ESPnet2 pretrained model, kamo-naoyuki/mini_an4_asr_train_raw_bpe_valid.acc.best, fs=16k, lang=en

This model was trained by kamo-naoyuki using mini_an4 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 2a994ca15dba7b4b97c3608e1105e1ec52921384
  pip install -e .
  cd egs2/mini_an4/asr1
  ./run.sh --skip_data_prep false --skip_train true --download_model kamo-naoyuki/mini_an4_asr_train_raw_bpe_valid.acc.best
  

• Results

  
  # RESULTS
  ## Environments
  - date: `Thu Jul 23 03:18:58 JST 2020`
  - python version: `3.7.3 (default, Mar 27 2019, 22:11:17)  [GCC 7.3.0]`
  - espnet version: `espnet 0.8.0`
  - pytorch version: `pytorch 1.0.1`
  - Git hash: `31794e2aec89e6159c1ac32643d4c09989200a30`
    - Commit date: `Thu Jul 23 03:17:45 2020 +0900`
  
  ## asr_train_raw_bpe
  ### WER
  
  |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
  |---|---|---|---|---|---|---|---|---|
  |decode_test_decode_lm_train_bpe_valid.loss.best_asr_model_valid.acc.best|2|10|0.0|20.0|80.0|0.0|100.0|100.0|
  |decode_test_seg_decode_lm_train_bpe_valid.loss.best_asr_model_valid.acc.best|2|10|0.0|20.0|80.0|0.0|100.0|100.0|
  |decode_train_dev_decode_lm_train_bpe_valid.loss.best_asr_model_valid.acc.best|1|1|0.0|100.0|0.0|0.0|100.0|100.0|
  
  ### CER
  
  |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
  |---|---|---|---|---|---|---|---|---|
  |decode_test_decode_lm_train_bpe_valid.loss.best_asr_model_valid.acc.best|2|67|53.7|44.8|1.5|206.0|252.2|100.0|
  |decode_test_seg_decode_lm_train_bpe_valid.loss.best_asr_model_valid.acc.best|2|67|53.7|44.8|1.5|206.0|252.2|100.0|
  |decode_train_dev_decode_lm_train_bpe_valid.loss.best_asr_model_valid.acc.best|1|3|33.3|66.7|0.0|1200.0|1266.7|100.0|
  
  ### TER
  
  |dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
  |---|---|---|---|---|---|---|---|---|
  |decode_test_decode_lm_train_bpe_valid.loss.best_asr_model_valid.acc.best|2|45|15.6|82.2|2.2|208.9|293.3|100.0|
  |decode_test_seg_decode_lm_train_bpe_valid.loss.best_asr_model_valid.acc.best|2|45|15.6|82.2|2.2|208.9|293.3|100.0|
  |decode_train_dev_decode_lm_train_bpe_valid.loss.best_asr_model_valid.acc.best|1|4|25.0|75.0|0.0|550.0|625.0|100.0|

