_drop=1, layers_to_drop={1} +2023-03-10 00:22:26,463 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.11 vs. limit=5.0 +2023-03-10 00:22:30,149 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9950, 5.0912, 5.2536, 5.2906, 4.9855, 5.7790, 5.3958, 5.0785], + device='cuda:0'), covar=tensor([0.1094, 0.0740, 0.0814, 0.0738, 0.1338, 0.0702, 0.0732, 0.1845], + device='cuda:0'), in_proj_covar=tensor([0.0377, 0.0309, 0.0337, 0.0340, 0.0343, 0.0451, 0.0305, 0.0443], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-10 00:23:03,151 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105486.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:23:03,939 INFO [train.py:898] (0/4) Epoch 30, batch 100, loss[loss=0.1311, simple_loss=0.2148, pruned_loss=0.02374, over 18393.00 frames. ], tot_loss[loss=0.1548, simple_loss=0.2452, pruned_loss=0.03214, over 1427916.33 frames. ], batch size: 43, lr: 3.77e-03, grad_scale: 2.0 +2023-03-10 00:23:21,387 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105502.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:23:36,854 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.707e+02 2.490e+02 2.952e+02 3.451e+02 6.193e+02, threshold=5.904e+02, percent-clipped=1.0 +2023-03-10 00:23:50,600 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6678, 2.2941, 2.6539, 2.6549, 3.1484, 4.5585, 4.6275, 3.1941], + device='cuda:0'), covar=tensor([0.2144, 0.2741, 0.3058, 0.2097, 0.2561, 0.0329, 0.0370, 0.1182], + device='cuda:0'), in_proj_covar=tensor([0.0339, 0.0370, 0.0423, 0.0297, 0.0404, 0.0273, 0.0306, 0.0278], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0002], + device='cuda:0') +2023-03-10 00:24:02,949 INFO [train.py:898] (0/4) Epoch 30, batch 150, loss[loss=0.1808, simple_loss=0.2737, pruned_loss=0.0439, over 17963.00 frames. ], tot_loss[loss=0.1556, simple_loss=0.2462, pruned_loss=0.03252, over 1899500.54 frames. ], batch size: 65, lr: 3.77e-03, grad_scale: 2.0 +2023-03-10 00:24:14,936 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105547.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:24:23,111 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.84 vs. limit=5.0 +2023-03-10 00:24:33,654 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105563.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:25:02,171 INFO [train.py:898] (0/4) Epoch 30, batch 200, loss[loss=0.1462, simple_loss=0.2325, pruned_loss=0.02996, over 18509.00 frames. ], tot_loss[loss=0.1538, simple_loss=0.2439, pruned_loss=0.03185, over 2273353.14 frames. ], batch size: 47, lr: 3.76e-03, grad_scale: 2.0 +2023-03-10 00:25:06,921 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8789, 4.1065, 2.4769, 3.9636, 5.2218, 2.7495, 3.8651, 3.9804], + device='cuda:0'), covar=tensor([0.0234, 0.1352, 0.1685, 0.0727, 0.0111, 0.1202, 0.0685, 0.0789], + device='cuda:0'), in_proj_covar=tensor([0.0188, 0.0285, 0.0212, 0.0203, 0.0147, 0.0187, 0.0224, 0.0233], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-10 00:25:34,560 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.818e+02 2.661e+02 3.049e+02 3.613e+02 6.015e+02, threshold=6.098e+02, percent-clipped=1.0 +2023-03-10 00:25:34,959 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5107, 3.3837, 2.1550, 4.2983, 3.0759, 4.0247, 2.4827, 3.6772], + device='cuda:0'), covar=tensor([0.0716, 0.0883, 0.1581, 0.0531, 0.0878, 0.0295, 0.1265, 0.0503], + device='cuda:0'), in_proj_covar=tensor([0.0226, 0.0237, 0.0200, 0.0305, 0.0201, 0.0276, 0.0212, 0.0213], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-10 00:26:01,240 INFO [train.py:898] (0/4) Epoch 30, batch 250, loss[loss=0.1495, simple_loss=0.2446, pruned_loss=0.02719, over 18486.00 frames. ], tot_loss[loss=0.1559, simple_loss=0.2465, pruned_loss=0.03265, over 2557126.92 frames. ], batch size: 47, lr: 3.76e-03, grad_scale: 2.0 +2023-03-10 00:26:16,614 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.35 vs. limit=2.0 +2023-03-10 00:26:29,610 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.0676, 5.3731, 2.8851, 5.2578, 5.0541, 5.3547, 5.1675, 2.4476], + device='cuda:0'), covar=tensor([0.0246, 0.0151, 0.0889, 0.0106, 0.0136, 0.0175, 0.0169, 0.1503], + device='cuda:0'), in_proj_covar=tensor([0.0093, 0.0083, 0.0097, 0.0099, 0.0090, 0.0080, 0.0087, 0.0098], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0004, 0.0005], + device='cuda:0') +2023-03-10 00:26:56,638 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9099, 4.7236, 4.7025, 3.6686, 3.8672, 3.5806, 2.9168, 2.8820], + device='cuda:0'), covar=tensor([0.0230, 0.0112, 0.0075, 0.0272, 0.0326, 0.0249, 0.0632, 0.0716], + device='cuda:0'), in_proj_covar=tensor([0.0076, 0.0064, 0.0071, 0.0073, 0.0093, 0.0072, 0.0079, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0006, 0.0005, 0.0006, 0.0006], + device='cuda:0') +2023-03-10 00:26:59,687 INFO [train.py:898] (0/4) Epoch 30, batch 300, loss[loss=0.133, simple_loss=0.2126, pruned_loss=0.02665, over 18494.00 frames. ], tot_loss[loss=0.1552, simple_loss=0.2459, pruned_loss=0.03221, over 2797448.05 frames. ], batch size: 44, lr: 3.76e-03, grad_scale: 2.0 +2023-03-10 00:27:07,316 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.0626, 2.5360, 3.3000, 3.0731, 2.3882, 3.4099, 3.1959, 2.5277], + device='cuda:0'), covar=tensor([0.0606, 0.1324, 0.0498, 0.0472, 0.1457, 0.0373, 0.0687, 0.0961], + device='cuda:0'), in_proj_covar=tensor([0.0225, 0.0247, 0.0239, 0.0176, 0.0228, 0.0223, 0.0261, 0.0201], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-10 00:27:32,634 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.857e+02 2.459e+02 2.793e+02 3.329e+02 5.201e+02, threshold=5.586e+02, percent-clipped=0.0 +2023-03-10 00:27:58,848 INFO [train.py:898] (0/4) Epoch 30, batch 350, loss[loss=0.142, simple_loss=0.2256, pruned_loss=0.02923, over 18516.00 frames. ], tot_loss[loss=0.1549, simple_loss=0.2461, pruned_loss=0.03186, over 2963494.49 frames. ], batch size: 44, lr: 3.76e-03, grad_scale: 2.0 +2023-03-10 00:28:08,218 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=105744.0, num_to_drop=1, layers_to_drop={0} +2023-03-10 00:28:21,623 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.8200, 4.1607, 2.5006, 3.8907, 5.2377, 2.7555, 3.8740, 4.0336], + device='cuda:0'), covar=tensor([0.0230, 0.1220, 0.1781, 0.0822, 0.0091, 0.1347, 0.0718, 0.0726], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0288, 0.0214, 0.0205, 0.0149, 0.0189, 0.0226, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-10 00:28:28,666 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.19 vs. limit=2.0 +2023-03-10 00:28:40,232 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.92 vs. limit=2.0 +2023-03-10 00:28:57,667 INFO [train.py:898] (0/4) Epoch 30, batch 400, loss[loss=0.1394, simple_loss=0.2335, pruned_loss=0.02259, over 18484.00 frames. ], tot_loss[loss=0.1546, simple_loss=0.2459, pruned_loss=0.03161, over 3103035.53 frames. ], batch size: 53, lr: 3.76e-03, grad_scale: 4.0 +2023-03-10 00:29:03,556 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105792.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:29:12,521 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=105799.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:29:30,416 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.921e+02 2.502e+02 2.978e+02 3.750e+02 7.883e+02, threshold=5.955e+02, percent-clipped=2.0 +2023-03-10 00:29:35,934 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.79 vs. limit=5.0 +2023-03-10 00:29:47,179 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6657, 4.3570, 4.3050, 3.2935, 3.5686, 3.2989, 2.4804, 2.4212], + device='cuda:0'), covar=tensor([0.0240, 0.0156, 0.0096, 0.0347, 0.0337, 0.0284, 0.0763, 0.0911], + device='cuda:0'), in_proj_covar=tensor([0.0077, 0.0064, 0.0071, 0.0073, 0.0093, 0.0072, 0.0080, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0007, 0.0005, 0.0006, 0.0006], + device='cuda:0') +2023-03-10 00:29:55,604 INFO [train.py:898] (0/4) Epoch 30, batch 450, loss[loss=0.1497, simple_loss=0.2358, pruned_loss=0.03182, over 18288.00 frames. ], tot_loss[loss=0.1538, simple_loss=0.2451, pruned_loss=0.03121, over 3224339.95 frames. ], batch size: 49, lr: 3.76e-03, grad_scale: 4.0 +2023-03-10 00:30:01,520 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=105842.