decoding_results/greedy_search/log-decode-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model-2023-11-01-15-11-56

2023-11-01 15:11:56,345 INFO [decode.py:666] Decoding started
2023-11-01 15:11:56,346 INFO [decode.py:672] Device: cuda:0
2023-11-01 15:11:56,353 INFO [decode.py:682] {'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': '1.24.3', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '821ebc378e7fb99b8adc81950227963332821e01', 'k2-git-date': 'Wed Jul 19 15:38:25 2023', 'lhotse-version': '1.16.0.dev+git.1db4d97a.clean', 'torch-version': '1.11.0+cu102', 'torch-cuda-available': True, 'torch-cuda-version': '10.2', 'python-version': '3.9', 'icefall-git-branch': 'dev_zipformer_cn', 'icefall-git-sha1': '5b9014f7-dirty', 'icefall-git-date': 'Tue Oct 24 16:08:39 2023', 'icefall-path': '/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/icefall-1.0-py3.9.egg', 'k2-path': '/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/k2-1.24.3.dev20230721+cuda10.2.torch1.11.0-py3.9-linux-x86_64.egg/k2/__init__.py', 'lhotse-path': '/star-home/jinzengrui/lib/miniconda3/envs/dev39/lib/python3.9/site-packages/lhotse-1.16.0.dev0+git.1db4d97a.clean-py3.9.egg/lhotse/__init__.py', 'hostname': 'de-74279-k2-train-9-0208143539-7dbf569d4f-r7nrb', 'IP address': '10.177.13.150'}, 'epoch': 20, 'iter': 0, 'avg': 1, 'use_averaged_model': True, 'exp_dir': PosixPath('zipformer/exp-w-ctc-streaming'), 'bpe_model': 'data/lang_bpe_2000/bpe.model', 'lang_dir': PosixPath('data/lang_bpe_2000'), 'decoding_method': 'greedy_search', 'beam_size': 4, 'beam': 20.0, 'ngram_lm_scale': 0.01, 'max_contexts': 8, 'max_states': 64, 'context_size': 2, 'max_sym_per_frame': 1, 'num_paths': 200, 'nbest_scale': 0.5, 'num_encoder_layers': '2,2,3,4,3,2', 'downsampling_factor': '1,2,4,8,4,2', 'feedforward_dim': '512,768,1024,1536,1024,768', 'num_heads': '4,4,4,8,4,4', 'encoder_dim': '192,256,384,512,384,256', 'query_head_dim': '32', 'value_head_dim': '12', 'pos_head_dim': '4', 'pos_dim': 48, 'encoder_unmasked_dim': '192,192,256,256,256,192', 'cnn_module_kernel': '31,31,15,15,15,31', 'decoder_dim': 512, 'joiner_dim': 512, 'causal': True, 'chunk_size': '32', 'left_context_frames': '256', 'use_transducer': True, 'use_ctc': True, 'manifest_dir': PosixPath('data/fbank'), 'max_duration': 300.0, 'bucketing_sampler': True, 'num_buckets': 30, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': True, 'drop_last': True, 'return_cuts': True, 'num_workers': 2, 'enable_spec_aug': True, 'spec_aug_time_warp_factor': 80, 'enable_musan': True, 'input_strategy': 'PrecomputedFeatures', 'res_dir': PosixPath('zipformer/exp-w-ctc-streaming/greedy_search'), 'suffix': 'epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model', 'blank_id': 0, 'unk_id': 2, 'vocab_size': 2000}
2023-11-01 15:11:56,353 INFO [decode.py:684] About to create model
2023-11-01 15:11:57,917 INFO [decode.py:751] Calculating the averaged model over epoch range from 19 (excluded) to 20
2023-11-01 15:12:13,614 INFO [decode.py:785] Number of model parameters: 70213431
2023-11-01 15:12:13,614 INFO [multi_dataset.py:221] About to get multidataset test cuts
2023-11-01 15:12:13,614 INFO [multi_dataset.py:224] Loading Aidatatang_200zh set in lazy mode
2023-11-01 15:12:13,638 INFO [multi_dataset.py:233] Loading Aishell set in lazy mode
2023-11-01 15:12:13,642 INFO [multi_dataset.py:242] Loading Aishell-2 set in lazy mode
2023-11-01 15:12:13,646 INFO [multi_dataset.py:251] Loading Aishell-4 TEST set in lazy mode
2023-11-01 15:12:13,649 INFO [multi_dataset.py:257] Loading Ali-Meeting set in lazy mode
2023-11-01 15:12:13,653 INFO [multi_dataset.py:266] Loading MagicData set in lazy mode
2023-11-01 15:12:13,659 INFO [multi_dataset.py:275] Loading KeSpeech set in lazy mode
2023-11-01 15:12:13,673 INFO [multi_dataset.py:287] Loading WeNetSpeech set in lazy mode
2023-11-01 15:12:26,734 WARNING [decode.py:795] Excluding cut with ID: TEST_NET_Y0000000004_0ub4ZzdHzBc_S00023 from decoding, num_frames: 8
2023-11-01 15:12:28,710 INFO [decode.py:809] Start decoding test set: aidatatang_test
2023-11-01 15:12:31,492 INFO [decode.py:563] batch 0/?, cuts processed until now is 80
2023-11-01 15:12:38,609 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([2.1541, 1.0672, 1.6325, 1.4503, 1.7398, 1.4591, 1.6340, 1.3796],
       device='cuda:0')
2023-11-01 15:13:01,016 INFO [decode.py:563] batch 50/?, cuts processed until now is 4543
2023-11-01 15:13:31,552 INFO [decode.py:563] batch 100/?, cuts processed until now is 9084
2023-11-01 15:14:01,372 INFO [decode.py:563] batch 150/?, cuts processed until now is 13880
2023-11-01 15:14:14,535 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.5692, 1.1975, 1.5079, 2.5953], device='cuda:0')
2023-11-01 15:14:31,508 INFO [decode.py:563] batch 200/?, cuts processed until now is 18516
2023-11-01 15:14:50,293 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([2.7910, 1.7359, 2.0341, 2.0215, 2.0379, 1.9213, 2.0914, 2.0142],
       device='cuda:0')
2023-11-01 15:14:59,764 INFO [decode.py:563] batch 250/?, cuts processed until now is 22994
2023-11-01 15:15:30,128 INFO [decode.py:563] batch 300/?, cuts processed until now is 28179
2023-11-01 15:15:48,646 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([2.4999, 2.3366, 2.5058, 2.8586], device='cuda:0')
2023-11-01 15:16:01,293 INFO [decode.py:563] batch 350/?, cuts processed until now is 32886
2023-11-01 15:16:30,531 INFO [decode.py:563] batch 400/?, cuts processed until now is 37667
2023-11-01 15:16:31,239 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.0964, 1.0647, 1.5409, 1.5783, 1.3994, 1.4800, 1.5416, 1.2848],
       device='cuda:0')
2023-11-01 15:17:01,327 INFO [decode.py:563] batch 450/?, cuts processed until now is 42122
2023-11-01 15:17:27,429 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.2224, 2.7105, 2.0970, 1.6589], device='cuda:0')
2023-11-01 15:17:33,538 INFO [decode.py:563] batch 500/?, cuts processed until now is 46172
2023-11-01 15:17:57,296 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.0637, 0.9826, 1.5031, 1.5560, 1.3731, 1.4742, 1.5735, 1.2702],
       device='cuda:0')
2023-11-01 15:17:59,572 INFO [decode.py:579] The transcripts are stored in zipformer/exp-w-ctc-streaming/greedy_search/recogs-aidatatang_test-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:18:01,472 INFO [utils.py:565] [aidatatang_test-greedy_search] %WER 5.18% [24311 / 468933, 3808 ins, 1531 del, 18972 sub ]
2023-11-01 15:18:05,139 INFO [decode.py:592] Wrote detailed error stats to zipformer/exp-w-ctc-streaming/greedy_search/errs-aidatatang_test-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:18:05,144 INFO [decode.py:608] 
For aidatatang_test, WER of different settings are:
greedy_search	5.18	best for aidatatang_test

