File size: 8,349 Bytes
0a3b2c5 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 |
2023-06-15 11:40:32,789 INFO [decode.py:661] Decoding started
2023-06-15 11:40:32,789 INFO [decode.py:667] Device: cuda:0
2023-06-15 11:40:34,355 INFO [lexicon.py:168] Loading pre-compiled data/lang_char/Linv.pt
2023-06-15 11:40:34,616 INFO [decode.py:678] {'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': 6, 'iter': 0, 'avg': 1, '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-6-avg-1-modified_beam_search-beam-size-4-blank-penalty-2.0-use-averaged-model', 'blank_id': 0, 'vocab_size': 5537}
2023-06-15 11:40:34,616 INFO [decode.py:680] About to create model
2023-06-15 11:40:35,276 INFO [decode.py:747] Calculating the averaged model over epoch range from 5 (excluded) to 6
2023-06-15 11:40:45,498 INFO [decode.py:778] Number of model parameters: 75879898
2023-06-15 11:40:45,498 INFO [asr_datamodule.py:398] About to get dev cuts
2023-06-15 11:40:45,515 INFO [asr_datamodule.py:336] About to create dev dataset
2023-06-15 11:40:46,070 INFO [asr_datamodule.py:354] About to create dev dataloader
2023-06-15 11:40:46,071 INFO [asr_datamodule.py:403] About to get TEST_NET cuts
2023-06-15 11:40:46,083 INFO [asr_datamodule.py:367] About to create test dataset
2023-06-15 11:40:46,134 WARNING [decode.py:787] Exclude cut with ID TEST_NET_Y0000000004_0ub4ZzdHzBc_S00023 from decoding, num_frames : 8.
2023-06-15 11:40:46,652 INFO [asr_datamodule.py:408] About to get TEST_MEETING cuts
2023-06-15 11:40:46,659 INFO [asr_datamodule.py:367] About to create test dataset
2023-06-15 11:40:56,593 INFO [decode.py:557] batch 0/?, cuts processed until now is 130
2023-06-15 11:43:30,433 INFO [decode.py:557] batch 20/?, cuts processed until now is 3192
2023-06-15 11:46:03,811 INFO [decode.py:557] batch 40/?, cuts processed until now is 6421
2023-06-15 11:48:32,741 INFO [decode.py:557] batch 60/?, cuts processed until now is 10176
2023-06-15 11:50:34,863 INFO [decode.py:557] batch 80/?, cuts processed until now is 13727
2023-06-15 11:50:43,316 INFO [decode.py:573] The transcripts are stored in zipformer/exp_L_context_2/modified_beam_search/recogs-DEV-beam_size_4_blank_penalty_2.0-epoch-6-avg-1-modified_beam_search-beam-size-4-blank-penalty-2.0-use-averaged-model.txt
2023-06-15 11:50:43,710 INFO [utils.py:562] [DEV-beam_size_4_blank_penalty_2.0] %WER 7.62% [25198 / 330498, 2645 ins, 10239 del, 12314 sub ]
2023-06-15 11:50:44,735 INFO [decode.py:586] Wrote detailed error stats to zipformer/exp_L_context_2/modified_beam_search/errs-DEV-beam_size_4_blank_penalty_2.0-epoch-6-avg-1-modified_beam_search-beam-size-4-blank-penalty-2.0-use-averaged-model.txt
2023-06-15 11:50:44,738 INFO [decode.py:602]
For DEV, WER of different settings are:
beam_size_4_blank_penalty_2.0 7.62 best for DEV
2023-06-15 11:50:45,004 WARNING [decode.py:787] Exclude cut with ID TEST_NET_Y0000000004_0ub4ZzdHzBc_S00023 from decoding, num_frames : 8.