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:30:15,345 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105853.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:30:20,854 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=105858.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:30:23,365 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=105860.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:30:54,158 INFO [train.py:898] (0/4) Epoch 30, batch 500, loss[loss=0.1638, simple_loss=0.2594, pruned_loss=0.03411, over 18290.00 frames. ], tot_loss[loss=0.1544, simple_loss=0.246, pruned_loss=0.03138, over 3309973.47 frames. ], batch size: 57, lr: 3.76e-03, grad_scale: 4.0 +2023-03-10 00:31:21,507 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.5897, 5.5466, 5.2244, 5.5370, 5.4592, 4.9552, 5.4196, 5.1385], + device='cuda:0'), covar=tensor([0.0402, 0.0416, 0.1237, 0.0665, 0.0629, 0.0394, 0.0408, 0.1072], + device='cuda:0'), in_proj_covar=tensor([0.0519, 0.0588, 0.0725, 0.0456, 0.0491, 0.0535, 0.0566, 0.0707], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-10 00:31:25,881 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.730e+02 2.387e+02 2.835e+02 3.555e+02 6.672e+02, threshold=5.669e+02, percent-clipped=2.0 +2023-03-10 00:31:51,016 INFO [train.py:898] (0/4) Epoch 30, batch 550, loss[loss=0.1386, simple_loss=0.2266, pruned_loss=0.02532, over 18445.00 frames. ], tot_loss[loss=0.1541, simple_loss=0.2458, pruned_loss=0.03124, over 3378521.90 frames. ], batch size: 43, lr: 3.76e-03, grad_scale: 4.0 +2023-03-10 00:32:20,095 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6517, 4.7170, 4.7952, 4.4972, 4.4514, 4.5265, 4.8599, 4.8646], + device='cuda:0'), covar=tensor([0.0089, 0.0081, 0.0074, 0.0139, 0.0083, 0.0193, 0.0082, 0.0102], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0077, 0.0082, 0.0103, 0.0082, 0.0112, 0.0094, 0.0094], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-10 00:32:47,914 INFO [train.py:898] (0/4) Epoch 30, batch 600, loss[loss=0.1666, simple_loss=0.2554, pruned_loss=0.03892, over 18465.00 frames. ], tot_loss[loss=0.1541, simple_loss=0.2456, pruned_loss=0.03128, over 3434649.53 frames. ], batch size: 59, lr: 3.76e-03, grad_scale: 4.0 +2023-03-10 00:33:04,565 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-106000.pt +2023-03-10 00:33:26,793 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.547e+02 2.468e+02 2.895e+02 3.494e+02 7.378e+02, threshold=5.790e+02, percent-clipped=2.0 +2023-03-10 00:33:51,606 INFO [train.py:898] (0/4) Epoch 30, batch 650, loss[loss=0.1605, simple_loss=0.2546, pruned_loss=0.03323, over 18209.00 frames. ], tot_loss[loss=0.1544, simple_loss=0.2461, pruned_loss=0.03135, over 3466258.76 frames. ], batch size: 60, lr: 3.76e-03, grad_scale: 4.0 +2023-03-10 00:34:00,457 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106044.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:34:29,761 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-10 00:34:38,558 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=106077.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:34:49,502 INFO [train.py:898] (0/4) Epoch 30, batch 700, loss[loss=0.1465, simple_loss=0.2372, pruned_loss=0.02792, over 18509.00 frames. ], tot_loss[loss=0.154, simple_loss=0.2459, pruned_loss=0.03105, over 3505175.71 frames. ], batch size: 47, lr: 3.76e-03, grad_scale: 4.0 +2023-03-10 00:34:55,338 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=106092.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:35:14,100 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.86 vs. limit=2.0 +2023-03-10 00:35:22,424 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.614e+02 2.517e+02 2.950e+02 3.495e+02 6.202e+02, threshold=5.900e+02, percent-clipped=1.0 +2023-03-10 00:35:48,487 INFO [train.py:898] (0/4) Epoch 30, batch 750, loss[loss=0.179, simple_loss=0.2619, pruned_loss=0.04808, over 12438.00 frames. ], tot_loss[loss=0.1535, simple_loss=0.2452, pruned_loss=0.03094, over 3521873.92 frames. ], batch size: 130, lr: 3.75e-03, grad_scale: 4.0 +2023-03-10 00:35:49,966 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106138.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:35:54,444 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106142.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:36:01,497 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106148.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:36:09,951 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106155.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:36:13,459 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106158.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:36:46,887 INFO [train.py:898] (0/4) Epoch 30, batch 800, loss[loss=0.1414, simple_loss=0.233, pruned_loss=0.02488, over 18517.00 frames. ], tot_loss[loss=0.1537, simple_loss=0.2454, pruned_loss=0.03093, over 3544597.20 frames. ], batch size: 47, lr: 3.75e-03, grad_scale: 8.0 +2023-03-10 00:36:50,258 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=106190.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:37:09,714 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=106206.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:37:19,671 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.545e+02 2.490e+02 2.877e+02 3.308e+02 6.408e+02, threshold=5.753e+02, percent-clipped=1.0 +2023-03-10 00:37:45,816 INFO [train.py:898] (0/4) Epoch 30, batch 850, loss[loss=0.1351, simple_loss=0.2235, pruned_loss=0.02335, over 18492.00 frames. ], tot_loss[loss=0.1535, simple_loss=0.2451, pruned_loss=0.03099, over 3568527.09 frames. ], batch size: 47, lr: 3.75e-03, grad_scale: 8.0 +2023-03-10 00:38:11,907 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.63 vs. limit=2.0 +2023-03-10 00:38:19,520 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5668, 2.8405, 4.4460, 3.8162, 2.6890, 4.6160, 3.9627, 2.7368], + device='cuda:0'), covar=tensor([0.0560, 0.1560, 0.0248, 0.0427, 0.1580, 0.0207, 0.0545, 0.1148], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0245, 0.0238, 0.0173, 0.0225, 0.0220, 0.0259, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-10 00:38:21,717 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=106268.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:38:42,941 INFO [train.py:898] (0/4) Epoch 30, batch 900, loss[loss=0.14, simple_loss=0.2175, pruned_loss=0.03131, over 18421.00 frames. ], tot_loss[loss=0.1541, simple_loss=0.2455, pruned_loss=0.03131, over 3579914.65 frames. ], batch size: 43, lr: 3.75e-03, grad_scale: 8.0 +2023-03-10 00:39:06,038 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8862, 5.2238, 2.8686, 5.0859, 4.9578, 5.2128, 4.9775, 2.5396], + device='cuda:0'), covar=tensor([0.0237, 0.0069, 0.0737, 0.0073, 0.0079, 0.0073, 0.0108, 0.1091], + device='cuda:0'), in_proj_covar=tensor([0.0094, 0.0085, 0.0099, 0.0101, 0.0092, 0.0081, 0.0088, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-10 00:39:15,254 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.738e+02 2.494e+02 3.069e+02 3.732e+02 1.058e+03, threshold=6.138e+02, percent-clipped=4.0 +2023-03-10 00:39:27,404 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.83 vs. limit=2.0 +2023-03-10 00:39:31,391 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106329.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:39:39,803 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=106336.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:39:40,562 INFO [train.py:898] (0/4) Epoch 30, batch 950, loss[loss=0.1708, simple_loss=0.2648, pruned_loss=0.03842, over 18091.00 frames. ], tot_loss[loss=0.1542, simple_loss=0.2457, pruned_loss=0.03135, over 3574487.02 frames. ], batch size: 62, lr: 3.75e-03, grad_scale: 8.0 +2023-03-10 00:39:40,912 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5263, 3.3577, 2.0668, 4.4054, 2.9164, 3.8431, 2.4068, 3.6732], + device='cuda:0'), covar=tensor([0.0672, 0.0947, 0.1672, 0.0470, 0.0972, 0.0411, 0.1337, 0.0572], + device='cuda:0'), in_proj_covar=tensor([0.0224, 0.0234, 0.0196, 0.0300, 0.0198, 0.0272, 0.0208, 0.0210], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-10 00:40:39,192 INFO [train.py:898] (0/4) Epoch 30, batch 1000, loss[loss=0.1693, simple_loss=0.2599, pruned_loss=0.03938, over 18205.00 frames. ], tot_loss[loss=0.1543, simple_loss=0.2458, pruned_loss=0.0314, over 3575000.79 frames. ], batch size: 60, lr: 3.75e-03, grad_scale: 8.0 +2023-03-10 00:40:50,677 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106397.