2023-11-01 15:18:05,145 INFO [decode.py:809] Start decoding test set: aidatatang_dev
2023-11-01 15:18:08,028 INFO [decode.py:563] batch 0/?, cuts processed until now is 81
2023-11-01 15:18:38,656 INFO [decode.py:563] batch 50/?, cuts processed until now is 4556
2023-11-01 15:19:09,338 INFO [decode.py:563] batch 100/?, cuts processed until now is 9077
2023-11-01 15:19:30,020 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.5282, 2.8812, 2.3507, 1.9345], device='cuda:0')
2023-11-01 15:19:39,844 INFO [decode.py:563] batch 150/?, cuts processed until now is 13773
2023-11-01 15:20:00,060 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([2.5989, 2.8339, 2.4733, 2.9824], device='cuda:0')
2023-11-01 15:20:09,908 INFO [decode.py:563] batch 200/?, cuts processed until now is 18432
2023-11-01 15:20:43,300 INFO [decode.py:563] batch 250/?, cuts processed until now is 22590
2023-11-01 15:20:54,974 INFO [decode.py:579] The transcripts are stored in zipformer/exp-w-ctc-streaming/greedy_search/recogs-aidatatang_dev-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:20:55,845 INFO [utils.py:565] [aidatatang_dev-greedy_search] %WER 4.71% [11037 / 234524, 1858 ins, 804 del, 8375 sub ]
2023-11-01 15:20:57,528 INFO [decode.py:592] Wrote detailed error stats to zipformer/exp-w-ctc-streaming/greedy_search/errs-aidatatang_dev-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:20:57,533 INFO [decode.py:608] 
For aidatatang_dev, WER of different settings are:
greedy_search	4.71	best for aidatatang_dev

