config: conf/train_asr_wav2vec_small.yaml print_config: false log_level: INFO dry_run: false iterator_type: sequence output_dir: exp/asr_train_asr_wav2vec_small_raw_en_us_bpe300_sp ngpu: 1 seed: 0 num_workers: 1 num_att_plot: 3 dist_backend: nccl dist_init_method: env:// dist_world_size: null dist_rank: null local_rank: 0 dist_master_addr: null dist_master_port: null dist_launcher: null multiprocessing_distributed: false unused_parameters: true sharded_ddp: false cudnn_enabled: true cudnn_benchmark: false cudnn_deterministic: true collect_stats: false write_collected_feats: false max_epoch: 150 patience: null val_scheduler_criterion: - valid - loss early_stopping_criterion: - valid - loss - min best_model_criterion: - - valid - acc - max keep_nbest_models: 10 nbest_averaging_interval: 0 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 use_matplotlib: true use_tensorboard: true use_wandb: false wandb_project: null wandb_id: null wandb_entity: null wandb_name: null wandb_model_log_interval: -1 detect_anomaly: false pretrain_path: null init_param: [] ignore_init_mismatch: false freeze_param: - frontend.upstream num_iters_per_epoch: null batch_size: 20 valid_batch_size: null batch_bins: 1600000 valid_batch_bins: null train_shape_file: - exp/asr_stats_raw_en_us_bpe300_sp/train/speech_shape - exp/asr_stats_raw_en_us_bpe300_sp/train/text_shape.bpe valid_shape_file: - exp/asr_stats_raw_en_us_bpe300_sp/valid/speech_shape - exp/asr_stats_raw_en_us_bpe300_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_en_us_sp/wav.scp - speech - sound - - dump/raw/train_en_us_sp/text - text - text valid_data_path_and_name_and_type: - - dump/raw/dev_en_us/wav.scp - speech - sound - - dump/raw/dev_en_us/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: 40000 token_list: - - - S - E - T - ▁ - A - ▁THE - D - I - U - R - O - ▁A - M - C - N - Y - P - ▁S - ER - ED - ING - G - ▁IN - ▁TO - ▁OF - AL - L - AR - ▁AND - EN - IN - ▁C - AN - V - B - ▁F - W - OR - H - K - F - ▁BE - RE - ▁B - LY - CH - ▁IS - UR - ▁RE - LE - IC - ES - IT - 'ON' - ▁G - ST - ▁ON - ENT - ▁FOR - CE - ▁T - ATION - IL - ▁M - ▁W - ▁E - ▁ST - ▁ARE - ▁H - ▁THAT - ▁MA - TH - AM - ▁IT - '''' - IR - ▁N - LI - ▁OR - VER - UN - '-' - ▁YOU - ▁WITH - ▁P - RO - ▁CO - ▁WAS - UL - ID - ▁TH - OL - ▁DE - TION - ▁AN - ▁CON - ATE - ▁AS - ET - NE - ▁HE - ▁O - ▁1 - OM - IVE - CK - SE - ▁HO - ▁DI - ▁LE - J - EL - ▁CAN - ▁FROM - ION - ▁PO - ▁AT - IF - VE - ▁PRO - LA - ▁BY - LL - '0' - ▁HAVE - ▁WHI - US - ▁EX - ▁UN - IS - IGHT - ▁MO - X - ▁PA - ITY - ▁WE - ▁DO - ▁US - EST - ▁LO - IST - ▁THIS - ▁BO - MENT - AND - ▁SP - ERS - ▁SOME - ▁RO - ▁K - '9' - ▁NOT - IES - ▁WHE - ▁LA - ▁VI - OP - ▁2 - ▁NO - ACT - ▁THEY - ▁HAS - ▁WERE - LAND - Z - OUS - ▁ALL - ▁ME - . - AGE - ▁RA - ▁TRA - '5' - '00' - DAY - ▁WILL - '4' - ATED - ▁OTHER - THER - ▁BUT - ▁PRE - IAN - ABLE - ISH - ▁MORE - '1' - ANCE - ▁HAD - '6' - ▁MOST - LSO - ▁TIME - ▁PER - MANY - ICAL - '3' - ▁COMP - '8' - IZ - URE - ILL - ▁CAR - ▁PH - ▁GO - ▁HIS - ▁CHA - ▁EVEN - WAY - ▁PEOPLE - ▁NEW - ▁OVER - ▁DR - ▁AFTER - ▁JU - OSE - ▁PART - ▁RI - ▁PRI - ▁WHO - ▁WORLD - TAIN - '7' - '2' - ▁UP - ▁YEAR - ▁FIRST - QUE - OOD - LONG - ▁DA - ▁ABOUT - ▁SAID - ▁OFTEN - ▁KNOW - ▁INTO - ▁APP - CAUSE - ▁TRAVEL - ▁COUNTRY - ▁DUR - ',' - ▁WOULD - ▁LIKE - ▁WORK - ▁TRI - ▁DIFFERENT - ▁TWO - THOUGH - ▁MAKE - ▁JO - ▁NORTH - ▁PLACE - ▁SHOULD - ▁WATER - ▁SOUTH - ▁REPORT - ▁INCLUD - ▁IMP - ▁COUNTRIES - ▁SMALL - ▁PLAN - ▁REQUIRE - ▁AROUND - ▁TAKE - / - ':' - '"' - '%' - $ - á - í - — - ; - ü - é - + - '&' - '[' - ']' - ° - ç - – - ” - ã - ’ - £ - ó - õ - ú - ¥ - '!' - Q - ö - İ - ł - ‘ - init: null input_size: null ctc_conf: dropout_rate: 0.0 ctc_type: builtin reduce: true ignore_nan_grad: true joint_net_conf: null use_preprocessor: true token_type: bpe bpemodel: data/en_us_token_list/bpe_unigram300/bpe.model non_linguistic_symbols: null cleaner: null g2p: null speech_volume_normalize: null rir_scp: null rir_apply_prob: 1.0 noise_scp: null noise_apply_prob: 1.0 noise_db_range: '13_15' frontend: s3prl frontend_conf: frontend_conf: upstream: wav2vec2_large_960 download_dir: ./hub multilayer_feature: true fs: 16k specaug: specaug specaug_conf: apply_time_warp: true time_warp_window: 5 time_warp_mode: bicubic apply_freq_mask: true freq_mask_width_range: - 0 - 30 num_freq_mask: 2 apply_time_mask: true time_mask_width_range: - 0 - 40 num_time_mask: 2 normalize: utterance_mvn normalize_conf: {} model: espnet model_conf: ctc_weight: 0.3 lsm_weight: 0.1 length_normalized_loss: false extract_feats_in_collect_stats: false preencoder: linear preencoder_conf: input_size: 1024 output_size: 80 encoder: transformer encoder_conf: output_size: 256 attention_heads: 4 linear_units: 2048 num_blocks: 12 dropout_rate: 0.1 positional_dropout_rate: 0.1 attention_dropout_rate: 0.0 input_layer: conv2d normalize_before: true postencoder: null postencoder_conf: {} decoder: transformer decoder_conf: attention_heads: 4 linear_units: 2048 num_blocks: 6 dropout_rate: 0.1 positional_dropout_rate: 0.1 self_attention_dropout_rate: 0.0 src_attention_dropout_rate: 0.0 required: - output_dir - token_list version: '202205' distributed: false