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:41:00,942 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.4536, 5.3671, 5.7409, 5.7726, 5.3829, 6.2414, 5.9120, 5.4154], + device='cuda:0'), covar=tensor([0.1124, 0.0643, 0.0747, 0.0747, 0.1533, 0.0721, 0.0742, 0.1688], + device='cuda:0'), in_proj_covar=tensor([0.0378, 0.0308, 0.0337, 0.0339, 0.0345, 0.0448, 0.0305, 0.0446], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-10 00:41:06,559 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-03-10 00:41:11,526 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.702e+02 2.421e+02 2.824e+02 3.495e+02 5.852e+02, threshold=5.647e+02, percent-clipped=0.0 +2023-03-10 00:41:11,786 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0365, 5.1515, 5.3733, 5.4263, 5.0249, 5.9152, 5.5052, 5.1341], + device='cuda:0'), covar=tensor([0.1297, 0.0709, 0.0844, 0.0808, 0.1505, 0.0739, 0.0764, 0.1712], + device='cuda:0'), in_proj_covar=tensor([0.0378, 0.0308, 0.0337, 0.0339, 0.0345, 0.0449, 0.0306, 0.0446], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-10 00:41:20,491 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0503, 5.1062, 5.1242, 4.8647, 4.8606, 4.9611, 5.2398, 5.2232], + device='cuda:0'), covar=tensor([0.0077, 0.0065, 0.0059, 0.0120, 0.0069, 0.0151, 0.0064, 0.0083], + device='cuda:0'), in_proj_covar=tensor([0.0104, 0.0077, 0.0083, 0.0104, 0.0083, 0.0113, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-10 00:41:32,611 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106433.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:41:37,594 INFO [train.py:898] (0/4) Epoch 30, batch 1050, loss[loss=0.1639, simple_loss=0.2601, pruned_loss=0.03382, over 18334.00 frames. ], tot_loss[loss=0.1542, simple_loss=0.2456, pruned_loss=0.03141, over 3578234.20 frames. ], batch size: 56, lr: 3.75e-03, grad_scale: 8.0 +2023-03-10 00:41:51,021 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106448.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:41:57,884 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=106454.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:41:58,928 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106455.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:42:23,828 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4558, 2.7577, 4.1634, 3.5865, 2.3987, 4.3415, 3.8609, 2.8263], + device='cuda:0'), covar=tensor([0.0588, 0.1486, 0.0342, 0.0469, 0.1771, 0.0258, 0.0536, 0.0976], + device='cuda:0'), in_proj_covar=tensor([0.0223, 0.0245, 0.0237, 0.0173, 0.0225, 0.0220, 0.0258, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-10 00:42:33,692 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6198, 3.2882, 4.5301, 3.2119, 3.9937, 2.4981, 2.8330, 2.0064], + device='cuda:0'), covar=tensor([0.1338, 0.1104, 0.0262, 0.0769, 0.0583, 0.2745, 0.2629, 0.2313], + device='cuda:0'), in_proj_covar=tensor([0.0236, 0.0257, 0.0237, 0.0211, 0.0269, 0.0285, 0.0342, 0.0251], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-10 00:42:35,946 INFO [train.py:898] (0/4) Epoch 30, batch 1100, loss[loss=0.1413, simple_loss=0.2332, pruned_loss=0.02473, over 18474.00 frames. ], tot_loss[loss=0.1545, simple_loss=0.2456, pruned_loss=0.03167, over 3547113.80 frames. ], batch size: 47, lr: 3.75e-03, grad_scale: 8.0 +2023-03-10 00:42:36,425 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9698, 3.7675, 5.0271, 4.2720, 3.4457, 2.9495, 4.3787, 5.2085], + device='cuda:0'), covar=tensor([0.0790, 0.1388, 0.0268, 0.0464, 0.1004, 0.1321, 0.0451, 0.0354], + device='cuda:0'), in_proj_covar=tensor([0.0156, 0.0287, 0.0179, 0.0189, 0.0201, 0.0197, 0.0204, 0.0218], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-10 00:42:46,761 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=106496.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:42:54,768 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=106503.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:43:08,579 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.832e+02 2.448e+02 2.852e+02 3.322e+02 5.869e+02, threshold=5.705e+02, percent-clipped=1.0 +2023-03-10 00:43:09,030 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=106515.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:43:33,595 INFO [train.py:898] (0/4) Epoch 30, batch 1150, loss[loss=0.1324, simple_loss=0.2168, pruned_loss=0.02399, over 18425.00 frames. ], tot_loss[loss=0.1551, simple_loss=0.2462, pruned_loss=0.03201, over 3555942.68 frames. ], batch size: 43, lr: 3.75e-03, grad_scale: 8.0 +2023-03-10 00:43:51,944 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7402, 3.5892, 4.8240, 4.1953, 3.2035, 2.8904, 4.2226, 5.0123], + device='cuda:0'), covar=tensor([0.0803, 0.1262, 0.0223, 0.0425, 0.1017, 0.1265, 0.0425, 0.0322], + device='cuda:0'), in_proj_covar=tensor([0.0157, 0.0288, 0.0180, 0.0190, 0.0202, 0.0198, 0.0205, 0.0219], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-10 00:44:32,011 INFO [train.py:898] (0/4) Epoch 30, batch 1200, loss[loss=0.1581, simple_loss=0.255, pruned_loss=0.0306, over 18389.00 frames. ], tot_loss[loss=0.1554, simple_loss=0.2465, pruned_loss=0.03209, over 3561806.83 frames. ], batch size: 52, lr: 3.75e-03, grad_scale: 8.0 +2023-03-10 00:45:04,657 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.878e+02 2.700e+02 3.058e+02 3.812e+02 6.890e+02, threshold=6.116e+02, percent-clipped=3.0 +2023-03-10 00:45:15,356 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106624.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:45:30,330 INFO [train.py:898] (0/4) Epoch 30, batch 1250, loss[loss=0.157, simple_loss=0.2541, pruned_loss=0.02996, over 18400.00 frames. ], tot_loss[loss=0.1548, simple_loss=0.246, pruned_loss=0.0318, over 3576485.03 frames. ], batch size: 52, lr: 3.75e-03, grad_scale: 8.0 +2023-03-10 00:45:55,845 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8513, 4.5564, 4.5493, 3.5211, 3.7862, 3.4976, 2.7270, 2.6325], + device='cuda:0'), covar=tensor([0.0238, 0.0151, 0.0082, 0.0299, 0.0353, 0.0236, 0.0670, 0.0805], + device='cuda:0'), in_proj_covar=tensor([0.0077, 0.0064, 0.0071, 0.0073, 0.0092, 0.0071, 0.0079, 0.0087], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0006, 0.0005, 0.0006, 0.0006], + device='cuda:0') +2023-03-10 00:46:00,752 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.22 vs. limit=2.0 +2023-03-10 00:46:28,920 INFO [train.py:898] (0/4) Epoch 30, batch 1300, loss[loss=0.1562, simple_loss=0.2581, pruned_loss=0.02719, over 18137.00 frames. ], tot_loss[loss=0.1546, simple_loss=0.2457, pruned_loss=0.03176, over 3580608.98 frames. ], batch size: 62, lr: 3.74e-03, grad_scale: 8.0 +2023-03-10 00:46:34,899 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106692.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:46:44,265 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5868, 2.7639, 2.5788, 2.8629, 3.6161, 3.4959, 3.1211, 2.8521], + device='cuda:0'), covar=tensor([0.0182, 0.0328, 0.0567, 0.0425, 0.0211, 0.0203, 0.0391, 0.0372], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0152, 0.0170, 0.0170, 0.0146, 0.0133, 0.0165, 0.0168], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0003, 0.0004], + device='cuda:0') +2023-03-10 00:47:01,417 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.810e+02 2.359e+02 2.806e+02 3.557e+02 5.057e+02, threshold=5.611e+02, percent-clipped=0.0 +2023-03-10 00:47:22,038 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106733.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:47:26,895 INFO [train.py:898] (0/4) Epoch 30, batch 1350, loss[loss=0.1357, simple_loss=0.2234, pruned_loss=0.02401, over 18505.00 frames. ], tot_loss[loss=0.154, simple_loss=0.245, pruned_loss=0.03153, over 3582284.28 frames. ], batch size: 47, lr: 3.74e-03, grad_scale: 8.0 +2023-03-10 00:47:27,322 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5600, 2.3646, 2.5721, 2.5655, 3.1043, 4.6927, 4.6920, 3.3008], + device='cuda:0'), covar=tensor([0.2227, 0.2760, 0.3214, 0.2208, 0.2757, 0.0334, 0.0384, 0.1073], + device='cuda:0'), in_proj_covar=tensor([0.0339, 0.0367, 0.0422, 0.0294, 0.0402, 0.0272, 0.0305, 0.0278], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0002], + device='cuda:0') +2023-03-10 00:48:18,860 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=106781.