2023-11-01 15:20:57,534 INFO [decode.py:809] Start decoding test set: alimeeting_test
2023-11-01 15:21:00,984 INFO [decode.py:563] batch 0/?, cuts processed until now is 44
2023-11-01 15:21:19,797 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([2.0249, 1.7796, 2.1354, 1.8768, 1.9674, 2.0313, 1.6473, 1.9614],
       device='cuda:0')
2023-11-01 15:21:31,341 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([0.7993, 4.1113, 3.7712, 1.4562], device='cuda:0')
2023-11-01 15:21:38,966 WARNING [decode.py:795] Excluding cut with ID: R8008_M8016-8062-123 from decoding, num_frames: 6
2023-11-01 15:21:47,476 INFO [decode.py:563] batch 50/?, cuts processed until now is 3819
2023-11-01 15:21:49,558 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.5677, 1.1293, 1.4885, 2.8541], device='cuda:0')
2023-11-01 15:22:37,340 INFO [decode.py:563] batch 100/?, cuts processed until now is 7625
2023-11-01 15:22:44,839 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.2946, 3.2570, 2.3165, 1.7768], device='cuda:0')
2023-11-01 15:23:25,992 INFO [decode.py:563] batch 150/?, cuts processed until now is 11772
2023-11-01 15:23:39,445 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([2.2975, 1.9610, 2.3670, 2.1581, 2.2752, 2.2335, 1.9825, 2.1924],
       device='cuda:0')
2023-11-01 15:23:49,699 INFO [decode.py:579] The transcripts are stored in zipformer/exp-w-ctc-streaming/greedy_search/recogs-alimeeting_test-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:23:50,837 INFO [utils.py:565] [alimeeting_test-greedy_search] %WER 35.20% [73865 / 209845, 22160 ins, 10381 del, 41324 sub ]
2023-11-01 15:23:52,898 INFO [decode.py:592] Wrote detailed error stats to zipformer/exp-w-ctc-streaming/greedy_search/errs-alimeeting_test-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:23:52,903 INFO [decode.py:608] 
For alimeeting_test, WER of different settings are:
greedy_search	35.2	best for alimeeting_test

2023-11-01 15:23:52,904 INFO [decode.py:809] Start decoding test set: alimeeting_eval
2023-11-01 15:23:56,420 INFO [decode.py:563] batch 0/?, cuts processed until now is 35
2023-11-01 15:24:44,664 INFO [decode.py:563] batch 50/?, cuts processed until now is 3467
2023-11-01 15:25:01,871 INFO [decode.py:579] The transcripts are stored in zipformer/exp-w-ctc-streaming/greedy_search/recogs-alimeeting_eval-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:25:02,196 INFO [utils.py:565] [alimeeting_eval-greedy_search] %WER 33.78% [27396 / 81111, 8312 ins, 3583 del, 15501 sub ]
2023-11-01 15:25:02,844 INFO [decode.py:592] Wrote detailed error stats to zipformer/exp-w-ctc-streaming/greedy_search/errs-alimeeting_eval-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:25:02,848 INFO [decode.py:608] 
For alimeeting_eval, WER of different settings are:
greedy_search	33.78	best for alimeeting_eval

2023-11-01 15:25:02,849 INFO [decode.py:809] Start decoding test set: aishell_test
2023-11-01 15:25:05,568 INFO [decode.py:563] batch 0/?, cuts processed until now is 47
2023-11-01 15:25:37,400 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.9304, 1.4712, 1.8339, 2.8694], device='cuda:0')
2023-11-01 15:25:45,280 INFO [decode.py:563] batch 50/?, cuts processed until now is 2712
2023-11-01 15:26:15,114 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([2.5488, 1.3408, 1.9544, 1.8357, 2.0528, 1.8807, 2.1311, 1.8428],
       device='cuda:0')
2023-11-01 15:26:23,383 INFO [decode.py:563] batch 100/?, cuts processed until now is 5468
2023-11-01 15:26:34,730 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([2.6095, 1.9175, 2.3976, 3.4924], device='cuda:0')
2023-11-01 15:26:49,827 INFO [decode.py:579] The transcripts are stored in zipformer/exp-w-ctc-streaming/greedy_search/recogs-aishell_test-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:26:50,824 INFO [utils.py:565] [aishell_test-greedy_search] %WER 3.63% [3806 / 104765, 196 ins, 113 del, 3497 sub ]
2023-11-01 15:26:51,494 INFO [decode.py:592] Wrote detailed error stats to zipformer/exp-w-ctc-streaming/greedy_search/errs-aishell_test-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:26:51,498 INFO [decode.py:608] 
For aishell_test, WER of different settings are:
greedy_search	3.63	best for aishell_test

