decoding_result/greedy_search/log-decode-epoch-34-avg-19-context-2-max-sym-per-frame-1-use-averaged-model-2023-10-05-17-02-28

2023-10-05 17:02:28,745 INFO [decode.py:688] Decoding started
2023-10-05 17:02:28,746 INFO [decode.py:694] Device: cuda:0
2023-10-05 17:02:28,753 INFO [decode.py:704] {'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/bilingual', 'icefall-git-sha1': '4897f2c0-dirty', 'icefall-git-date': 'Thu Sep 28 11:38:28 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': 34, 'iter': 0, 'avg': 19, 'use_averaged_model': True, 'exp_dir': PosixPath('zipformer/exp-w-tal-csasr'), 'bpe_model': 'data/lang_bbpe_2000/bbpe.model', 'lang_dir': PosixPath('data/lang_bbpe_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, 'use_tal_csasr': False, 'use_librispeech': False, 'use_aishell2': False, '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': False, 'chunk_size': '16,32,64,-1', 'left_context_frames': '64,128,256,-1', 'use_transducer': True, 'use_ctc': False, '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-tal-csasr/greedy_search'), 'suffix': 'epoch-34-avg-19-context-2-max-sym-per-frame-1-use-averaged-model', 'blank_id': 0, 'unk_id': 2, 'vocab_size': 2000}
2023-10-05 17:02:28,753 INFO [decode.py:706] About to create model
2023-10-05 17:02:29,595 INFO [decode.py:773] Calculating the averaged model over epoch range from 15 (excluded) to 34
2023-10-05 17:02:43,800 INFO [decode.py:807] Number of model parameters: 68625511
2023-10-05 17:02:43,800 INFO [multi_dataset.py:142] About to get multidataset test cuts
2023-10-05 17:02:43,800 INFO [multi_dataset.py:145] Loading Aishell-2 set in lazy mode
2023-10-05 17:02:43,819 INFO [multi_dataset.py:211] About to get test-clean cuts
2023-10-05 17:02:43,825 INFO [multi_dataset.py:218] About to get test-other cuts
2023-10-05 17:02:43,828 INFO [multi_dataset.py:157] Loading TAL-CSASR set in lazy mode
2023-10-05 17:02:46,342 INFO [decode.py:831] Start decoding test set: aishell-2_test
2023-10-05 17:02:47,933 INFO [decode.py:585] batch 0/?, cuts processed until now is 83
2023-10-05 17:03:06,003 INFO [zipformer.py:1853] name=None, attn_weights_entropy = tensor([3.3845, 3.1985, 2.7148, 2.7275], device='cuda:0')
2023-10-05 17:03:07,309 INFO [decode.py:585] batch 50/?, cuts processed until now is 4841
2023-10-05 17:03:09,972 INFO [decode.py:601] The transcripts are stored in zipformer/exp-w-tal-csasr/greedy_search/recogs-aishell-2_test-greedy_search-epoch-34-avg-19-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-10-05 17:03:10,142 INFO [utils.py:564] [aishell-2_test-greedy_search] %WER 7.03% [3482 / 49532, 41 ins, 392 del, 3049 sub ]
2023-10-05 17:03:10,471 INFO [decode.py:614] Wrote detailed error stats to zipformer/exp-w-tal-csasr/greedy_search/errs-aishell-2_test-greedy_search-epoch-34-avg-19-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-10-05 17:03:10,475 INFO [decode.py:630] 
For aishell-2_test, WER of different settings are:
greedy_search	7.03	best for aishell-2_test

2023-10-05 17:03:10,475 INFO [decode.py:831] Start decoding test set: aishell-2_dev
2023-10-05 17:03:11,609 INFO [decode.py:585] batch 0/?, cuts processed until now is 81
2023-10-05 17:03:23,274 INFO [decode.py:601] The transcripts are stored in zipformer/exp-w-tal-csasr/greedy_search/recogs-aishell-2_dev-greedy_search-epoch-34-avg-19-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-10-05 17:03:23,355 INFO [utils.py:564] [aishell-2_dev-greedy_search] %WER 6.57% [1630 / 24802, 18 ins, 174 del, 1438 sub ]
2023-10-05 17:03:23,524 INFO [decode.py:614] Wrote detailed error stats to zipformer/exp-w-tal-csasr/greedy_search/errs-aishell-2_dev-greedy_search-epoch-34-avg-19-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-10-05 17:03:23,528 INFO [decode.py:630] 
For aishell-2_dev, WER of different settings are:
greedy_search	6.57	best for aishell-2_dev

2023-10-05 17:03:23,528 INFO [decode.py:831] Start decoding test set: librispeech_test_clean
2023-10-05 17:03:25,148 INFO [decode.py:585] batch 0/?, cuts processed until now is 22
2023-10-05 17:04:01,148 INFO [decode.py:585] batch 50/?, cuts processed until now is 1656
2023-10-05 17:04:19,094 INFO [decode.py:601] The transcripts are stored in zipformer/exp-w-tal-csasr/greedy_search/recogs-librispeech_test_clean-greedy_search-epoch-34-avg-19-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-10-05 17:04:19,263 INFO [utils.py:564] [librispeech_test_clean-greedy_search] %WER 2.43% [1277 / 52576, 141 ins, 103 del, 1033 sub ]
2023-10-05 17:04:19,603 INFO [decode.py:614] Wrote detailed error stats to zipformer/exp-w-tal-csasr/greedy_search/errs-librispeech_test_clean-greedy_search-epoch-34-avg-19-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-10-05 17:04:19,607 INFO [decode.py:630] 
For librispeech_test_clean, WER of different settings are:
greedy_search	2.43	best for librispeech_test_clean