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:48:25,445 INFO [train.py:898] (0/4) Epoch 30, batch 1400, loss[loss=0.1284, simple_loss=0.2168, pruned_loss=0.02004, over 18152.00 frames. ], tot_loss[loss=0.1542, simple_loss=0.2456, pruned_loss=0.03142, over 3579674.45 frames. ], batch size: 44, lr: 3.74e-03, grad_scale: 8.0 +2023-03-10 00:48:44,988 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.24 vs. limit=2.0 +2023-03-10 00:48:52,859 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=106810.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:48:58,341 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.725e+02 2.645e+02 3.103e+02 3.852e+02 9.390e+02, threshold=6.206e+02, percent-clipped=1.0 +2023-03-10 00:49:23,850 INFO [train.py:898] (0/4) Epoch 30, batch 1450, loss[loss=0.1673, simple_loss=0.2576, pruned_loss=0.03851, over 18273.00 frames. ], tot_loss[loss=0.1549, simple_loss=0.2462, pruned_loss=0.03177, over 3570001.06 frames. ], batch size: 57, lr: 3.74e-03, grad_scale: 8.0 +2023-03-10 00:49:42,381 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.52 vs. limit=2.0 +2023-03-10 00:49:55,922 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0298, 5.1553, 4.3871, 5.0066, 5.1054, 4.5709, 4.8476, 4.5920], + device='cuda:0'), covar=tensor([0.0826, 0.0694, 0.2501, 0.1086, 0.0823, 0.0520, 0.0843, 0.1438], + device='cuda:0'), in_proj_covar=tensor([0.0528, 0.0599, 0.0740, 0.0463, 0.0498, 0.0543, 0.0578, 0.0722], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0004, 0.0005, 0.0004, 0.0005, 0.0006], + device='cuda:0') +2023-03-10 00:50:22,115 INFO [train.py:898] (0/4) Epoch 30, batch 1500, loss[loss=0.1501, simple_loss=0.2301, pruned_loss=0.03503, over 18411.00 frames. ], tot_loss[loss=0.1551, simple_loss=0.2466, pruned_loss=0.03185, over 3564302.67 frames. ], batch size: 42, lr: 3.74e-03, grad_scale: 8.0 +2023-03-10 00:50:23,762 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9564, 3.4988, 5.0455, 2.9659, 4.3862, 2.5798, 3.0946, 1.7149], + device='cuda:0'), covar=tensor([0.1189, 0.1061, 0.0206, 0.1020, 0.0568, 0.2697, 0.2646, 0.2429], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0257, 0.0236, 0.0211, 0.0269, 0.0284, 0.0342, 0.0251], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-10 00:50:43,224 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9078, 3.5717, 5.0593, 2.9386, 4.3934, 2.6094, 3.0509, 1.6985], + device='cuda:0'), covar=tensor([0.1248, 0.1035, 0.0161, 0.1068, 0.0502, 0.2751, 0.2721, 0.2433], + device='cuda:0'), in_proj_covar=tensor([0.0236, 0.0258, 0.0237, 0.0211, 0.0270, 0.0285, 0.0344, 0.0252], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-10 00:50:55,405 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.852e+02 2.528e+02 2.938e+02 3.558e+02 8.053e+02, threshold=5.876e+02, percent-clipped=1.0 +2023-03-10 00:51:05,761 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106924.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:51:20,623 INFO [train.py:898] (0/4) Epoch 30, batch 1550, loss[loss=0.1444, simple_loss=0.2321, pruned_loss=0.0283, over 18357.00 frames. ], tot_loss[loss=0.1559, simple_loss=0.2473, pruned_loss=0.03219, over 3569189.77 frames. ], batch size: 46, lr: 3.74e-03, grad_scale: 8.0 +2023-03-10 00:52:01,402 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=106972.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:52:18,620 INFO [train.py:898] (0/4) Epoch 30, batch 1600, loss[loss=0.1695, simple_loss=0.2603, pruned_loss=0.03931, over 18235.00 frames. ], tot_loss[loss=0.1558, simple_loss=0.2469, pruned_loss=0.03229, over 3565061.34 frames. ], batch size: 60, lr: 3.74e-03, grad_scale: 8.0 +2023-03-10 00:52:24,668 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=106992.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:52:31,267 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5216, 2.8357, 2.4939, 2.7905, 3.6411, 3.4748, 3.0969, 2.8378], + device='cuda:0'), covar=tensor([0.0215, 0.0309, 0.0641, 0.0461, 0.0179, 0.0176, 0.0420, 0.0428], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0151, 0.0169, 0.0169, 0.0145, 0.0132, 0.0163, 0.0168], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0003, 0.0004], + device='cuda:0') +2023-03-10 00:52:42,200 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9009, 3.8793, 3.7174, 3.3872, 3.6314, 3.1253, 3.0861, 3.8430], + device='cuda:0'), covar=tensor([0.0073, 0.0096, 0.0087, 0.0144, 0.0109, 0.0178, 0.0206, 0.0081], + device='cuda:0'), in_proj_covar=tensor([0.0162, 0.0180, 0.0150, 0.0200, 0.0161, 0.0190, 0.0195, 0.0140], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-10 00:52:50,787 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.531e+02 2.519e+02 3.033e+02 3.655e+02 1.046e+03, threshold=6.066e+02, percent-clipped=4.0 +2023-03-10 00:53:11,290 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6944, 2.9856, 4.4772, 3.6936, 2.8009, 4.7201, 4.0149, 3.0421], + device='cuda:0'), covar=tensor([0.0594, 0.1569, 0.0333, 0.0542, 0.1616, 0.0226, 0.0592, 0.0993], + device='cuda:0'), in_proj_covar=tensor([0.0224, 0.0246, 0.0239, 0.0174, 0.0226, 0.0220, 0.0259, 0.0199], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-10 00:53:16,397 INFO [train.py:898] (0/4) Epoch 30, batch 1650, loss[loss=0.1229, simple_loss=0.2152, pruned_loss=0.01533, over 18489.00 frames. ], tot_loss[loss=0.1555, simple_loss=0.2468, pruned_loss=0.03213, over 3572667.63 frames. ], batch size: 47, lr: 3.74e-03, grad_scale: 8.0 +2023-03-10 00:53:20,439 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=107040.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:54:14,367 INFO [train.py:898] (0/4) Epoch 30, batch 1700, loss[loss=0.1346, simple_loss=0.2213, pruned_loss=0.02398, over 18273.00 frames. ], tot_loss[loss=0.1552, simple_loss=0.2462, pruned_loss=0.03214, over 3576133.90 frames. ], batch size: 45, lr: 3.74e-03, grad_scale: 8.0 +2023-03-10 00:54:27,669 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.9444, 4.6663, 4.7455, 3.6778, 3.8456, 3.6389, 2.7966, 2.7420], + device='cuda:0'), covar=tensor([0.0230, 0.0145, 0.0079, 0.0280, 0.0341, 0.0209, 0.0682, 0.0766], + device='cuda:0'), in_proj_covar=tensor([0.0077, 0.0064, 0.0071, 0.0073, 0.0093, 0.0071, 0.0079, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0005, 0.0005, 0.0005, 0.0007, 0.0005, 0.0006, 0.0006], + device='cuda:0') +2023-03-10 00:54:39,210 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9958, 5.1574, 5.2998, 5.2940, 4.9230, 5.7561, 5.3801, 5.0467], + device='cuda:0'), covar=tensor([0.1237, 0.0698, 0.0777, 0.0870, 0.1447, 0.0701, 0.0742, 0.1687], + device='cuda:0'), in_proj_covar=tensor([0.0380, 0.0311, 0.0339, 0.0343, 0.0345, 0.0454, 0.0306, 0.0449], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-10 00:54:41,575 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=107110.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:54:46,858 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.859e+02 2.555e+02 2.944e+02 3.831e+02 7.585e+02, threshold=5.887e+02, percent-clipped=1.0 +2023-03-10 00:55:12,544 INFO [train.py:898] (0/4) Epoch 30, batch 1750, loss[loss=0.1605, simple_loss=0.2581, pruned_loss=0.03146, over 18358.00 frames. ], tot_loss[loss=0.1545, simple_loss=0.2454, pruned_loss=0.03182, over 3588777.36 frames. ], batch size: 56, lr: 3.74e-03, grad_scale: 8.0 +2023-03-10 00:55:16,977 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107140.0, num_to_drop=1, layers_to_drop={1} +2023-03-10 00:55:30,522 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107152.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:55:37,676 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=107158.