2023-11-01 15:26:51,498 INFO [decode.py:809] Start decoding test set: aishell_dev
2023-11-01 15:26:54,596 INFO [decode.py:563] batch 0/?, cuts processed until now is 53
2023-11-01 15:27:19,619 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.3457, 0.9338, 2.0064, 1.7320, 1.9318, 1.9831, 2.2106, 1.6197],
       device='cuda:0')
2023-11-01 15:27:20,353 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.7572, 1.9079, 1.9443, 1.7160, 1.7852, 1.7694, 1.5095, 1.7908],
       device='cuda:0')
2023-11-01 15:27:27,868 INFO [decode.py:563] batch 50/?, cuts processed until now is 3030
2023-11-01 15:27:55,773 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([2.7890, 1.6776, 2.2641, 2.8151], device='cuda:0')
2023-11-01 15:28:02,361 INFO [decode.py:563] batch 100/?, cuts processed until now is 6034
2023-11-01 15:28:15,406 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([4.4835, 3.9338, 2.2097, 4.1590], device='cuda:0')
2023-11-01 15:28:23,717 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.1489, 3.3002, 3.3014, 1.5704], device='cuda:0')
2023-11-01 15:28:38,285 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([4.4926, 3.9478, 2.2479, 4.1604], device='cuda:0')
2023-11-01 15:28:38,740 INFO [decode.py:563] batch 150/?, cuts processed until now is 9180
2023-11-01 15:28:43,249 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.9999, 2.0582, 2.1861, 2.8321], device='cuda:0')
2023-11-01 15:29:14,872 INFO [decode.py:563] batch 200/?, cuts processed until now is 12198
2023-11-01 15:29:28,640 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([2.1650, 1.3434, 2.2038, 2.2589, 2.1754, 2.3933, 2.6168, 2.1641],
       device='cuda:0')
2023-11-01 15:29:41,061 INFO [decode.py:579] The transcripts are stored in zipformer/exp-w-ctc-streaming/greedy_search/recogs-aishell_dev-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:29:41,787 INFO [utils.py:565] [aishell_dev-greedy_search] %WER 3.28% [6745 / 205341, 370 ins, 224 del, 6151 sub ]
2023-11-01 15:29:43,107 INFO [decode.py:592] Wrote detailed error stats to zipformer/exp-w-ctc-streaming/greedy_search/errs-aishell_dev-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:29:43,111 INFO [decode.py:608] 
For aishell_dev, WER of different settings are:
greedy_search	3.28	best for aishell_dev

2023-11-01 15:29:43,112 INFO [decode.py:809] Start decoding test set: aishell-2_test
2023-11-01 15:29:45,364 INFO [decode.py:563] batch 0/?, cuts processed until now is 83
2023-11-01 15:29:52,053 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([2.6256, 1.5043, 2.0287, 2.6426], device='cuda:0')
2023-11-01 15:30:06,335 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.0368, 2.6843, 2.0226, 1.5319], device='cuda:0')
2023-11-01 15:30:15,344 INFO [decode.py:563] batch 50/?, cuts processed until now is 4841
2023-11-01 15:30:18,510 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.1510, 1.8514, 1.8388, 1.6843, 1.4989, 1.8295, 1.8016, 1.5533],
       device='cuda:0')
2023-11-01 15:30:19,075 INFO [decode.py:579] The transcripts are stored in zipformer/exp-w-ctc-streaming/greedy_search/recogs-aishell-2_test-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:30:19,241 INFO [utils.py:565] [aishell-2_test-greedy_search] %WER 4.76% [2359 / 49532, 137 ins, 64 del, 2158 sub ]
2023-11-01 15:30:19,584 INFO [decode.py:592] Wrote detailed error stats to zipformer/exp-w-ctc-streaming/greedy_search/errs-aishell-2_test-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:30:19,588 INFO [decode.py:608] 
For aishell-2_test, WER of different settings are:
greedy_search	4.76	best for aishell-2_test

