2024-09-26 10:19:26,383 INFO [ctc_decode_with_stats.py:784] Decoding started
2024-09-26 10:19:26,384 INFO [ctc_decode_with_stats.py:790] Device: cuda:0
2024-09-26 10:19:26,384 INFO [ctc_decode_with_stats.py:791] {'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, 'ignore_id': -1, 'label_smoothing': 0.1, 'warm_step': 2000, 'env_info': {'k2-version': '1.24.4', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': '44a9d5682af9fd3ef77074777e15278ec6d390eb', 'k2-git-date': 'Wed Sep 27 11:22:55 2023', 'lhotse-version': '1.17.0.dev+git.ccfc5b2c.dirty', 'torch-version': '1.10.0+cu102', 'torch-cuda-available': True, 'torch-cuda-version': '10.2', 'python-version': '3.8', 'icefall-git-branch': 'cr-ctc', 'icefall-git-sha1': 'a6eead6c-clean', 'icefall-git-date': 'Mon Sep 9 10:10:08 2024', 'icefall-path': '/star-zw/workspace/zipformer/icefall_cr_ctc', 'k2-path': '/star-zw/workspace/k2/k2/k2/python/k2/__init__.py', 'lhotse-path': '/star-zw/workspace/lhotse/lhotse/lhotse/__init__.py', 'hostname': 'de-74279-k2-train-5-0904174858-644c565fdb-d6n2c', 'IP address': '10.30.46.238'}, 'frame_shift_ms': 10, 'search_beam': 20, 'output_beam': 8, 'min_active_states': 30, 'max_active_states': 10000, 'use_double_scores': True, 'epoch': 50, 'iter': 0, 'avg': 24, 'use_averaged_model': True, 'exp_dir': PosixPath('zipformer/exp-cr-loss-scale-0.2-time-mask-ratio-2.5-scaled-masked-1-4'), 'bpe_model': 'data/lang_bpe_500/bpe.model', 'lang_dir': PosixPath('data/lang_bpe_500'), 'context_size': 2, 'decoding_method': 'ctc-greedy-search', 'num_paths': 100, 'nbest_scale': 1.0, 'hlg_scale': 0.6, 'lm_dir': PosixPath('data/lm'), '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, 'attention_decoder_dim': 512, 'attention_decoder_num_layers': 6, 'attention_decoder_attention_dim': 512, 'attention_decoder_num_heads': 8, 'attention_decoder_feedforward_dim': 2048, 'causal': False, 'chunk_size': '16,32,64,-1', 'left_context_frames': '64,128,256,-1', 'use_transducer': False, 'use_ctc': True, 'use_attention_decoder': False, 'use_cr_ctc': True, 'full_libri': True, 'mini_libri': False, 'manifest_dir': PosixPath('data/fbank'), 'max_duration': 300, '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-cr-loss-scale-0.2-time-mask-ratio-2.5-scaled-masked-1-4/ctc-greedy-search'), 'suffix': 'epoch-50-avg-24-use-averaged-model'}
2024-09-26 10:19:26,734 INFO [lexicon.py:168] Loading pre-compiled data/lang_bpe_500/Linv.pt
2024-09-26 10:19:31,369 INFO [ctc_decode_with_stats.py:876] About to create model
2024-09-26 10:19:31,995 INFO [ctc_decode_with_stats.py:943] Calculating the averaged model over epoch range from 26 (excluded) to 50
2024-09-26 10:19:34,611 INFO [ctc_decode_with_stats.py:960] Number of model parameters: 64250603
2024-09-26 10:19:34,611 INFO [asr_datamodule.py:467] About to get test-clean cuts
2024-09-26 10:19:34,748 INFO [asr_datamodule.py:474] About to get test-other cuts
2024-09-26 10:19:35,471 INFO [ctc_decode_with_stats.py:689] batch 0/?, cuts processed until now is 21
2024-09-26 10:19:36,831 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([5.6097, 4.6140, 5.3613, 5.0853], device='cuda:0')
2024-09-26 10:19:45,624 INFO [ctc_decode_with_stats.py:712] The transcripts are stored in zipformer/exp-cr-loss-scale-0.2-time-mask-ratio-2.5-scaled-masked-1-4/ctc-greedy-search/recogs-test-clean-epoch-50-avg-24-use-averaged-model.txt
2024-09-26 10:19:45,887 INFO [utils.py:657] [test-clean-ctc-greedy-search] %WER 2.12% [1116 / 52576, 113 ins, 85 del, 918 sub ]
2024-09-26 10:19:46,101 INFO [ctc_decode_with_stats.py:721] Wrote detailed error stats to zipformer/exp-cr-loss-scale-0.2-time-mask-ratio-2.5-scaled-masked-1-4/ctc-greedy-search/errs-test-clean-epoch-50-avg-24-use-averaged-model.txt
2024-09-26 10:19:46,106 INFO [ctc_decode_with_stats.py:735] 
For test-clean, WER of different settings are:
ctc-greedy-search	2.12	best for test-clean

2024-09-26 10:19:46,589 INFO [ctc_decode_with_stats.py:689] batch 0/?, cuts processed until now is 26
2024-09-26 10:19:51,927 INFO [zipformer.py:1858] name=None, attn_weights_entropy = tensor([3.3742, 3.0918, 2.6845, 2.9870], device='cuda:0')
2024-09-26 10:19:56,892 INFO [ctc_decode_with_stats.py:712] The transcripts are stored in zipformer/exp-cr-loss-scale-0.2-time-mask-ratio-2.5-scaled-masked-1-4/ctc-greedy-search/recogs-test-other-epoch-50-avg-24-use-averaged-model.txt
2024-09-26 10:19:57,025 INFO [utils.py:657] [test-other-ctc-greedy-search] %WER 4.62% [2416 / 52343, 235 ins, 197 del, 1984 sub ]
2024-09-26 10:19:57,306 INFO [ctc_decode_with_stats.py:721] Wrote detailed error stats to zipformer/exp-cr-loss-scale-0.2-time-mask-ratio-2.5-scaled-masked-1-4/ctc-greedy-search/errs-test-other-epoch-50-avg-24-use-averaged-model.txt
2024-09-26 10:19:57,312 INFO [ctc_decode_with_stats.py:735] 
For test-other, WER of different settings are:
ctc-greedy-search	4.62	best for test-other

2024-09-26 10:19:57,313 INFO [ctc_decode_with_stats.py:1000] Averaged non-blank duration=1.284059427189012
2024-09-26 10:19:57,317 INFO [ctc_decode_with_stats.py:1003] Averaged non-blank probs=0.8942277432591206
2024-09-26 10:19:57,323 INFO [ctc_decode_with_stats.py:1006] Averaged blank probs=0.9419012471026017
2024-09-26 10:19:57,323 INFO [ctc_decode_with_stats.py:1009] Done!