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 00:56:08,257 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.79 vs. limit=2.0 +2023-03-10 00:56:12,004 INFO [train.py:898] (0/4) Epoch 30, batch 1800, loss[loss=0.1716, simple_loss=0.2617, pruned_loss=0.04073, over 17922.00 frames. ], tot_loss[loss=0.1546, simple_loss=0.2453, pruned_loss=0.03193, over 3594872.65 frames. ], batch size: 65, lr: 3.74e-03, grad_scale: 8.0 +2023-03-10 00:56:28,959 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107201.0, num_to_drop=1, layers_to_drop={3} +2023-03-10 00:56:30,234 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.6046, 2.4078, 2.5347, 2.5472, 3.0435, 4.5729, 4.5792, 3.0614], + device='cuda:0'), covar=tensor([0.2236, 0.2676, 0.3357, 0.2226, 0.2751, 0.0339, 0.0400, 0.1240], + device='cuda:0'), in_proj_covar=tensor([0.0339, 0.0368, 0.0423, 0.0295, 0.0402, 0.0273, 0.0306, 0.0278], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0002], + device='cuda:0') +2023-03-10 00:56:43,041 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107213.0, num_to_drop=1, layers_to_drop={0} +2023-03-10 00:56:44,846 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.514e+02 2.402e+02 2.925e+02 3.436e+02 5.429e+02, threshold=5.850e+02, percent-clipped=0.0 +2023-03-10 00:57:09,816 INFO [train.py:898] (0/4) Epoch 30, batch 1850, loss[loss=0.1515, simple_loss=0.2422, pruned_loss=0.03037, over 18398.00 frames. ], tot_loss[loss=0.1552, simple_loss=0.2461, pruned_loss=0.03211, over 3594004.92 frames. ], batch size: 48, lr: 3.74e-03, grad_scale: 8.0 +2023-03-10 00:58:08,615 INFO [train.py:898] (0/4) Epoch 30, batch 1900, loss[loss=0.1582, simple_loss=0.25, pruned_loss=0.03315, over 17359.00 frames. ], tot_loss[loss=0.1545, simple_loss=0.2452, pruned_loss=0.03187, over 3591002.97 frames. ], batch size: 78, lr: 3.73e-03, grad_scale: 8.0 +2023-03-10 00:58:41,094 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.738e+02 2.430e+02 2.951e+02 3.799e+02 8.497e+02, threshold=5.903e+02, percent-clipped=5.0 +2023-03-10 00:58:41,812 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.26 vs. limit=2.0 +2023-03-10 00:59:07,027 INFO [train.py:898] (0/4) Epoch 30, batch 1950, loss[loss=0.1622, simple_loss=0.2561, pruned_loss=0.03417, over 17069.00 frames. ], tot_loss[loss=0.155, simple_loss=0.2459, pruned_loss=0.03206, over 3577762.80 frames. ], batch size: 78, lr: 3.73e-03, grad_scale: 8.0 +2023-03-10 01:00:04,528 INFO [train.py:898] (0/4) Epoch 30, batch 2000, loss[loss=0.1639, simple_loss=0.2579, pruned_loss=0.03501, over 16204.00 frames. ], tot_loss[loss=0.1545, simple_loss=0.2455, pruned_loss=0.03175, over 3574848.24 frames. ], batch size: 95, lr: 3.73e-03, grad_scale: 8.0 +2023-03-10 01:00:38,307 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.734e+02 2.377e+02 2.711e+02 3.457e+02 7.129e+02, threshold=5.422e+02, percent-clipped=2.0 +2023-03-10 01:01:03,315 INFO [train.py:898] (0/4) Epoch 30, batch 2050, loss[loss=0.1444, simple_loss=0.2372, pruned_loss=0.02575, over 18352.00 frames. ], tot_loss[loss=0.1545, simple_loss=0.2456, pruned_loss=0.03174, over 3568889.69 frames. ], batch size: 56, lr: 3.73e-03, grad_scale: 8.0 +2023-03-10 01:01:47,039 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.0890, 5.1320, 5.4200, 5.3557, 5.0194, 5.9742, 5.5262, 5.1484], + device='cuda:0'), covar=tensor([0.1164, 0.0674, 0.0771, 0.0810, 0.1320, 0.0652, 0.0709, 0.1650], + device='cuda:0'), in_proj_covar=tensor([0.0379, 0.0310, 0.0337, 0.0343, 0.0342, 0.0452, 0.0306, 0.0447], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-10 01:01:51,935 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7188, 2.2870, 2.5635, 2.6353, 2.9858, 4.5586, 4.4967, 3.0419], + device='cuda:0'), covar=tensor([0.2111, 0.2780, 0.3218, 0.2067, 0.2772, 0.0329, 0.0402, 0.1213], + device='cuda:0'), in_proj_covar=tensor([0.0339, 0.0368, 0.0424, 0.0294, 0.0401, 0.0272, 0.0306, 0.0278], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0001, 0.0001, 0.0002], + device='cuda:0') +2023-03-10 01:02:02,304 INFO [train.py:898] (0/4) Epoch 30, batch 2100, loss[loss=0.1543, simple_loss=0.2427, pruned_loss=0.03299, over 18287.00 frames. ], tot_loss[loss=0.1544, simple_loss=0.2458, pruned_loss=0.03153, over 3578484.64 frames. ], batch size: 49, lr: 3.73e-03, grad_scale: 8.0 +2023-03-10 01:02:13,395 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=107496.0, num_to_drop=1, layers_to_drop={0} +2023-03-10 01:02:15,933 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9734, 3.8616, 5.1688, 3.0577, 4.4980, 2.6262, 3.2743, 1.8841], + device='cuda:0'), covar=tensor([0.1192, 0.0909, 0.0150, 0.0954, 0.0495, 0.2655, 0.2650, 0.2293], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0257, 0.0237, 0.0211, 0.0270, 0.0286, 0.0342, 0.0251], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-10 01:02:27,081 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=107508.0, num_to_drop=1, layers_to_drop={0} +2023-03-10 01:02:29,361 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107510.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:02:36,352 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.605e+02 2.608e+02 2.930e+02 3.296e+02 8.592e+02, threshold=5.861e+02, percent-clipped=2.0 +2023-03-10 01:02:44,992 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.80 vs. limit=5.0 +2023-03-10 01:03:00,488 INFO [train.py:898] (0/4) Epoch 30, batch 2150, loss[loss=0.1506, simple_loss=0.2419, pruned_loss=0.02968, over 18574.00 frames. ], tot_loss[loss=0.1545, simple_loss=0.2457, pruned_loss=0.03165, over 3586540.07 frames. ], batch size: 54, lr: 3.73e-03, grad_scale: 8.0 +2023-03-10 01:03:14,235 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9919, 5.5019, 5.4844, 5.4703, 5.0045, 5.4034, 4.8523, 5.3687], + device='cuda:0'), covar=tensor([0.0240, 0.0248, 0.0171, 0.0436, 0.0322, 0.0210, 0.0960, 0.0309], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0282, 0.0283, 0.0364, 0.0290, 0.0291, 0.0324, 0.0283], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-10 01:03:40,493 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107571.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:03:43,792 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.9279, 5.4381, 5.4374, 5.4032, 4.9472, 5.3482, 4.8491, 5.3161], + device='cuda:0'), covar=tensor([0.0241, 0.0256, 0.0168, 0.0477, 0.0341, 0.0214, 0.0964, 0.0303], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0282, 0.0283, 0.0363, 0.0289, 0.0291, 0.0324, 0.0282], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-10 01:03:58,555 INFO [train.py:898] (0/4) Epoch 30, batch 2200, loss[loss=0.1475, simple_loss=0.2467, pruned_loss=0.02418, over 18314.00 frames. ], tot_loss[loss=0.1545, simple_loss=0.2458, pruned_loss=0.03164, over 3584809.17 frames. ], batch size: 54, lr: 3.73e-03, grad_scale: 8.0 +2023-03-10 01:04:32,915 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.698e+02 2.424e+02 2.775e+02 3.608e+02 1.106e+03, threshold=5.550e+02, percent-clipped=4.0 +2023-03-10 01:04:56,686 INFO [train.py:898] (0/4) Epoch 30, batch 2250, loss[loss=0.1489, simple_loss=0.2441, pruned_loss=0.02689, over 18616.00 frames. ], tot_loss[loss=0.1543, simple_loss=0.2456, pruned_loss=0.03156, over 3594950.22 frames. ], batch size: 52, lr: 3.73e-03, grad_scale: 8.0 +2023-03-10 01:05:22,624 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107659.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:05:54,193 INFO [train.py:898] (0/4) Epoch 30, batch 2300, loss[loss=0.1736, simple_loss=0.2645, pruned_loss=0.04133, over 18286.00 frames. ], tot_loss[loss=0.1537, simple_loss=0.2447, pruned_loss=0.03134, over 3603630.36 frames. ], batch size: 57, lr: 3.73e-03, grad_scale: 8.0 +2023-03-10 01:06:05,920 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.3110, 5.2545, 5.5932, 5.5325, 5.1616, 6.1048, 5.6966, 5.3832], + device='cuda:0'), covar=tensor([0.1113, 0.0626, 0.0795, 0.0833, 0.1361, 0.0629, 0.0686, 0.1643], + device='cuda:0'), in_proj_covar=tensor([0.0380, 0.0309, 0.0338, 0.0343, 0.0343, 0.0451, 0.0304, 0.0447], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-10 01:06:27,333 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.765e+02 2.455e+02 2.885e+02 3.451e+02 1.072e+03, threshold=5.770e+02, percent-clipped=4.0 +2023-03-10 01:06:32,228 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107720.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:06:52,530 INFO [train.py:898] (0/4) Epoch 30, batch 2350, loss[loss=0.1652, simple_loss=0.2609, pruned_loss=0.03472, over 18345.00 frames. ], tot_loss[loss=0.1541, simple_loss=0.245, pruned_loss=0.03161, over 3582304.56 frames. ], batch size: 55, lr: 3.73e-03, grad_scale: 8.0 +2023-03-10 01:06:52,779 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=107737.