ESPnet2 pretrained model, Shinji Watanabe/librispeech_asr_train_asr_conformer_raw_bpe_batch_bins30000000_accum_grad3_optim_conflr0.001_sp_valid.acc.ave, fs=16k, lang=en

This model was trained by Shinji Watanabe using librispeech recipe in espnet. Python API See https://github.com/espnet/espnet_model_zoo Evaluate in the recipe git clone https://github.com/espnet/espnet cd espnet git checkout eda997f9e97ad959c6b13df1b34eb24fb8c52768 pip install -e . cd egs2/librispeech/asr1 ./run.sh --skip_data_prep false --skip_train true --download_model Shinji Watanabe/librispeech_asr_train_asr_conformer_raw_bpe_batch_bins30000000_accum_grad3_optim_conflr0.001_sp_valid.acc.ave Results # RESULTS ## Environments - date: `Mon Nov 16 18:59:34 EST 2020` - python version: `3.7.3 (default, Mar 27 2019, 22:11:17) [GCC 7.3.0]` - espnet version: `espnet 0.9.2` - pytorch version: `pytorch 1.4.0` - Git hash: `eda997f9e97ad959c6b13df1b34eb24fb8c52768` - Commit date: `Thu Oct 8 07:32:49 2020 -0400` ## asr_train_asr_conformer_raw_bpe_batch_bins30000000_accum_grad3_optim_conflr0.001_sp ### WER |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err| |---|---|---|---|---|---|---|---|---| |decode_asr_lm_lm_train_lm_adam_bpe_valid.loss.ave_asr_model_valid.acc.ave/dev_clean|2703|54402|98.0|1.8|0.2|0.3|2.3|27.6| |decode_asr_lm_lm_train_lm_adam_bpe_valid.loss.ave_asr_model_valid.acc.ave/dev_other|2864|50948|95.0|4.3|0.6|0.5|5.5|45.1| |decode_asr_lm_lm_train_lm_adam_bpe_valid.loss.ave_asr_model_valid.acc.ave/test_clean|2620|52576|97.9|1.9|0.2|0.3|2.4|28.2| |decode_asr_lm_lm_train_lm_adam_bpe_valid.loss.ave_asr_model_valid.acc.ave/test_other|2939|52343|94.9|4.5|0.7|0.6|5.8|48.6| ### CER |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err| |---|---|---|---|---|---|---|---|---| |decode_asr_lm_lm_train_lm_adam_bpe_valid.loss.ave_asr_model_valid.acc.ave/dev_clean|2703|288456|99.3|0.3|0.3|0.2|0.9|27.6| |decode_asr_lm_lm_train_lm_adam_bpe_valid.loss.ave_asr_model_valid.acc.ave/dev_other|2864|265951|98.0|1.0|1.0|0.6|2.6|45.1| |decode_asr_lm_lm_train_lm_adam_bpe_valid.loss.ave_asr_model_valid.acc.ave/test_clean|2620|281530|99.4|0.3|0.3|0.2|0.9|28.2| |decode_asr_lm_lm_train_lm_adam_bpe_valid.loss.ave_asr_model_valid.acc.ave/test_other|2939|272758|98.0|1.0|1.0|0.6|2.6|48.6| ### TER |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err| |---|---|---|---|---|---|---|---|---| |decode_asr_lm_lm_train_lm_adam_bpe_valid.loss.ave_asr_model_valid.acc.ave/dev_clean|2703|69307|97.3|1.7|1.1|0.4|3.1|27.6| |decode_asr_lm_lm_train_lm_adam_bpe_valid.loss.ave_asr_model_valid.acc.ave/dev_other|2864|64239|93.7|4.1|2.2|0.9|7.2|45.1| |decode_asr_lm_lm_train_lm_adam_bpe_valid.loss.ave_asr_model_valid.acc.ave/test_clean|2620|66712|97.2|1.7|1.1|0.3|3.1|28.2| |decode_asr_lm_lm_train_lm_adam_bpe_valid.loss.ave_asr_model_valid.acc.ave/test_other|2939|66329|93.4|4.1|2.5|0.7|7.3|48.6| ASR config config: conf/tuning/train_asr_conformer.yaml print_config: false log_level: INFO dry_run: false iterator_type: sequence output_dir: exp/asr_train_asr_conformer_raw_bpe_batch_bins30000000_accum_grad3_optim_conflr0.001_sp ngpu: 1 seed: 0 num_workers: 1 num_att_plot: 3 dist_backend: nccl dist_init_method: env:// dist_world_size: 4 dist_rank: 0 local_rank: 0 dist_master_addr: localhost dist_master_port: 40183 dist_launcher: null multiprocessing_distributed: true cudnn_enabled: true cudnn_benchmark: false cudnn_deterministic: true collect_stats: false write_collected_feats: false max_epoch: 60 patience: null val_scheduler_criterion: - valid - loss early_stopping_criterion: - valid - loss - min best_model_criterion: - - valid - acc - max keep_nbest_models: 10 grad_clip: 5.0 grad_clip_type: 2.0 grad_noise: false accum_grad: 3 no_forward_run: false resume: true train_dtype: float32 use_amp: false log_interval: null pretrain_path: [] pretrain_key: [] num_iters_per_epoch: null batch_size: 20 valid_batch_size: null batch_bins: 30000000 valid_batch_bins: null train_shape_file: - exp/asr_stats_raw_sp/train/speech_shape - exp/asr_stats_raw_sp/train/text_shape.bpe valid_shape_file: - exp/asr_stats_raw_sp/valid/speech_shape - exp/asr_stats_raw_sp/valid/text_shape.bpe batch_type: numel valid_batch_type: null fold_length: - 80000 - 150 sort_in_batch: descending sort_batch: descending multiple_iterator: false chunk_length: 500 chunk_shift_ratio: 0.5 num_cache_chunks: 1024 train_data_path_and_name_and_type: - - dump/raw/train_960_sp/wav.scp - speech - sound - - dump/raw/train_960_sp/text - text - text valid_data_path_and_name_and_type: - - dump/raw/dev/wav.scp - speech - sound - - dump/raw/dev/text - text - text allow_variable_data_keys: false max_cache_size: 0.0 max_cache_fd: 32 valid_max_cache_size: null optim: adam optim_conf: lr: 0.001 scheduler: warmuplr scheduler_conf: warmup_steps: 25000 token_list: - - - "\u2581EXCLAIM" - "\u2581OPPORTUNITIES" - "\u2581REMEDY" - "\u2581DEFENSE" - "\u2581ETERNITY" - "\u2581SKULL" - "\u2581PLEADED" - "\u2581INSTINCTIVELY" - "\u2581SLAUGHTER" - "\u2581RATIONAL" - "\u2581PULSE" - "\u2581PARALLEL" - "\u2581SCOUNDREL" - "\u2581PRUDENCE" - "\u2581PROBABILITY" - "\u2581DERIVED" - "\u2581MONSTROUS" - "\u2581POTATOES" - "\u2581IMPRESSIVE" - "\u2581DAINTY" - "\u2581SULTAN" - "\u2581CARPENTER" - "\u2581INNUMERABLE" - "\u2581INVITE" - "\u2581CIRCULAT" - "\u2581ELOQUENCE" - "\u2581DISCIPLE" - "\u2581ATTIRE" - "\u2581OBSTINATE" - "\u2581STREAK" - "\u2581WOLVES" - "\u2581GRINNED" - "\u2581ORCHARD" - "\u2581JAPANESE" - "\u2581ANNUAL" - "\u2581SHAWL" - "\u2581PACIFIC" - "\u2581VEHICLE" - "\u2581APOSTLE" - "\u2581CONGREGATION" - "\u2581AMAZING" - "\u2581OCCURRENCE" - "\u2581CONFERENCE" - "\u2581MIXTURE" - "\u2581EXAMINING" - "\u2581SAUCE" - "\u2581ADMIRING" - "\u2581AMBASSADOR" - "\u2581DEVICE" - "\u2581INCAPABLE" - "\u2581WHEREUPON" - "\u2581IMPERFECT" - "\u2581PERCEPTION" - "\u2581LOUNG" - "\u2581VACANT" - "\u2581EXCURSION" - "\u2581DISCOURAGE" - "\u2581FANTASTIC" - "\u2581REBELLION" - "\u2581CONVINCE" - "\u2581DEFIANCE" - "\u2581CONNECT" - "\u2581EMPHASIS" - "\u2581MEXICO" - "\u2581OPPONENT" - "\u2581DETERMINE" - "\u2581MANUSCRIPT" - "\u2581INCESSANT" - "\u2581BRONZE" - "\u2581COURTEOUS" - "\u2581COFFIN" - "\u2581CONSTRUCTION" - "\u2581BLUNDER" - "\u2581SENATE" - "\u2581CIRCUM" - "\u2581DANIEL" - "\u2581ELOQUENT" - "\u2581QUIVERING" - "\u2581VIGIL" - "\u2581HAZARD" - "\u2581UNWORTHY" - "\u2581TAB" - "\u2581ILLUSION" - "\u2581AGITATED" - "\u2581CHAMPION" - "\u2581DIMINISH" - "\u2581STUMP" - "\u2581CONFIDE" - "\u2581PHILADELPHIA" - "\u2581DOUGLAS" - "\u2581BUMP" - "\u2581COMMERCE" - "\u2581FACULTY" - "\u2581CONFEDERATE" - "\u2581EMBARRASSMENT" - "\u2581EXPLORE" - "\u2581MAGGIE" - "\u2581PHILOSOPHIC" - "\u2581ADMINISTRATION" - "\u2581HEADQUARTERS" - "\u2581SOLUTION" - "\u2581REFRAIN" - "\u2581ELDEST" - "\u2581FORMIDABLE" - "\u2581VERANDA" - "\u2581DISMAL" - "\u2581ESTHER" - "\u2581PRUDENT" - "\u2581BLAZING" - "\u2581RESOLVE" - "\u2581ELSIE" - "\u2581TURKEY" - "\u2581DECREE" - "\u2581CONVERSE" - "\u2581GRAVITY" - "\u2581MIRTH" - "\u2581RESEMBLANCE" - "\u2581GULF" - "\u2581SHRUB" - "\u2581EXHIBITION" - "\u2581AUSTRALIA" - "\u2581ELEANOR" - "\u2581UNCOMMON" - "\u2581RACHEL" - "\u2581TOMORROW" - "\u2581INJUSTICE" - "\u2581WISTFUL" - "\u2581WREATH" - "\u2581DISDAIN" - "\u2581CRUMB" - "\u2581CLINGING" - "\u2581COMMEND" - "\u2581SUPERSTITION" - "\u2581CRISIS" - "\u2581MAXIM" - "\u2581DESIRABLE" - "\u2581GIGANTIC" - "\u2581JUNGLE" - "\u2581DIGNIFIED" - "\u2581INVALID" - "\u2581UNNECESSARY" - "\u2581SUBLIME" - "\u2581PLOUGH" - "\u2581SUFFICE" - "\u2581BUNK" - "\u2581LUNCHEON" - "\u2581DRAUGHT" - "\u2581COLONY" - "\u2581PARLOUR" - "\u2581TERRIFIED" - "\u2581LOATH" - "\u2581SIGNIFICANCE" - "\u2581EXTENSIVE" - "\u2581HORACE" - "\u2581SERENE" - "\u2581CHEESE" - "\u2581PRECEDING" - "\u2581LEVI" - "\u2581INVARIABLY" - "\u2581OBSERVING" - "\u2581EARLIEST" - "\u2581WHEAT" - "\u2581DEMOCRAT" - "\u2581YOURSELVES" - "\u2581FEMININE" - "\u2581ARTIFICIAL" - "\u2581IDIOT" - "\u2581TORRENT" - "\u2581CONVICT" - "\u2581CONSUME" - "\u2581EMBROIDER" - "\u2581CONQUEST" - "\u2581CALCULATED" - "\u2581HAPPIER" - "\u2581DECAY" - "\u2581LITERALLY" - "\u2581RADIANT" - ENNI - "\u2581AMAZED" - "\u2581SPLIT" - "\u2581SUPPOSING" - "\u2581CANADA" - "\u2581PAVEMENT" - "\u2581ANTHONY" - "\u2581BULK" - "\u2581MEDIUM" - "\u2581MAURICE" - "\u2581SALOON" - "\u2581BARRIER" - "\u2581SWORE" - GUARD - "\u2581TEMPORARY" - "\u2581STALK" - "\u2581IRREGULAR" - "\u2581FRANTIC" - "\u2581BLISS" - "\u2581CONSPICUOUS" - "\u2581GERALD" - "\u2581EXCITING" - "\u2581SMASH" - "\u2581EXTERNAL" - "\u2581HESITATE" - "\u2581PATHETIC" - "\u2581NINTH" - "\u2581HAMILTON" - "\u2581UNSEEN" - "\u2581DEFECT" - "\u2581ACCURATE" - "\u2581LIQUOR" - "\u2581ENLIGHTEN" - WICH - "\u2581CLARK" - "\u2581REVERSE" - "\u2581PRIMITIVE" - "\u2581BLUFF" - "\u2581PRECAUTION" - "\u2581AWHILE" - "\u2581SPOON" - "\u2581TEMPERAMENT" - "\u2581SCHOONER" - "\u2581FREDERICK" - "\u2581REMORSE" - "\u2581CUSHION" - "\u2581EXCLUSIVE" - "\u2581CONTEMPLATE" - "\u2581SYLVIA" - "\u2581SITUATED" - "\u2581SKIPPER" - "\u2581DESOLATE" - "\u2581WORRIED" - "\u2581DWELT" - "\u2581TROUSERS" - "\u2581MARTYR" - "\u2581RIV" - "\u2581MEXICAN" - "\u2581DIVIDE" - "\u2581AMIABLE" - "\u2581PRUSSIA" - "\u2581COMMERCIAL" - "\u2581CONFRONT" - MPTON - "\u2581PROSPERITY" - "\u2581FEBRUARY" - "\u2581ADJUST" - "\u2581FUGITIVE" - "\u2581ABUNDANT" - "\u2581CRUELTY" - "\u2581UNCOMFORTABLE" - "\u2581THUMB" - "\u2581LAWRENCE" - "\u2581INTERCOURSE" - "\u2581STUDIO" - "\u2581PROMINENT" - "\u2581CHARLOTTE" - SHAW - "\u2581BRETHREN" - "\u2581COMPLEXION" - "\u2581TRAITOR" - "\u2581UNJUST" - BOAT - "\u2581TICK" - "\u2581KNELT" - "\u2581PROPRIETOR" - "\u2581ELABORATE" - "\u2581WRIT" - "\u2581OBSTACLE" - "\u2581CONSOLATION" - "\u2581DETAIN" - "\u2581FLANK" - "\u2581SAXON" - "\u2581SCRIPTURE" - "\u2581PERFUME" - "\u2581TRAGIC" - "\u2581EGYPTIAN" - "\u2581FOWL" - Q - "\u2581NOWHERE" - ARIA - "\u2581ATTENTIVE" - "\u2581STAIRCASE" - "\u2581BRITAIN" - "\u2581NORMAL" - "\u2581INFLICT" - "\u2581ECONOMI" - "\u2581OXFORD" - OTTE - "\u2581TUMULT

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.