2023-06-15 14:39:07,470 INFO [decode.py:639] Decoding started
2023-06-15 14:39:07,470 INFO [decode.py:645] Device: cuda:0
2023-06-15 14:39:09,072 INFO [lexicon.py:168] Loading pre-compiled data/lang_char/Linv.pt
2023-06-15 14:39:09,331 INFO [decode.py:656] {'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': '1.24.1', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': 'c51a0b9684442a88ee37f3ce0af686a04b66855b', 'k2-git-date': 'Mon May 1 21:38:03 2023', 'lhotse-version': '1.14.0.dev+git.0f812851.dirty', 'torch-version': '1.10.0+cu102', 'torch-cuda-available': True, 'torch-cuda-version': '10.2', 'python-version': '3.8', 'icefall-git-branch': 'zipformer_wenetspeech', 'icefall-git-sha1': '28d3f6d-dirty', 'icefall-git-date': 'Thu Jun 15 10:30:34 2023', 'icefall-path': '/star-kw/kangwei/code/icefall_wenetspeech', 'k2-path': '/ceph-hw/kangwei/code/k2_release/k2/k2/python/k2/__init__.py', 'lhotse-path': '/ceph-hw/kangwei/dev_tools/anaconda3/envs/rnnt2/lib/python3.8/site-packages/lhotse-1.14.0.dev0+git.0f812851.dirty-py3.8.egg/lhotse/__init__.py', 'hostname': 'de-74279-k2-train-10-0221105906-5745685d6b-t8zzx', 'IP address': '10.177.57.19'}, 'epoch': 12, 'iter': 0, 'avg': 4, 'use_averaged_model': True, 'exp_dir': PosixPath('zipformer/exp_L_context_2'), 'lang_dir': PosixPath('data/lang_char'), 'decoding_method': 'modified_beam_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, 'blank_penalty': 2.0, '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', 'manifest_dir': PosixPath('data/fbank'), 'max_duration': 1000, 'bucketing_sampler': True, 'num_buckets': 30, 'concatenate_cuts': False, 'duration_factor': 1.0, 'gap': 1.0, 'on_the_fly_feats': False, 'shuffle': True, 'return_cuts': True, 'num_workers': 2, 'enable_spec_aug': True, 'spec_aug_time_warp_factor': 80, 'enable_musan': True, 'training_subset': 'L', 'res_dir': PosixPath('zipformer/exp_L_context_2/modified_beam_search'), 'suffix': 'epoch-12-avg-4-modified_beam_search-beam-size-4-blank-penalty-2.0-use-averaged-model', 'blank_id': 0, 'vocab_size': 5537}
2023-06-15 14:39:09,331 INFO [decode.py:658] About to create model
2023-06-15 14:39:10,067 INFO [decode.py:725] Calculating the averaged model over epoch range from 8 (excluded) to 12
2023-06-15 14:39:21,286 INFO [decode.py:756] Number of model parameters: 75879898
2023-06-15 14:39:21,286 INFO [asr_datamodule.py:398] About to get dev cuts
2023-06-15 14:39:21,300 INFO [asr_datamodule.py:336] About to create dev dataset
2023-06-15 14:39:21,870 INFO [asr_datamodule.py:354] About to create dev dataloader
2023-06-15 14:39:21,871 INFO [asr_datamodule.py:403] About to get TEST_NET cuts
2023-06-15 14:39:21,873 INFO [asr_datamodule.py:367] About to create test dataset
2023-06-15 14:39:21,924 WARNING [decode.py:765] Exclude cut with ID TEST_NET_Y0000000004_0ub4ZzdHzBc_S00023 from decoding, num_frames : 8.
2023-06-15 14:39:22,451 INFO [asr_datamodule.py:408] About to get TEST_MEETING cuts
2023-06-15 14:39:22,459 INFO [asr_datamodule.py:367] About to create test dataset
2023-06-15 14:39:32,543 INFO [decode.py:536] batch 0/?, cuts processed until now is 130
2023-06-15 14:42:07,613 INFO [decode.py:536] batch 20/?, cuts processed until now is 3192
2023-06-15 14:44:39,669 INFO [decode.py:536] batch 40/?, cuts processed until now is 6421
2023-06-15 14:47:09,561 INFO [decode.py:536] batch 60/?, cuts processed until now is 10176
2023-06-15 14:49:08,809 INFO [decode.py:536] batch 80/?, cuts processed until now is 13727
2023-06-15 14:49:16,349 INFO [decode.py:552] The transcripts are stored in zipformer/exp_L_context_2/modified_beam_search/recogs-DEV-beam_size_4_blank_penalty_2.0-epoch-12-avg-4-modified_beam_search-beam-size-4-blank-penalty-2.0-use-averaged-model.txt
2023-06-15 14:49:16,722 INFO [utils.py:562] [DEV-beam_size_4_blank_penalty_2.0] %WER 7.32% [24180 / 330498, 2991 ins, 9738 del, 11451 sub ]
2023-06-15 14:49:17,743 INFO [decode.py:565] Wrote detailed error stats to zipformer/exp_L_context_2/modified_beam_search/errs-DEV-beam_size_4_blank_penalty_2.0-epoch-12-avg-4-modified_beam_search-beam-size-4-blank-penalty-2.0-use-averaged-model.txt
2023-06-15 14:49:17,747 INFO [decode.py:581] 
For DEV, WER of different settings are:
beam_size_4_blank_penalty_2.0	7.32	best for DEV