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:07:51,009 INFO [train.py:898] (0/4) Epoch 30, batch 2400, loss[loss=0.1608, simple_loss=0.2634, pruned_loss=0.02911, over 18354.00 frames. ], tot_loss[loss=0.1544, simple_loss=0.2457, pruned_loss=0.03156, over 3594779.27 frames. ], batch size: 55, lr: 3.73e-03, grad_scale: 8.0 +2023-03-10 01:08:01,167 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=107796.0, num_to_drop=1, layers_to_drop={1} +2023-03-10 01:08:03,448 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=107798.0, num_to_drop=1, layers_to_drop={0} +2023-03-10 01:08:14,952 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=107808.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:08:24,298 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.608e+02 2.477e+02 2.910e+02 3.480e+02 7.076e+02, threshold=5.821e+02, percent-clipped=2.0 +2023-03-10 01:08:28,924 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7674, 5.3150, 5.2697, 5.2631, 4.7968, 5.1670, 4.6629, 5.1541], + device='cuda:0'), covar=tensor([0.0286, 0.0273, 0.0202, 0.0456, 0.0366, 0.0225, 0.1092, 0.0359], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0285, 0.0286, 0.0367, 0.0292, 0.0293, 0.0325, 0.0285], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-10 01:08:48,117 INFO [train.py:898] (0/4) Epoch 30, batch 2450, loss[loss=0.1823, simple_loss=0.2578, pruned_loss=0.05341, over 12650.00 frames. ], tot_loss[loss=0.1547, simple_loss=0.2457, pruned_loss=0.03182, over 3583586.92 frames. ], batch size: 129, lr: 3.72e-03, grad_scale: 8.0 +2023-03-10 01:08:51,470 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.5905, 3.3921, 5.0901, 3.1269, 4.4204, 2.4764, 2.9083, 1.7123], + device='cuda:0'), covar=tensor([0.1431, 0.1184, 0.0165, 0.0848, 0.0472, 0.2702, 0.2629, 0.2448], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0258, 0.0238, 0.0212, 0.0270, 0.0286, 0.0343, 0.0252], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-10 01:08:56,618 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=107844.0, num_to_drop=1, layers_to_drop={1} +2023-03-10 01:09:01,140 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2590, 5.7453, 5.3236, 5.5245, 5.3709, 5.1528, 5.8136, 5.7593], + device='cuda:0'), covar=tensor([0.1162, 0.0871, 0.0738, 0.0749, 0.1533, 0.0709, 0.0668, 0.0760], + device='cuda:0'), in_proj_covar=tensor([0.0649, 0.0576, 0.0411, 0.0597, 0.0796, 0.0593, 0.0811, 0.0629], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0004, 0.0003, 0.0004, 0.0005, 0.0004, 0.0006, 0.0004], + device='cuda:0') +2023-03-10 01:09:04,781 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7834, 5.1505, 2.5675, 5.0164, 4.8363, 5.1429, 4.9363, 2.3991], + device='cuda:0'), covar=tensor([0.0300, 0.0086, 0.0949, 0.0108, 0.0096, 0.0106, 0.0117, 0.1300], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0085, 0.0099, 0.0101, 0.0091, 0.0081, 0.0088, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-10 01:09:06,220 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=4.29 vs. limit=5.0 +2023-03-10 01:09:10,475 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=107856.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:09:18,710 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8448, 3.8454, 5.1008, 4.4498, 3.4867, 3.1355, 4.5752, 5.3673], + device='cuda:0'), covar=tensor([0.0822, 0.1350, 0.0200, 0.0398, 0.0916, 0.1175, 0.0376, 0.0227], + device='cuda:0'), in_proj_covar=tensor([0.0157, 0.0288, 0.0181, 0.0190, 0.0201, 0.0200, 0.0206, 0.0219], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0004, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-10 01:09:21,688 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=107866.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:09:46,132 INFO [train.py:898] (0/4) Epoch 30, batch 2500, loss[loss=0.1368, simple_loss=0.2213, pruned_loss=0.02618, over 18508.00 frames. ], tot_loss[loss=0.1551, simple_loss=0.2459, pruned_loss=0.03213, over 3579699.37 frames. ], batch size: 44, lr: 3.72e-03, grad_scale: 8.0 +2023-03-10 01:10:19,340 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.763e+02 2.392e+02 2.985e+02 3.620e+02 9.433e+02, threshold=5.970e+02, percent-clipped=3.0 +2023-03-10 01:10:44,306 INFO [train.py:898] (0/4) Epoch 30, batch 2550, loss[loss=0.1352, simple_loss=0.2147, pruned_loss=0.02783, over 18157.00 frames. ], tot_loss[loss=0.1551, simple_loss=0.246, pruned_loss=0.03214, over 3580094.06 frames. ], batch size: 44, lr: 3.72e-03, grad_scale: 8.0 +2023-03-10 01:11:15,939 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.8648, 3.5302, 5.1249, 3.0025, 4.4445, 2.5766, 3.1982, 1.8323], + device='cuda:0'), covar=tensor([0.1323, 0.1125, 0.0165, 0.1073, 0.0501, 0.2884, 0.2460, 0.2435], + device='cuda:0'), in_proj_covar=tensor([0.0236, 0.0258, 0.0237, 0.0212, 0.0270, 0.0286, 0.0343, 0.0252], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-10 01:11:42,236 INFO [train.py:898] (0/4) Epoch 30, batch 2600, loss[loss=0.1415, simple_loss=0.2265, pruned_loss=0.02828, over 18242.00 frames. ], tot_loss[loss=0.155, simple_loss=0.2461, pruned_loss=0.03201, over 3588174.02 frames. ], batch size: 45, lr: 3.72e-03, grad_scale: 8.0 +2023-03-10 01:11:52,539 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8857, 4.5341, 4.5414, 3.4694, 3.8236, 3.5310, 2.7741, 2.7471], + device='cuda:0'), covar=tensor([0.0237, 0.0154, 0.0089, 0.0325, 0.0321, 0.0219, 0.0695, 0.0772], + device='cuda:0'), in_proj_covar=tensor([0.0078, 0.0065, 0.0073, 0.0074, 0.0094, 0.0072, 0.0080, 0.0089], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0007, 0.0005, 0.0006, 0.0006], + device='cuda:0') +2023-03-10 01:11:57,567 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/checkpoint-108000.pt +2023-03-10 01:12:19,587 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108015.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:12:20,528 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.532e+02 2.560e+02 3.005e+02 3.499e+02 6.456e+02, threshold=6.010e+02, percent-clipped=1.0 +2023-03-10 01:12:44,878 INFO [train.py:898] (0/4) Epoch 30, batch 2650, loss[loss=0.1545, simple_loss=0.2473, pruned_loss=0.0309, over 18384.00 frames. ], tot_loss[loss=0.1549, simple_loss=0.2458, pruned_loss=0.03198, over 3584899.69 frames. ], batch size: 55, lr: 3.72e-03, grad_scale: 8.0 +2023-03-10 01:13:41,191 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.4822, 2.8387, 4.1685, 3.4884, 2.6684, 4.3538, 3.7652, 3.0138], + device='cuda:0'), covar=tensor([0.0608, 0.1473, 0.0314, 0.0532, 0.1553, 0.0245, 0.0581, 0.0883], + device='cuda:0'), in_proj_covar=tensor([0.0222, 0.0243, 0.0238, 0.0174, 0.0225, 0.0219, 0.0258, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-10 01:13:42,698 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=96, metric=1.28 vs. limit=2.0 +2023-03-10 01:13:44,283 INFO [train.py:898] (0/4) Epoch 30, batch 2700, loss[loss=0.1528, simple_loss=0.2491, pruned_loss=0.0283, over 18485.00 frames. ], tot_loss[loss=0.1544, simple_loss=0.2458, pruned_loss=0.03154, over 3592139.60 frames. ], batch size: 53, lr: 3.72e-03, grad_scale: 8.0 +2023-03-10 01:13:49,846 INFO [scaling.py:679] (0/4) Whitening: num_groups=1, num_channels=384, metric=2.78 vs. limit=5.0 +2023-03-10 01:13:51,614 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108093.0, num_to_drop=1, layers_to_drop={0} +2023-03-10 01:14:17,645 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.687e+02 2.532e+02 2.946e+02 3.512e+02 1.150e+03, threshold=5.891e+02, percent-clipped=1.0 +2023-03-10 01:14:38,629 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.6201, 3.9510, 2.3433, 3.8792, 4.8135, 2.7954, 3.4503, 3.5047], + device='cuda:0'), covar=tensor([0.0276, 0.1459, 0.1679, 0.0675, 0.0141, 0.1065, 0.0823, 0.1025], + device='cuda:0'), in_proj_covar=tensor([0.0190, 0.0289, 0.0215, 0.0205, 0.0150, 0.0188, 0.0225, 0.0235], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0004, 0.0003, 0.0003, 0.0002, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-10 01:14:42,591 INFO [train.py:898] (0/4) Epoch 30, batch 2750, loss[loss=0.1468, simple_loss=0.2335, pruned_loss=0.03006, over 18501.00 frames. ], tot_loss[loss=0.1546, simple_loss=0.2457, pruned_loss=0.03172, over 3588213.22 frames. ], batch size: 47, lr: 3.72e-03, grad_scale: 8.0 +2023-03-10 01:14:50,121 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8444, 4.4483, 4.3838, 3.3462, 3.6499, 3.4060, 2.5024, 2.7439], + device='cuda:0'), covar=tensor([0.0221, 0.0138, 0.0083, 0.0334, 0.0340, 0.0230, 0.0750, 0.0707], + device='cuda:0'), in_proj_covar=tensor([0.0077, 0.0065, 0.0072, 0.0073, 0.0094, 0.0072, 0.0080, 0.0088], + device='cuda:0'), out_proj_covar=tensor([0.0006, 0.0005, 0.0005, 0.0005, 0.0007, 0.0005, 0.0006, 0.0006], + device='cuda:0') +2023-03-10 01:15:16,462 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108166.