• ASR config

  config: null
  print_config: false
  log_level: INFO
  dry_run: false
  iterator_type: sequence
  output_dir: exp/asr_train_raw_bpe
  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
  cudnn_enabled: true
  cudnn_benchmark: false
  cudnn_deterministic: true
  collect_stats: false
  write_collected_feats: false
  max_epoch: 40
  patience: null
  val_scheduler_criterion:
  - valid
  - loss
  early_stopping_criterion:
  - valid
  - loss
  - min
  best_model_criterion:
  -   - train
      - loss
      - min
  -   - valid
      - loss
      - min
  -   - train
      - acc
      - max
  -   - valid
      - acc
      - max
  keep_nbest_models: 10
  grad_clip: 5.0
  grad_noise: false
  accum_grad: 1
  no_forward_run: false
  resume: true
  train_dtype: float32
  log_interval: null
  pretrain_path: []
  pretrain_key: []
  num_iters_per_epoch: null
  batch_size: 20
  valid_batch_size: null
  batch_bins: 1000000
  valid_batch_bins: null
  train_shape_file:
  - exp/asr_stats_raw/train/speech_shape
  - exp/asr_stats_raw/train/text_shape.bpe
  valid_shape_file:
  - exp/asr_stats_raw/valid/speech_shape
  - exp/asr_stats_raw/valid/text_shape.bpe
  batch_type: folded
  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_nodev/wav.scp
      - speech
      - sound
  -   - dump/raw/train_nodev/text
      - text
      - text
  valid_data_path_and_name_and_type:
  -   - dump/raw/train_dev/wav.scp
      - speech
      - sound
  -   - dump/raw/train_dev/text
      - text
      - text
  allow_variable_data_keys: false
  max_cache_size: 0.0
  valid_max_cache_size: null
  optim: adadelta
  optim_conf: {}
  scheduler: null
  scheduler_conf: {}
  token_list:
  - 
  - 
  - T
  - "\u2581"
  - I
  - H
  - G
  - O
  - AR
  - "\u2581T"
  - NE
  - E
  - EN
  - Y
  - "\u2581E"
  - "\u2581S"
  - EVEN
  - F
  - M
  - C
  - R
  - D
  - N
  - W
  - ENT
  - L
  - 
  init: null
  input_size: null
  ctc_conf:
      dropout_rate: 0.0
      ctc_type: builtin
      reduce: true
  model_conf:
      ctc_weight: 0.5
      ignore_id: -1
      lsm_weight: 0.0
      length_normalized_loss: false
      report_cer: true
      report_wer: true
      sym_space: 
      sym_blank: 
  use_preprocessor: true
  token_type: bpe
  bpemodel: data/token_list/bpe_unigram30/bpe.model
  non_linguistic_symbols: null
  cleaner: null
  g2p: null
  frontend: default
  frontend_conf:
      fs: 16k
  specaug: null
  specaug_conf: {}
  normalize: global_mvn
  normalize_conf:
      stats_file: exp/asr_stats_raw/train/feats_stats.npz
  encoder: rnn
  encoder_conf: {}
  decoder: rnn
  decoder_conf: {}
  required:
  - output_dir
  - token_list
  distributed: false

• LM config

  config: null
  print_config: false
  log_level: INFO
  dry_run: false
  iterator_type: sequence
  output_dir: exp/lm_train_bpe
  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
  cudnn_enabled: true
  cudnn_benchmark: false
  cudnn_deterministic: true
  collect_stats: false
  write_collected_feats: false
  max_epoch: 40
  patience: null
  val_scheduler_criterion:
  - valid
  - loss
  early_stopping_criterion:
  - valid
  - loss
  - min
  best_model_criterion:
  -   - train
      - loss
      - min
  -   - valid
      - loss
      - min
  -   - train
      - acc
      - max
  -   - valid
      - acc
      - max
  keep_nbest_models: 10
  grad_clip: 5.0
  grad_noise: false
  accum_grad: 1
  no_forward_run: false
  resume: true
  train_dtype: float32
  log_interval: null
  pretrain_path: []
  pretrain_key: []
  num_iters_per_epoch: null
  batch_size: 20
  valid_batch_size: null
  batch_bins: 1000000
  valid_batch_bins: null
  train_shape_file:
  - exp/lm_stats/train/text_shape.bpe
  valid_shape_file:
  - exp/lm_stats/valid/text_shape.bpe
  batch_type: folded
  valid_batch_type: null
  fold_length:
  - 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/srctexts
      - text
      - text
  valid_data_path_and_name_and_type:
  -   - dump/raw/train_dev/text
      - text
      - text
  allow_variable_data_keys: false
  max_cache_size: 0.0
  valid_max_cache_size: null
  optim: adadelta
  optim_conf: {}
  scheduler: null
  scheduler_conf: {}
  token_list:
  - 
  - 
  - T
  - "\u2581"
  - I
  - H
  - G
  - O
  - AR
  - "\u2581T"
  - NE
  - E
  - EN
  - Y
  - "\u2581E"
  - "\u2581S"
  - EVEN
  - F
  - M
  - C
  - R
  - D
  - N
  - W
  - ENT
  - L
  - 
  init: null
  model_conf:
      ignore_id: 0
  use_preprocessor: true
  token_type: bpe
  bpemodel: data/token_list/bpe_unigram30/bpe.model
  non_linguistic_symbols: null
  cleaner: null
  g2p: null
  lm: seq_rnn
  lm_conf: {}
  required:
  - output_dir
  - token_list
  distributed: false