2023-06-15 11:50:54,313 INFO [decode.py:557] batch 0/?, cuts processed until now is 146
2023-06-15 11:53:13,713 INFO [decode.py:557] batch 20/?, cuts processed until now is 4116
2023-06-15 11:54:55,405 INFO [zipformer.py:1728] name=None, attn_weights_entropy = tensor([2.0992, 2.4837, 2.1656, 2.6700, 3.0641, 3.2619, 3.1530, 2.6695],
device='cuda:0')
2023-06-15 11:55:38,606 INFO [decode.py:557] batch 40/?, cuts processed until now is 8601
2023-06-15 11:55:44,356 INFO [zipformer.py:1728] name=None, attn_weights_entropy = tensor([3.7238, 3.2068, 1.8849, 1.8835], device='cuda:0')
2023-06-15 11:57:59,597 INFO [decode.py:557] batch 60/?, cuts processed until now is 14082
2023-06-15 12:00:22,741 INFO [decode.py:557] batch 80/?, cuts processed until now is 18750
2023-06-15 12:01:02,708 INFO [zipformer.py:1728] name=None, attn_weights_entropy = tensor([1.1849, 1.6059, 2.4053, 2.4054], device='cuda:0')
2023-06-15 12:02:09,686 INFO [decode.py:557] batch 100/?, cuts processed until now is 24487
2023-06-15 12:02:24,637 INFO [decode.py:573] The transcripts are stored in zipformer/exp_L_context_2/modified_beam_search/recogs-TEST_NET-beam_size_4_blank_penalty_2.0-epoch-6-avg-1-modified_beam_search-beam-size-4-blank-penalty-2.0-use-averaged-model.txt
2023-06-15 12:02:25,145 INFO [utils.py:562] [TEST_NET-beam_size_4_blank_penalty_2.0] %WER 8.14% [33828 / 415746, 3611 ins, 8022 del, 22195 sub ]
2023-06-15 12:02:26,497 INFO [decode.py:586] Wrote detailed error stats to zipformer/exp_L_context_2/modified_beam_search/errs-TEST_NET-beam_size_4_blank_penalty_2.0-epoch-6-avg-1-modified_beam_search-beam-size-4-blank-penalty-2.0-use-averaged-model.txt
2023-06-15 12:02:26,500 INFO [decode.py:602]
For TEST_NET, WER of different settings are:
beam_size_4_blank_penalty_2.0 8.14 best for TEST_NET
2023-06-15 12:02:35,913 INFO [decode.py:557] batch 0/?, cuts processed until now is 93
2023-06-15 12:05:03,600 INFO [decode.py:557] batch 20/?, cuts processed until now is 2345
2023-06-15 12:07:35,564 INFO [decode.py:557] batch 40/?, cuts processed until now is 4929
2023-06-15 12:07:43,131 INFO [zipformer.py:1728] name=None, attn_weights_entropy = tensor([1.8584, 3.2715, 4.4633, 4.5427], device='cuda:0')
2023-06-15 12:09:40,941 INFO [decode.py:557] batch 60/?, cuts processed until now is 7955
2023-06-15 12:10:10,608 INFO [decode.py:573] The transcripts are stored in zipformer/exp_L_context_2/modified_beam_search/recogs-TEST_MEETING-beam_size_4_blank_penalty_2.0-epoch-6-avg-1-modified_beam_search-beam-size-4-blank-penalty-2.0-use-averaged-model.txt
2023-06-15 12:10:10,874 INFO [utils.py:562] [TEST_MEETING-beam_size_4_blank_penalty_2.0] %WER 13.06% [28790 / 220385, 2961 ins, 13112 del, 12717 sub ]
2023-06-15 12:10:11,576 INFO [decode.py:586] Wrote detailed error stats to zipformer/exp_L_context_2/modified_beam_search/errs-TEST_MEETING-beam_size_4_blank_penalty_2.0-epoch-6-avg-1-modified_beam_search-beam-size-4-blank-penalty-2.0-use-averaged-model.txt
2023-06-15 12:10:11,580 INFO [decode.py:602]
For TEST_MEETING, WER of different settings are:
beam_size_4_blank_penalty_2.0 13.06 best for TEST_MEETING
2023-06-15 12:10:11,580 INFO [decode.py:825] Done!
|