2023-11-01 15:30:19,588 INFO [decode.py:809] Start decoding test set: aishell-2_dev
2023-11-01 15:30:21,209 INFO [decode.py:563] batch 0/?, cuts processed until now is 81
2023-11-01 15:30:22,458 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.7206, 2.3090, 1.6930, 2.4259], device='cuda:0')
2023-11-01 15:30:39,005 INFO [decode.py:579] The transcripts are stored in zipformer/exp-w-ctc-streaming/greedy_search/recogs-aishell-2_dev-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:30:39,092 INFO [utils.py:565] [aishell-2_dev-greedy_search] %WER 4.34% [1077 / 24802, 62 ins, 31 del, 984 sub ]
2023-11-01 15:30:39,271 INFO [decode.py:592] Wrote detailed error stats to zipformer/exp-w-ctc-streaming/greedy_search/errs-aishell-2_dev-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:30:39,275 INFO [decode.py:608] 
For aishell-2_dev, WER of different settings are:
greedy_search	4.34	best for aishell-2_dev

2023-11-01 15:30:39,275 INFO [decode.py:809] Start decoding test set: aishell-4
2023-11-01 15:30:42,988 INFO [decode.py:563] batch 0/?, cuts processed until now is 33
2023-11-01 15:31:35,224 INFO [decode.py:563] batch 50/?, cuts processed until now is 2510
2023-11-01 15:31:57,421 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([5.1531, 4.6024, 2.4798, 4.9686], device='cuda:0')
2023-11-01 15:32:25,880 INFO [decode.py:563] batch 100/?, cuts processed until now is 5000
2023-11-01 15:32:54,517 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([2.2818, 2.1334, 2.5860, 2.8771], device='cuda:0')
2023-11-01 15:32:54,527 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([0.7411, 4.0579, 3.8294, 1.3754], device='cuda:0')
2023-11-01 15:33:08,784 INFO [decode.py:563] batch 150/?, cuts processed until now is 8249
2023-11-01 15:33:29,931 INFO [decode.py:579] The transcripts are stored in zipformer/exp-w-ctc-streaming/greedy_search/recogs-aishell-4-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:33:30,556 INFO [utils.py:565] [aishell-4-greedy_search] %WER 21.99% [39722 / 180665, 13506 ins, 3824 del, 22392 sub ]
2023-11-01 15:33:32,881 INFO [decode.py:592] Wrote detailed error stats to zipformer/exp-w-ctc-streaming/greedy_search/errs-aishell-4-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:33:32,887 INFO [decode.py:608] 
For aishell-4, WER of different settings are:
greedy_search	21.99	best for aishell-4

2023-11-01 15:33:32,888 INFO [decode.py:809] Start decoding test set: magicdata_test
2023-11-01 15:33:35,493 INFO [decode.py:563] batch 0/?, cuts processed until now is 57
2023-11-01 15:33:45,940 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([4.4134, 3.9453, 4.0638, 4.5766], device='cuda:0')
2023-11-01 15:33:49,742 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.9656, 2.2676, 2.1918, 2.2144, 2.0468, 2.2360, 2.3176, 1.9597],
       device='cuda:0')
2023-11-01 15:34:10,853 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([4.3569, 3.9111, 2.3004, 4.1492], device='cuda:0')
2023-11-01 15:34:11,876 INFO [decode.py:563] batch 50/?, cuts processed until now is 3245
2023-11-01 15:34:45,916 INFO [decode.py:563] batch 100/?, cuts processed until now is 6425
2023-11-01 15:35:18,166 INFO [decode.py:563] batch 150/?, cuts processed until now is 9787
2023-11-01 15:35:19,475 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.8893, 2.1565, 1.4386, 2.5974], device='cuda:0')
2023-11-01 15:35:51,293 INFO [decode.py:563] batch 200/?, cuts processed until now is 13211
2023-11-01 15:36:22,476 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.6952, 2.0973, 1.8146, 1.5153, 1.8879, 1.5104, 1.9810, 1.9615],
       device='cuda:0')
2023-11-01 15:36:27,467 INFO [decode.py:563] batch 250/?, cuts processed until now is 16395
2023-11-01 15:36:58,933 INFO [decode.py:563] batch 300/?, cuts processed until now is 20136
2023-11-01 15:36:59,063 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.8314, 2.1446, 1.5254, 2.6259], device='cuda:0')
2023-11-01 15:37:29,843 INFO [decode.py:563] batch 350/?, cuts processed until now is 23001
2023-11-01 15:37:30,469 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([3.8589, 3.3207, 2.0877, 2.2580], device='cuda:0')
2023-11-01 15:37:57,742 INFO [decode.py:579] The transcripts are stored in zipformer/exp-w-ctc-streaming/greedy_search/recogs-magicdata_test-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:37:59,301 INFO [utils.py:565] [magicdata_test-greedy_search] %WER 3.98% [9508 / 239091, 1607 ins, 331 del, 7570 sub ]
2023-11-01 15:38:00,957 INFO [decode.py:592] Wrote detailed error stats to zipformer/exp-w-ctc-streaming/greedy_search/errs-magicdata_test-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:38:00,962 INFO [decode.py:608] 
For magicdata_test, WER of different settings are:
greedy_search	3.98	best for magicdata_test