2023-10-05 17:04:19,607 INFO [decode.py:831] Start decoding test set: librispeech_test_other
2023-10-05 17:04:21,250 INFO [decode.py:585] batch 0/?, cuts processed until now is 26
2023-10-05 17:04:59,132 INFO [decode.py:585] batch 50/?, cuts processed until now is 1884
2023-10-05 17:05:14,731 INFO [decode.py:601] The transcripts are stored in zipformer/exp-w-tal-csasr/greedy_search/recogs-librispeech_test_other-greedy_search-epoch-34-avg-19-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-10-05 17:05:14,895 INFO [utils.py:564] [librispeech_test_other-greedy_search] %WER 5.70% [2981 / 52343, 287 ins, 265 del, 2429 sub ]
2023-10-05 17:05:15,255 INFO [decode.py:614] Wrote detailed error stats to zipformer/exp-w-tal-csasr/greedy_search/errs-librispeech_test_other-greedy_search-epoch-34-avg-19-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-10-05 17:05:15,259 INFO [decode.py:630] 
For librispeech_test_other, WER of different settings are:
greedy_search	5.7	best for librispeech_test_other

2023-10-05 17:05:15,259 INFO [decode.py:831] Start decoding test set: tal_csasr_test
2023-10-05 17:05:17,761 INFO [decode.py:585] batch 0/?, cuts processed until now is 27
2023-10-05 17:05:29,437 INFO [zipformer.py:1853] name=None, attn_weights_entropy = tensor([5.8078, 5.5299, 5.2011, 5.0578], device='cuda:0')
2023-10-05 17:05:49,394 INFO [zipformer.py:1853] name=None, attn_weights_entropy = tensor([3.7014, 3.4656, 3.3183, 3.2987], device='cuda:0')
2023-10-05 17:05:49,799 INFO [zipformer.py:1853] name=None, attn_weights_entropy = tensor([1.6391, 2.8705, 4.1981, 2.5853], device='cuda:0')
2023-10-05 17:05:56,840 INFO [decode.py:585] batch 50/?, cuts processed until now is 2106
2023-10-05 17:06:33,953 INFO [decode.py:585] batch 100/?, cuts processed until now is 4226
2023-10-05 17:06:47,179 INFO [zipformer.py:1853] name=None, attn_weights_entropy = tensor([2.1115, 2.8246, 3.5938, 3.3386], device='cuda:0')
2023-10-05 17:06:50,556 INFO [zipformer.py:1853] name=None, attn_weights_entropy = tensor([5.6247, 5.5525, 5.4680, 4.9361], device='cuda:0')
2023-10-05 17:07:00,353 INFO [zipformer.py:1853] name=None, attn_weights_entropy = tensor([4.2325, 4.5643, 5.5795, 5.4254], device='cuda:0')
2023-10-05 17:07:06,771 INFO [decode.py:585] batch 150/?, cuts processed until now is 6623
2023-10-05 17:07:42,385 INFO [decode.py:585] batch 200/?, cuts processed until now is 8979
2023-10-05 17:08:21,579 INFO [decode.py:585] batch 250/?, cuts processed until now is 11092
2023-10-05 17:08:54,857 INFO [decode.py:585] batch 300/?, cuts processed until now is 13660
2023-10-05 17:09:14,273 INFO [decode.py:601] The transcripts are stored in zipformer/exp-w-tal-csasr/greedy_search/recogs-tal_csasr_test-greedy_search-epoch-34-avg-19-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-10-05 17:09:15,312 INFO [utils.py:564] [tal_csasr_test-greedy_search] %WER 6.80% [22796 / 335124, 3750 ins, 5415 del, 13631 sub ]
2023-10-05 17:09:17,424 INFO [decode.py:614] Wrote detailed error stats to zipformer/exp-w-tal-csasr/greedy_search/errs-tal_csasr_test-greedy_search-epoch-34-avg-19-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-10-05 17:09:17,428 INFO [decode.py:630] 
For tal_csasr_test, WER of different settings are:
greedy_search	6.8	best for tal_csasr_test

2023-10-05 17:09:17,429 INFO [decode.py:831] Start decoding test set: tal_csasr_dev
2023-10-05 17:09:19,645 INFO [decode.py:585] batch 0/?, cuts processed until now is 26
2023-10-05 17:09:58,549 INFO [decode.py:585] batch 50/?, cuts processed until now is 2076
2023-10-05 17:10:03,175 INFO [zipformer.py:1853] name=None, attn_weights_entropy = tensor([4.8078, 3.0778, 4.6410, 4.4266], device='cuda:0')
2023-10-05 17:10:24,059 INFO [zipformer.py:1853] name=None, attn_weights_entropy = tensor([1.8053, 2.2310, 2.7751, 2.7537], device='cuda:0')
2023-10-05 17:10:34,741 INFO [decode.py:585] batch 100/?, cuts processed until now is 4445
2023-10-05 17:10:41,813 INFO [decode.py:601] The transcripts are stored in zipformer/exp-w-tal-csasr/greedy_search/recogs-tal_csasr_dev-greedy_search-epoch-34-avg-19-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-10-05 17:10:42,138 INFO [utils.py:564] [tal_csasr_dev-greedy_search] %WER 6.73% [7670 / 113947, 1284 ins, 1810 del, 4576 sub ]
2023-10-05 17:10:42,816 INFO [decode.py:614] Wrote detailed error stats to zipformer/exp-w-tal-csasr/greedy_search/errs-tal_csasr_dev-greedy_search-epoch-34-avg-19-context-2-max-sym-per-frame-1-use-averaged-model.txt
2023-10-05 17:10:42,820 INFO [decode.py:630] 
For tal_csasr_dev, WER of different settings are:
greedy_search	6.73	best for tal_csasr_dev

2023-10-05 17:10:42,820 INFO [decode.py:848] Done!