2023-06-15 14:49:18,132 WARNING [decode.py:765] Exclude cut with ID TEST_NET_Y0000000004_0ub4ZzdHzBc_S00023 from decoding, num_frames : 8.
2023-06-15 14:49:27,537 INFO [decode.py:536] batch 0/?, cuts processed until now is 146
2023-06-15 14:51:09,310 INFO [zipformer.py:1728] name=None, attn_weights_entropy = tensor([3.9867, 4.0293, 1.9069, 3.7287], device='cuda:0')
2023-06-15 14:51:47,503 INFO [decode.py:536] batch 20/?, cuts processed until now is 4116
2023-06-15 14:53:29,615 INFO [zipformer.py:1728] name=None, attn_weights_entropy = tensor([2.0150, 3.3764, 2.6014, 4.2142], device='cuda:0')
2023-06-15 14:54:12,801 INFO [decode.py:536] batch 40/?, cuts processed until now is 8601
2023-06-15 14:54:54,261 INFO [zipformer.py:1728] name=None, attn_weights_entropy = tensor([4.7249, 4.0098, 2.2175, 2.1477], device='cuda:0')
2023-06-15 14:56:35,004 INFO [decode.py:536] batch 60/?, cuts processed until now is 14082
2023-06-15 14:58:59,317 INFO [decode.py:536] batch 80/?, cuts processed until now is 18750
2023-06-15 15:00:46,706 INFO [decode.py:536] batch 100/?, cuts processed until now is 24487
2023-06-15 15:01:01,722 INFO [decode.py:552] The transcripts are stored in zipformer/exp_L_context_2/modified_beam_search/recogs-TEST_NET-beam_size_4_blank_penalty_2.0-epoch-12-avg-4-modified_beam_search-beam-size-4-blank-penalty-2.0-use-averaged-model.txt
2023-06-15 15:01:02,228 INFO [utils.py:562] [TEST_NET-beam_size_4_blank_penalty_2.0] %WER 7.61% [31635 / 415746, 4259 ins, 7303 del, 20073 sub ]
2023-06-15 15:01:03,563 INFO [decode.py:565] Wrote detailed error stats to zipformer/exp_L_context_2/modified_beam_search/errs-TEST_NET-beam_size_4_blank_penalty_2.0-epoch-12-avg-4-modified_beam_search-beam-size-4-blank-penalty-2.0-use-averaged-model.txt
2023-06-15 15:01:03,566 INFO [decode.py:581] 
For TEST_NET, WER of different settings are:
beam_size_4_blank_penalty_2.0	7.61	best for TEST_NET

2023-06-15 15:01:12,722 INFO [decode.py:536] batch 0/?, cuts processed until now is 93
2023-06-15 15:03:42,004 INFO [decode.py:536] batch 20/?, cuts processed until now is 2345
2023-06-15 15:06:18,416 INFO [decode.py:536] batch 40/?, cuts processed until now is 4929
2023-06-15 15:07:27,506 INFO [zipformer.py:1728] name=None, attn_weights_entropy = tensor([2.3657, 2.7457, 2.4217, 2.8425, 3.0775, 3.1380, 3.1068, 2.8610],
       device='cuda:0')
2023-06-15 15:08:21,381 INFO [decode.py:536] batch 60/?, cuts processed until now is 7955
2023-06-15 15:08:51,193 INFO [decode.py:552] The transcripts are stored in zipformer/exp_L_context_2/modified_beam_search/recogs-TEST_MEETING-beam_size_4_blank_penalty_2.0-epoch-12-avg-4-modified_beam_search-beam-size-4-blank-penalty-2.0-use-averaged-model.txt
2023-06-15 15:08:51,478 INFO [utils.py:562] [TEST_MEETING-beam_size_4_blank_penalty_2.0] %WER 12.35% [27219 / 220385, 3372 ins, 12413 del, 11434 sub ]
2023-06-15 15:08:52,172 INFO [decode.py:565] Wrote detailed error stats to zipformer/exp_L_context_2/modified_beam_search/errs-TEST_MEETING-beam_size_4_blank_penalty_2.0-epoch-12-avg-4-modified_beam_search-beam-size-4-blank-penalty-2.0-use-averaged-model.txt
2023-06-15 15:08:52,177 INFO [decode.py:581] 
For TEST_MEETING, WER of different settings are:
beam_size_4_blank_penalty_2.0	12.35	best for TEST_MEETING

2023-06-15 15:08:52,177 INFO [decode.py:801] Done!