2023-11-01 15:38:00,962 INFO [decode.py:809] Start decoding test set: magicdata_dev
2023-11-01 15:38:04,084 INFO [decode.py:563] batch 0/?, cuts processed until now is 52
2023-11-01 15:38:15,577 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([4.3457, 4.0365, 2.3763, 4.1979], device='cuda:0')
2023-11-01 15:38:22,503 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([2.1163, 2.0328, 2.4166, 2.1896, 2.3399, 2.3751, 2.1030, 2.3192],
       device='cuda:0')
2023-11-01 15:38:25,117 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.7563, 2.2073, 1.9044, 1.6292, 1.9939, 1.5569, 2.0980, 2.0150],
       device='cuda:0')
2023-11-01 15:38:38,340 INFO [decode.py:563] batch 50/?, cuts processed until now is 2982
2023-11-01 15:39:12,080 INFO [decode.py:563] batch 100/?, cuts processed until now is 5919
2023-11-01 15:39:43,695 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([2.0203, 1.8654, 1.9170, 2.3662], device='cuda:0')
2023-11-01 15:39:44,916 INFO [decode.py:563] batch 150/?, cuts processed until now is 9079
2023-11-01 15:40:00,819 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([3.2241, 3.2611, 3.3693, 3.5728], device='cuda:0')
2023-11-01 15:40:17,983 INFO [decode.py:563] batch 200/?, cuts processed until now is 11646
2023-11-01 15:40:21,945 INFO [decode.py:579] The transcripts are stored in zipformer/exp-w-ctc-streaming/greedy_search/recogs-magicdata_dev-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:40:22,366 INFO [utils.py:565] [magicdata_dev-greedy_search] %WER 4.65% [5429 / 116800, 669 ins, 195 del, 4565 sub ]
2023-11-01 15:40:23,217 INFO [decode.py:592] Wrote detailed error stats to zipformer/exp-w-ctc-streaming/greedy_search/errs-magicdata_dev-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:40:23,221 INFO [decode.py:608] 
For magicdata_dev, WER of different settings are:
greedy_search	4.65	best for magicdata_dev

2023-11-01 15:40:23,222 INFO [decode.py:809] Start decoding test set: kespeech-asr_test
2023-11-01 15:40:25,923 INFO [decode.py:563] batch 0/?, cuts processed until now is 45
2023-11-01 15:40:26,013 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.5024, 2.9828, 2.4337, 2.0559], device='cuda:0')
2023-11-01 15:41:05,577 INFO [decode.py:563] batch 50/?, cuts processed until now is 2446
2023-11-01 15:41:43,337 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([4.8311, 4.3483, 2.7143, 4.6088], device='cuda:0')
2023-11-01 15:41:44,887 INFO [decode.py:563] batch 100/?, cuts processed until now is 4867
2023-11-01 15:42:09,880 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.1380, 1.9970, 1.7819, 1.6685, 1.5859, 1.7674, 1.9077, 1.2815],
       device='cuda:0')
2023-11-01 15:42:25,031 INFO [decode.py:563] batch 150/?, cuts processed until now is 7421
2023-11-01 15:42:54,421 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.9951, 2.5104, 2.0189, 2.6982], device='cuda:0')
2023-11-01 15:42:56,373 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([2.7823, 1.6581, 1.9859, 1.9323, 2.0820, 2.0876, 2.1005, 1.9966],
       device='cuda:0')
2023-11-01 15:43:05,117 INFO [decode.py:563] batch 200/?, cuts processed until now is 9965
2023-11-01 15:43:06,396 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([4.8379, 4.3883, 2.7998, 4.6366], device='cuda:0')
2023-11-01 15:43:18,507 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.9802, 2.5396, 2.1793, 2.3128, 2.1551, 2.1576, 2.3155, 1.7813],
       device='cuda:0')
2023-11-01 15:43:20,225 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.2670, 2.8465, 2.1663, 1.7759], device='cuda:0')
2023-11-01 15:43:37,797 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([0.9939, 3.6398, 3.5400, 1.5022], device='cuda:0')
2023-11-01 15:43:44,192 INFO [decode.py:563] batch 250/?, cuts processed until now is 12394
2023-11-01 15:44:20,196 INFO [decode.py:563] batch 300/?, cuts processed until now is 15124
2023-11-01 15:44:34,827 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.5672, 2.3577, 2.1580, 1.9644, 2.0547, 2.0573, 2.1825, 1.4833],
       device='cuda:0')
2023-11-01 15:44:59,164 INFO [decode.py:563] batch 350/?, cuts processed until now is 17516
2023-11-01 15:45:31,983 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([2.0662, 3.1975, 2.8471, 2.4349], device='cuda:0')
2023-11-01 15:45:34,769 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.9668, 2.0279, 2.2686, 2.6485], device='cuda:0')
2023-11-01 15:45:38,200 INFO [decode.py:563] batch 400/?, cuts processed until now is 19643
2023-11-01 15:45:40,067 INFO [decode.py:579] The transcripts are stored in zipformer/exp-w-ctc-streaming/greedy_search/recogs-kespeech-asr_test-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:45:41,239 INFO [utils.py:565] [kespeech-asr_test-greedy_search] %WER 12.54% [35571 / 283772, 2580 ins, 1518 del, 31473 sub ]
2023-11-01 15:45:43,137 INFO [decode.py:592] Wrote detailed error stats to zipformer/exp-w-ctc-streaming/greedy_search/errs-kespeech-asr_test-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:45:43,143 INFO [decode.py:608] 
For kespeech-asr_test, WER of different settings are:
greedy_search	12.54	best for kespeech-asr_test

