icefall-aishell-pruned-transducer-stateless3-2022-06-20
/
decoding_results
/greedy_search
/log-decode-epoch-29-avg-5-context-1-max-sym-per-frame-1-use-averaged-model-2022-06-19-12-27-02
2022-06-19 12:27:02,918 INFO [decode.py:497] Decoding started | |
2022-06-19 12:27:02,919 INFO [decode.py:503] Device: cuda:0 | |
2022-06-19 12:27:03,155 INFO [lexicon.py:176] Loading pre-compiled data/lang_char/Linv.pt | |
2022-06-19 12:27:03,183 INFO [decode.py:509] {'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': 1000, 'feature_dim': 80, 'subsampling_factor': 4, 'model_warm_step': 3000, 'env_info': {'k2-version': '1.15.1', 'k2-build-type': 'Release', 'k2-with-cuda': True, 'k2-git-sha1': 'f8d2dba06c000ffee36aab5b66f24e7c9809f116', 'k2-git-date': 'Thu Apr 21 12:20:34 2022', 'lhotse-version': '1.3.0.dev+missing.version.file', 'torch-version': '1.10.0+cu102', 'torch-cuda-available': True, 'torch-cuda-version': '10.2', 'python-version': '3.8', 'icefall-git-branch': 'pruned-rnnt-aishell', 'icefall-git-sha1': 'd0a5f1d-dirty', 'icefall-git-date': 'Mon Jun 13 20:40:46 2022', 'icefall-path': '/k2-dev/fangjun/open-source/icefall-aishell', 'k2-path': '/ceph-fj/fangjun/open-source-2/k2-multi-22/k2/python/k2/__init__.py', 'lhotse-path': '/ceph-fj/fangjun/open-source-2/lhotse-jsonl/lhotse/__init__.py', 'hostname': 'de-74279-k2-train-7-0616225511-78bf4545d8-tv52r', 'IP address': '10.177.77.9'}, 'epoch': 29, 'iter': 0, 'avg': 5, 'use_averaged_model': True, 'exp_dir': PosixPath('pruned_transducer_stateless3/exp-context-size-1'), 'lang_dir': PosixPath('data/lang_char'), 'decoding_method': 'greedy_search', 'beam_size': 4, 'beam': 4, 'max_contexts': 4, 'max_states': 8, 'context_size': 1, 'max_sym_per_frame': 1, 'num_encoder_layers': 12, 'dim_feedforward': 2048, 'nhead': 8, 'encoder_dim': 512, 'decoder_dim': 512, 'joiner_dim': 512, 'max_duration': 600, 'bucketing_sampler': True, 'num_buckets': 30, 'shuffle': True, 'return_cuts': True, 'num_workers': 2, 'enable_spec_aug': True, 'spec_aug_time_warp_factor': 80, 'enable_musan': True, 'manifest_dir': PosixPath('data/fbank'), 'on_the_fly_feats': False, 'res_dir': PosixPath('pruned_transducer_stateless3/exp-context-size-1/greedy_search'), 'suffix': 'epoch-29-avg-5-context-1-max-sym-per-frame-1-use-averaged-model', 'blank_id': 0, 'vocab_size': 4336} | |
2022-06-19 12:27:03,184 INFO [decode.py:511] About to create model | |
2022-06-19 12:27:03,987 INFO [decode.py:583] Calculating the averaged model over epoch range from 24 (excluded) to 29 | |
2022-06-19 12:27:11,513 INFO [decode.py:606] Number of model parameters: 96983734 | |
2022-06-19 12:27:11,514 INFO [aishell.py:51] About to get test cuts from data/fbank/aishell_cuts_test.jsonl.gz | |
2022-06-19 12:27:11,519 INFO [aishell.py:45] About to get valid cuts from data/fbank/aishell_cuts_dev.jsonl.gz | |
2022-06-19 12:27:13,818 INFO [decode.py:403] batch 0/?, cuts processed until now is 99 | |
2022-06-19 12:27:35,470 INFO [decode.py:403] batch 50/?, cuts processed until now is 5795 | |
2022-06-19 12:27:39,918 INFO [decode.py:420] The transcripts are stored in pruned_transducer_stateless3/exp-context-size-1/greedy_search/recogs-test-greedy_search-epoch-29-avg-5-context-1-max-sym-per-frame-1-use-averaged-model.txt | |
2022-06-19 12:27:40,257 INFO [utils.py:408] [test-greedy_search] %WER 5.39% [5644 / 104765, 159 ins, 835 del, 4650 sub ] | |
2022-06-19 12:27:40,773 INFO [decode.py:437] Wrote detailed error stats to pruned_transducer_stateless3/exp-context-size-1/greedy_search/errs-test-greedy_search-epoch-29-avg-5-context-1-max-sym-per-frame-1-use-averaged-model.txt | |
2022-06-19 12:27:40,774 INFO [decode.py:454] | |
For test, CER of different settings are: | |
greedy_search 5.39 best for test | |
2022-06-19 12:27:42,272 INFO [decode.py:403] batch 0/?, cuts processed until now is 111 | |
2022-06-19 12:28:03,745 INFO [decode.py:403] batch 50/?, cuts processed until now is 6472 | |
2022-06-19 12:28:24,542 INFO [decode.py:403] batch 100/?, cuts processed until now is 13159 | |
2022-06-19 12:28:28,444 INFO [decode.py:420] The transcripts are stored in pruned_transducer_stateless3/exp-context-size-1/greedy_search/recogs-dev-greedy_search-epoch-29-avg-5-context-1-max-sym-per-frame-1-use-averaged-model.txt | |
2022-06-19 12:28:29,125 INFO [utils.py:408] [dev-greedy_search] %WER 5.09% [10461 / 205341, 315 ins, 1465 del, 8681 sub ] | |
2022-06-19 12:28:30,183 INFO [decode.py:437] Wrote detailed error stats to pruned_transducer_stateless3/exp-context-size-1/greedy_search/errs-dev-greedy_search-epoch-29-avg-5-context-1-max-sym-per-frame-1-use-averaged-model.txt | |
2022-06-19 12:28:30,183 INFO [decode.py:454] | |
For dev, CER of different settings are: | |
greedy_search 5.09 best for dev | |
2022-06-19 12:28:30,198 INFO [decode.py:633] Done! | |