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:15:41,148 INFO [train.py:898] (0/4) Epoch 30, batch 2800, loss[loss=0.1511, simple_loss=0.2419, pruned_loss=0.03019, over 18305.00 frames. ], tot_loss[loss=0.1547, simple_loss=0.2457, pruned_loss=0.03186, over 3581958.35 frames. ], batch size: 54, lr: 3.72e-03, grad_scale: 8.0 +2023-03-10 01:16:12,749 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=108214.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:16:14,906 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.768e+02 2.463e+02 2.976e+02 3.777e+02 6.000e+02, threshold=5.952e+02, percent-clipped=1.0 +2023-03-10 01:16:40,171 INFO [train.py:898] (0/4) Epoch 30, batch 2850, loss[loss=0.1749, simple_loss=0.257, pruned_loss=0.04636, over 12306.00 frames. ], tot_loss[loss=0.1552, simple_loss=0.2464, pruned_loss=0.03198, over 3560748.26 frames. ], batch size: 133, lr: 3.72e-03, grad_scale: 8.0 +2023-03-10 01:16:47,259 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.5719, 3.1233, 4.4648, 3.6770, 2.8142, 4.6577, 3.9576, 3.1837], + device='cuda:0'), covar=tensor([0.0614, 0.1310, 0.0275, 0.0493, 0.1478, 0.0188, 0.0538, 0.0795], + device='cuda:0'), in_proj_covar=tensor([0.0222, 0.0243, 0.0238, 0.0174, 0.0226, 0.0218, 0.0258, 0.0197], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002], + device='cuda:0') +2023-03-10 01:17:07,817 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7891, 5.2365, 2.5526, 5.0867, 4.9802, 5.2434, 5.0409, 2.6720], + device='cuda:0'), covar=tensor([0.0253, 0.0066, 0.0863, 0.0078, 0.0080, 0.0074, 0.0096, 0.1079], + device='cuda:0'), in_proj_covar=tensor([0.0095, 0.0085, 0.0099, 0.0101, 0.0091, 0.0081, 0.0087, 0.0099], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0004, 0.0005, 0.0005, 0.0004, 0.0004, 0.0005, 0.0005], + device='cuda:0') +2023-03-10 01:17:09,648 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6392, 3.3337, 2.2670, 4.3565, 3.0472, 3.9677, 2.4467, 3.7517], + device='cuda:0'), covar=tensor([0.0656, 0.0915, 0.1530, 0.0503, 0.0836, 0.0406, 0.1264, 0.0478], + device='cuda:0'), in_proj_covar=tensor([0.0227, 0.0237, 0.0199, 0.0302, 0.0201, 0.0272, 0.0211, 0.0211], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-10 01:17:38,309 INFO [train.py:898] (0/4) Epoch 30, batch 2900, loss[loss=0.1555, simple_loss=0.2403, pruned_loss=0.03532, over 17231.00 frames. ], tot_loss[loss=0.1553, simple_loss=0.2465, pruned_loss=0.03203, over 3567699.59 frames. ], batch size: 38, lr: 3.72e-03, grad_scale: 8.0 +2023-03-10 01:18:11,358 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108315.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:18:12,786 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.844e+02 2.541e+02 3.023e+02 3.918e+02 9.927e+02, threshold=6.045e+02, percent-clipped=2.0 +2023-03-10 01:18:23,024 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108325.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:18:25,820 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2218, 5.1946, 5.5585, 5.5098, 5.1839, 6.0183, 5.6746, 5.2928], + device='cuda:0'), covar=tensor([0.1188, 0.0687, 0.0836, 0.0743, 0.1447, 0.0681, 0.0632, 0.1749], + device='cuda:0'), in_proj_covar=tensor([0.0379, 0.0308, 0.0336, 0.0339, 0.0342, 0.0448, 0.0302, 0.0443], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-10 01:18:36,946 INFO [train.py:898] (0/4) Epoch 30, batch 2950, loss[loss=0.1444, simple_loss=0.2418, pruned_loss=0.02351, over 18388.00 frames. ], tot_loss[loss=0.1548, simple_loss=0.2457, pruned_loss=0.03188, over 3573976.66 frames. ], batch size: 50, lr: 3.72e-03, grad_scale: 8.0 +2023-03-10 01:19:03,169 INFO [scaling.py:679] (0/4) Whitening: num_groups=8, num_channels=192, metric=1.60 vs. limit=2.0 +2023-03-10 01:19:03,897 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.2155, 5.2090, 5.4775, 5.4712, 5.1736, 5.9661, 5.6178, 5.1994], + device='cuda:0'), covar=tensor([0.1174, 0.0651, 0.0884, 0.0892, 0.1470, 0.0743, 0.0749, 0.1910], + device='cuda:0'), in_proj_covar=tensor([0.0379, 0.0308, 0.0336, 0.0339, 0.0342, 0.0449, 0.0303, 0.0444], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004], + device='cuda:0') +2023-03-10 01:19:07,152 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=108363.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:19:22,242 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.7824, 2.9364, 2.7208, 3.0691, 3.7752, 3.6942, 3.2938, 3.0812], + device='cuda:0'), covar=tensor([0.0182, 0.0325, 0.0605, 0.0381, 0.0174, 0.0171, 0.0382, 0.0410], + device='cuda:0'), in_proj_covar=tensor([0.0151, 0.0152, 0.0169, 0.0169, 0.0147, 0.0133, 0.0164, 0.0169], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0003, 0.0004], + device='cuda:0') +2023-03-10 01:19:29,425 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8916, 3.8996, 3.7058, 3.3992, 3.6383, 3.0924, 3.1827, 3.9513], + device='cuda:0'), covar=tensor([0.0071, 0.0091, 0.0097, 0.0145, 0.0103, 0.0201, 0.0199, 0.0065], + device='cuda:0'), in_proj_covar=tensor([0.0164, 0.0183, 0.0152, 0.0204, 0.0163, 0.0193, 0.0199, 0.0141], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0003, 0.0002, 0.0003, 0.0003, 0.0002], + device='cuda:0') +2023-03-10 01:19:34,057 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108386.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:19:34,878 INFO [train.py:898] (0/4) Epoch 30, batch 3000, loss[loss=0.1355, simple_loss=0.2152, pruned_loss=0.02792, over 18450.00 frames. ], tot_loss[loss=0.1551, simple_loss=0.2461, pruned_loss=0.03202, over 3580279.85 frames. ], batch size: 43, lr: 3.72e-03, grad_scale: 8.0 +2023-03-10 01:19:34,880 INFO [train.py:923] (0/4) Computing validation loss +2023-03-10 01:19:45,290 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.0833, 1.7615, 1.8025, 1.7471, 1.9697, 2.1357, 2.1103, 1.8718], + device='cuda:0'), covar=tensor([0.0276, 0.0239, 0.0524, 0.0472, 0.0270, 0.0247, 0.0419, 0.0319], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0152, 0.0168, 0.0169, 0.0147, 0.0133, 0.0163, 0.0168], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0003, 0.0004], + device='cuda:0') +2023-03-10 01:19:47,081 INFO [train.py:932] (0/4) Epoch 30, validation: loss=0.1491, simple_loss=0.2469, pruned_loss=0.02567, over 944034.00 frames. +2023-03-10 01:19:47,082 INFO [train.py:933] (0/4) Maximum memory allocated so far is 19897MB +2023-03-10 01:19:54,093 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108393.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:20:01,016 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([4.7889, 4.8552, 4.8966, 4.6407, 4.6395, 4.6741, 4.9512, 4.9469], + device='cuda:0'), covar=tensor([0.0084, 0.0077, 0.0082, 0.0127, 0.0071, 0.0181, 0.0100, 0.0097], + device='cuda:0'), in_proj_covar=tensor([0.0103, 0.0078, 0.0083, 0.0104, 0.0083, 0.0112, 0.0095, 0.0095], + device='cuda:0'), out_proj_covar=tensor([0.0004, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0004, 0.0004], + device='cuda:0') +2023-03-10 01:20:20,295 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.803e+02 2.459e+02 2.931e+02 3.720e+02 6.119e+02, threshold=5.863e+02, percent-clipped=1.0 +2023-03-10 01:20:39,344 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108432.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:20:45,382 INFO [train.py:898] (0/4) Epoch 30, batch 3050, loss[loss=0.1399, simple_loss=0.232, pruned_loss=0.02386, over 18347.00 frames. ], tot_loss[loss=0.154, simple_loss=0.2448, pruned_loss=0.03165, over 3590786.63 frames. ], batch size: 46, lr: 3.71e-03, grad_scale: 8.0 +2023-03-10 01:20:50,485 INFO [zipformer.py:625] (0/4) warmup_begin=666.7, warmup_end=1333.3, batch_count=108441.