2023-11-01 15:45:43,144 INFO [decode.py:809] Start decoding test set: kespeech-asr_dev_phase1
2023-11-01 15:45:46,208 INFO [decode.py:563] batch 0/?, cuts processed until now is 44
2023-11-01 15:46:20,420 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([4.0049, 3.2307, 2.1773, 2.1628], device='cuda:0')
2023-11-01 15:46:21,364 INFO [decode.py:579] The transcripts are stored in zipformer/exp-w-ctc-streaming/greedy_search/recogs-kespeech-asr_dev_phase1-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:46:21,480 INFO [utils.py:565] [kespeech-asr_dev_phase1-greedy_search] %WER 10.42% [3295 / 31634, 287 ins, 167 del, 2841 sub ]
2023-11-01 15:46:21,728 INFO [decode.py:592] Wrote detailed error stats to zipformer/exp-w-ctc-streaming/greedy_search/errs-kespeech-asr_dev_phase1-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:46:21,734 INFO [decode.py:608] 
For kespeech-asr_dev_phase1, WER of different settings are:
greedy_search	10.42	best for kespeech-asr_dev_phase1

2023-11-01 15:46:21,734 INFO [decode.py:809] Start decoding test set: kespeech-asr_dev_phase2
2023-11-01 15:46:23,188 INFO [decode.py:563] batch 0/?, cuts processed until now is 47
2023-11-01 15:46:55,432 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([4.8888, 4.4312, 2.6076, 4.7035], device='cuda:0')
2023-11-01 15:47:00,924 INFO [decode.py:579] The transcripts are stored in zipformer/exp-w-ctc-streaming/greedy_search/recogs-kespeech-asr_dev_phase2-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:47:01,081 INFO [utils.py:565] [kespeech-asr_dev_phase2-greedy_search] %WER 3.94% [1259 / 31928, 90 ins, 54 del, 1115 sub ]
2023-11-01 15:47:01,313 INFO [decode.py:592] Wrote detailed error stats to zipformer/exp-w-ctc-streaming/greedy_search/errs-kespeech-asr_dev_phase2-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:47:01,316 INFO [decode.py:608] 
For kespeech-asr_dev_phase2, WER of different settings are:
greedy_search	3.94	best for kespeech-asr_dev_phase2

2023-11-01 15:47:01,317 INFO [decode.py:809] Start decoding test set: wenetspeech-meeting_test
2023-11-01 15:47:04,192 INFO [decode.py:563] batch 0/?, cuts processed until now is 28
2023-11-01 15:47:57,608 INFO [decode.py:563] batch 50/?, cuts processed until now is 1884
2023-11-01 15:48:19,504 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([3.8353, 2.1305, 3.7659, 4.1871], device='cuda:0')
2023-11-01 15:48:58,096 INFO [decode.py:563] batch 100/?, cuts processed until now is 3776
2023-11-01 15:49:43,991 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.9118, 1.4416, 1.8381, 3.2348], device='cuda:0')
2023-11-01 15:49:50,508 INFO [decode.py:563] batch 150/?, cuts processed until now is 5887
2023-11-01 15:50:41,063 INFO [decode.py:563] batch 200/?, cuts processed until now is 8092
2023-11-01 15:50:50,425 INFO [decode.py:579] The transcripts are stored in zipformer/exp-w-ctc-streaming/greedy_search/recogs-wenetspeech-meeting_test-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:50:51,225 INFO [utils.py:565] [wenetspeech-meeting_test-greedy_search] %WER 9.73% [21438 / 220385, 4701 ins, 2165 del, 14572 sub ]
2023-11-01 15:50:52,605 INFO [decode.py:592] Wrote detailed error stats to zipformer/exp-w-ctc-streaming/greedy_search/errs-wenetspeech-meeting_test-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:50:52,609 INFO [decode.py:608] 
For wenetspeech-meeting_test, WER of different settings are:
greedy_search	9.73	best for wenetspeech-meeting_test

2023-11-01 15:50:52,609 INFO [decode.py:809] Start decoding test set: wenetspeech-net_test
2023-11-01 15:50:52,965 WARNING [decode.py:795] Excluding cut with ID: TEST_NET_Y0000000004_0ub4ZzdHzBc_S00023 from decoding, num_frames: 8
2023-11-01 15:50:56,216 INFO [decode.py:563] batch 0/?, cuts processed until now is 43
2023-11-01 15:51:38,788 INFO [decode.py:563] batch 50/?, cuts processed until now is 3497
2023-11-01 15:51:49,657 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([0.9579, 3.1483, 3.1855, 1.2969], device='cuda:0')
2023-11-01 15:52:18,468 INFO [decode.py:563] batch 100/?, cuts processed until now is 7009
2023-11-01 15:52:48,539 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([0.9854, 1.2879, 1.1733, 1.3545, 0.7731, 1.3166, 1.1098, 1.1947],
       device='cuda:0')
2023-11-01 15:53:05,656 INFO [decode.py:563] batch 150/?, cuts processed until now is 11030
2023-11-01 15:53:45,168 INFO [decode.py:563] batch 200/?, cuts processed until now is 14995
2023-11-01 15:53:58,796 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([3.2209, 1.6449, 1.9945, 3.2929], device='cuda:0')
2023-11-01 15:54:27,132 INFO [decode.py:563] batch 250/?, cuts processed until now is 18681
2023-11-01 15:55:03,366 INFO [decode.py:563] batch 300/?, cuts processed until now is 22693
2023-11-01 15:55:28,723 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([3.7829, 2.2991, 3.7043, 4.0532], device='cuda:0')
2023-11-01 15:55:30,578 INFO [decode.py:579] The transcripts are stored in zipformer/exp-w-ctc-streaming/greedy_search/recogs-wenetspeech-net_test-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:55:32,429 INFO [utils.py:565] [wenetspeech-net_test-greedy_search] %WER 9.66% [40161 / 415746, 4842 ins, 8154 del, 27165 sub ]
2023-11-01 15:55:36,170 INFO [decode.py:592] Wrote detailed error stats to zipformer/exp-w-ctc-streaming/greedy_search/errs-wenetspeech-net_test-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:55:36,176 INFO [decode.py:608] 
For wenetspeech-net_test, WER of different settings are:
greedy_search	9.66	best for wenetspeech-net_test

2023-11-01 15:55:36,177 INFO [decode.py:809] Start decoding test set: wenetspeech_dev
2023-11-01 15:55:39,027 INFO [decode.py:563] batch 0/?, cuts processed until now is 39
2023-11-01 15:56:22,888 INFO [decode.py:563] batch 50/?, cuts processed until now is 2499
2023-11-01 15:57:06,931 INFO [decode.py:563] batch 100/?, cuts processed until now is 4983
2023-11-01 15:57:27,232 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([4.7565, 3.9741, 4.4024, 4.7860], device='cuda:0')
2023-11-01 15:57:54,307 INFO [decode.py:563] batch 150/?, cuts processed until now is 7682
2023-11-01 15:58:40,854 INFO [decode.py:563] batch 200/?, cuts processed until now is 10268
2023-11-01 15:59:17,630 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([1.8558, 1.9246, 2.0962, 2.6251], device='cuda:0')
2023-11-01 15:59:25,763 INFO [decode.py:563] batch 250/?, cuts processed until now is 13004
2023-11-01 15:59:36,867 INFO [decode.py:579] The transcripts are stored in zipformer/exp-w-ctc-streaming/greedy_search/recogs-wenetspeech_dev-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:59:37,875 INFO [utils.py:565] [wenetspeech_dev-greedy_search] %WER 8.16% [26958 / 330498, 3014 ins, 9559 del, 14385 sub ]
2023-11-01 15:59:40,726 INFO [decode.py:592] Wrote detailed error stats to zipformer/exp-w-ctc-streaming/greedy_search/errs-wenetspeech_dev-greedy_search-epoch-20-avg-1-chunk-32-left-context-256-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-11-01 15:59:40,732 INFO [decode.py:608] 
For wenetspeech_dev, WER of different settings are:
greedy_search	8.16	best for wenetspeech_dev

2023-11-01 15:59:40,735 INFO [decode.py:826] Done!