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:21:43,452 INFO [train.py:898] (0/4) Epoch 30, batch 3100, loss[loss=0.1438, simple_loss=0.2261, pruned_loss=0.03077, over 18482.00 frames. ], tot_loss[loss=0.1543, simple_loss=0.2452, pruned_loss=0.03165, over 3587124.73 frames. ], batch size: 44, lr: 3.71e-03, grad_scale: 8.0 +2023-03-10 01:21:51,105 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108493.0, num_to_drop=1, layers_to_drop={0} +2023-03-10 01:22:17,722 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.743e+02 2.471e+02 2.835e+02 3.390e+02 1.271e+03, threshold=5.670e+02, percent-clipped=3.0 +2023-03-10 01:22:41,853 INFO [train.py:898] (0/4) Epoch 30, batch 3150, loss[loss=0.171, simple_loss=0.26, pruned_loss=0.04096, over 13083.00 frames. ], tot_loss[loss=0.1547, simple_loss=0.246, pruned_loss=0.03172, over 3577974.96 frames. ], batch size: 129, lr: 3.71e-03, grad_scale: 8.0 +2023-03-10 01:23:29,768 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108577.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:23:40,617 INFO [train.py:898] (0/4) Epoch 30, batch 3200, loss[loss=0.1338, simple_loss=0.2158, pruned_loss=0.02589, over 18375.00 frames. ], tot_loss[loss=0.1548, simple_loss=0.2464, pruned_loss=0.03161, over 3588188.29 frames. ], batch size: 42, lr: 3.71e-03, grad_scale: 8.0 +2023-03-10 01:24:16,501 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.665e+02 2.358e+02 2.876e+02 3.361e+02 7.357e+02, threshold=5.753e+02, percent-clipped=3.0 +2023-03-10 01:24:39,476 INFO [train.py:898] (0/4) Epoch 30, batch 3250, loss[loss=0.1557, simple_loss=0.2565, pruned_loss=0.02745, over 18358.00 frames. ], tot_loss[loss=0.1545, simple_loss=0.2458, pruned_loss=0.03164, over 3592507.35 frames. ], batch size: 55, lr: 3.71e-03, grad_scale: 8.0 +2023-03-10 01:24:40,973 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108638.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:25:31,362 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108681.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:25:37,772 INFO [train.py:898] (0/4) Epoch 30, batch 3300, loss[loss=0.1718, simple_loss=0.2601, pruned_loss=0.04172, over 12439.00 frames. ], tot_loss[loss=0.1542, simple_loss=0.2452, pruned_loss=0.03157, over 3579976.26 frames. ], batch size: 129, lr: 3.71e-03, grad_scale: 8.0 +2023-03-10 01:26:12,808 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.499e+02 2.588e+02 2.944e+02 3.535e+02 1.316e+03, threshold=5.888e+02, percent-clipped=3.0 +2023-03-10 01:26:36,468 INFO [train.py:898] (0/4) Epoch 30, batch 3350, loss[loss=0.1365, simple_loss=0.2125, pruned_loss=0.03029, over 17645.00 frames. ], tot_loss[loss=0.1539, simple_loss=0.2449, pruned_loss=0.03149, over 3586039.31 frames. ], batch size: 39, lr: 3.71e-03, grad_scale: 8.0 +2023-03-10 01:27:34,245 INFO [train.py:898] (0/4) Epoch 30, batch 3400, loss[loss=0.1688, simple_loss=0.2587, pruned_loss=0.03948, over 18461.00 frames. ], tot_loss[loss=0.1543, simple_loss=0.2453, pruned_loss=0.03166, over 3587679.39 frames. ], batch size: 59, lr: 3.71e-03, grad_scale: 8.0 +2023-03-10 01:27:35,543 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108788.0, num_to_drop=1, layers_to_drop={2} +2023-03-10 01:27:38,994 INFO [zipformer.py:625] (0/4) warmup_begin=2666.7, warmup_end=3333.3, batch_count=108791.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:27:43,590 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.6987, 3.5657, 2.5373, 4.4956, 3.2911, 4.1811, 2.7156, 4.0869], + device='cuda:0'), covar=tensor([0.0618, 0.0843, 0.1337, 0.0500, 0.0791, 0.0361, 0.1145, 0.0404], + device='cuda:0'), in_proj_covar=tensor([0.0225, 0.0236, 0.0199, 0.0300, 0.0200, 0.0272, 0.0210, 0.0211], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0003, 0.0004, 0.0003, 0.0004, 0.0003, 0.0003], + device='cuda:0') +2023-03-10 01:28:09,508 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.896e+02 2.642e+02 3.152e+02 3.667e+02 6.078e+02, threshold=6.305e+02, percent-clipped=1.0 +2023-03-10 01:28:33,118 INFO [train.py:898] (0/4) Epoch 30, batch 3450, loss[loss=0.1498, simple_loss=0.245, pruned_loss=0.02728, over 16112.00 frames. ], tot_loss[loss=0.1536, simple_loss=0.2448, pruned_loss=0.03118, over 3586118.22 frames. ], batch size: 94, lr: 3.71e-03, grad_scale: 8.0 +2023-03-10 01:28:50,243 INFO [zipformer.py:625] (0/4) warmup_begin=3333.3, warmup_end=4000.0, batch_count=108852.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:28:54,683 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([5.1030, 5.5604, 5.5645, 5.5618, 5.0113, 5.4832, 5.0268, 5.5116], + device='cuda:0'), covar=tensor([0.0232, 0.0258, 0.0170, 0.0381, 0.0392, 0.0220, 0.0902, 0.0285], + device='cuda:0'), in_proj_covar=tensor([0.0237, 0.0282, 0.0284, 0.0362, 0.0291, 0.0291, 0.0321, 0.0281], + device='cuda:0'), out_proj_covar=tensor([0.0005, 0.0006, 0.0006, 0.0008, 0.0005, 0.0006, 0.0006, 0.0006], + device='cuda:0') +2023-03-10 01:29:31,186 INFO [train.py:898] (0/4) Epoch 30, batch 3500, loss[loss=0.158, simple_loss=0.2458, pruned_loss=0.03512, over 18503.00 frames. ], tot_loss[loss=0.1537, simple_loss=0.2448, pruned_loss=0.03126, over 3587458.45 frames. ], batch size: 47, lr: 3.71e-03, grad_scale: 8.0 +2023-03-10 01:30:05,389 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.699e+02 2.567e+02 2.966e+02 3.432e+02 4.808e+02, threshold=5.933e+02, percent-clipped=0.0 +2023-03-10 01:30:07,507 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([3.8782, 3.0489, 2.7863, 3.2123, 3.8740, 3.8281, 3.3887, 3.2577], + device='cuda:0'), covar=tensor([0.0164, 0.0343, 0.0527, 0.0380, 0.0206, 0.0204, 0.0363, 0.0350], + device='cuda:0'), in_proj_covar=tensor([0.0150, 0.0152, 0.0169, 0.0168, 0.0147, 0.0134, 0.0163, 0.0168], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0003, 0.0004, 0.0004, 0.0003, 0.0003, 0.0003, 0.0004], + device='cuda:0') +2023-03-10 01:30:21,146 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.9890, 3.7608, 5.0122, 4.4576, 3.3633, 3.1888, 4.5442, 5.2856], + device='cuda:0'), covar=tensor([0.0810, 0.1418, 0.0260, 0.0436, 0.1063, 0.1251, 0.0419, 0.0405], + device='cuda:0'), in_proj_covar=tensor([0.0158, 0.0289, 0.0182, 0.0191, 0.0202, 0.0201, 0.0207, 0.0221], + device='cuda:0'), out_proj_covar=tensor([0.0003, 0.0005, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003, 0.0003], + device='cuda:0') +2023-03-10 01:30:23,069 INFO [zipformer.py:625] (0/4) warmup_begin=1333.3, warmup_end=2000.0, batch_count=108933.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:30:27,371 INFO [train.py:898] (0/4) Epoch 30, batch 3550, loss[loss=0.1426, simple_loss=0.2354, pruned_loss=0.02492, over 18299.00 frames. ], tot_loss[loss=0.1539, simple_loss=0.2451, pruned_loss=0.0313, over 3586985.97 frames. ], batch size: 49, lr: 3.71e-03, grad_scale: 8.0 +2023-03-10 01:30:33,954 INFO [zipformer.py:1455] (0/4) attn_weights_entropy = tensor([2.7993, 3.6188, 5.0650, 2.9990, 4.4141, 2.6548, 3.1019, 1.7898], + device='cuda:0'), covar=tensor([0.1299, 0.1017, 0.0149, 0.1013, 0.0508, 0.2637, 0.2662, 0.2434], + device='cuda:0'), in_proj_covar=tensor([0.0235, 0.0257, 0.0237, 0.0212, 0.0269, 0.0285, 0.0342, 0.0252], + device='cuda:0'), out_proj_covar=tensor([0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0002, 0.0003, 0.0002], + device='cuda:0') +2023-03-10 01:31:14,529 INFO [zipformer.py:625] (0/4) warmup_begin=2000.0, warmup_end=2666.7, batch_count=108981.0, num_to_drop=0, layers_to_drop=set() +2023-03-10 01:31:20,776 INFO [train.py:898] (0/4) Epoch 30, batch 3600, loss[loss=0.1469, simple_loss=0.2443, pruned_loss=0.02474, over 18306.00 frames. ], tot_loss[loss=0.1536, simple_loss=0.2446, pruned_loss=0.03132, over 3590038.21 frames. ], batch size: 54, lr: 3.71e-03, grad_scale: 8.0 +2023-03-10 01:31:52,858 INFO [optim.py:369] (0/4) Clipping_scale=2.0, grad-norm quartiles 1.673e+02 2.589e+02 3.051e+02 3.699e+02 9.146e+02, threshold=6.102e+02, percent-clipped=6.0 +2023-03-10 01:31:56,908 INFO [checkpoint.py:75] (0/4) Saving checkpoint to pruned_transducer_stateless7/exp_960h_no_paddingidx_ngpu4/epoch-30.pt +2023-03-10 01:32:06,466 INFO [train.py:1165] (0/4) Done!