ESPnet: end-to-end speech processing toolkit
Published August 3, 2021 | Version v1
Other Open

ESPnet2 pretrained model, simpleoier/wsj_asr_train_asr_conformer_s3prlfrontend_hubert_raw_en_char_valid.acc.best, fs=16k, lang=en

Authors/Creators

Description

This model was trained by simpleoier using wsj 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 05a7d399f37a54659f42739c859d5b85cad9cdc6
pip install -e .
cd egs2/wsj/asr1
./run.sh --skip_data_prep false --skip_train true --download_model simpleoier/wsj_asr_train_asr_conformer_s3prlfrontend_hubert_raw_en_char_valid.acc.best
  • Results 
    
# RESULTS
## Environments
- date: `Thu Jul 15 02:41:00 EDT 2021`
- python version: `3.7.10 (default, Feb 26 2021, 18:47:35)  [GCC 7.3.0]`
- espnet version: `espnet 0.10.0`
- pytorch version: `pytorch 1.8.0`
- Git hash: `3a898e0ef5afb47c7afcc9600f754908f90a9518`
  - Commit date: `Thu Jul 15 00:11:41 2021 -0400`

## asr_train_asr_conformer_s3prlfrontend_hubert_raw_en_char
### WER

|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
|---|---|---|---|---|---|---|---|---|
|decode_lm_lm_train_lm_transformer_en_char_valid.loss.ave_asr_model_valid.acc.ave/test_dev93|503|8234|97.2|2.5|0.3|0.4|3.1|33.2|
|decode_lm_lm_train_lm_transformer_en_char_valid.loss.ave_asr_model_valid.acc.ave/test_eval92|333|5643|98.4|1.6|0.1|0.2|1.8|22.2|

### CER

|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
|---|---|---|---|---|---|---|---|---|
|decode_lm_lm_train_lm_transformer_en_char_valid.loss.ave_asr_model_valid.acc.ave/test_dev93|503|48634|99.0|0.3|0.7|0.2|1.2|43.7|
|decode_lm_lm_train_lm_transformer_en_char_valid.loss.ave_asr_model_valid.acc.ave/test_eval92|333|33341|99.4|0.2|0.4|0.1|0.7|35.4|

### TER

|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err|
|---|---|---|---|---|---|---|---|---|
  • ASR config 
    config: conf/tuning/train_asr_conformer_s3prlfrontend_hubert.yaml
print_config: false
log_level: INFO
dry_run: false
iterator_type: sequence
output_dir: exp/asr_train_asr_conformer_s3prlfrontend_hubert_raw_en_char
ngpu: 1
seed: 0
num_workers: 8
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: false
sharded_ddp: false
cudnn_enabled: true
cudnn_benchmark: false
cudnn_deterministic: true
collect_stats: false
write_collected_feats: false
max_epoch: 100
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: 4
no_forward_run: false
resume: true
train_dtype: float32
use_amp: false
log_interval: null
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: 4000000
valid_batch_bins: null
train_shape_file:
- exp/asr_stats_raw_en_char/train/speech_shape
- exp/asr_stats_raw_en_char/train/text_shape.char
valid_shape_file:
- exp/asr_stats_raw_en_char/valid/speech_shape
- exp/asr_stats_raw_en_char/valid/text_shape.char
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_si284/wav.scp
    - speech
    - sound
-   - dump/raw/train_si284/text
    - text
    - text
valid_data_path_and_name_and_type:
-   - dump/raw/test_dev93/wav.scp
    - speech
    - sound
-   - dump/raw/test_dev93/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: 0.005
scheduler: warmuplr
scheduler_conf:
    warmup_steps: 30000
token_list:
- 
- 
- 
- E
- T
- A
- N
- I
- O
- S
- R
- H
- L
- D
- C
- U
- M
- P
- F
- G
- Y
- W
- B
- V
- K
- .
- X
- ''''
- J
- Q
- Z
- 
- ','
- '-'
- '"'
- '*'
- ':'
- (
- )
- '?'
- '!'
- '&'
- ;
- '1'
- '2'
- '0'
- /
- $
- '{'
- '}'
- '8'
- '9'
- '6'
- '3'
- '5'
- '7'
- '4'
- '~'
- '`'
- _
- <*IN*>
- <*MR.*>
- \
- ^
- 
init: xavier_uniform
input_size: null
ctc_conf:
    ignore_nan_grad: true
model_conf:
    ctc_weight: 0.3
    lsm_weight: 0.1
    length_normalized_loss: false
use_preprocessor: true
token_type: char
bpemodel: null
non_linguistic_symbols: data/nlsyms.txt
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: hubert_large_ll60k
    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: {}
preencoder: linear
preencoder_conf:
    input_size: 1024
    output_size: 80
encoder: conformer
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: conv2d2
    normalize_before: true
    macaron_style: true
    rel_pos_type: latest
    pos_enc_layer_type: rel_pos
    selfattention_layer_type: rel_selfattn
    activation_type: swish
    use_cnn_module: true
    cnn_module_kernel: 15
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: 0.10.0
distributed: false
  • LM config 
    config: conf/train_lm_transformer.yaml
print_config: false
log_level: INFO
dry_run: false
iterator_type: sequence
output_dir: exp/lm_train_lm_transformer_en_char
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: false
sharded_ddp: false
cudnn_enabled: true
cudnn_benchmark: false
cudnn_deterministic: true
collect_stats: false
write_collected_feats: false
max_epoch: 25
patience: null
val_scheduler_criterion:
- valid
- loss
early_stopping_criterion:
- valid
- loss
- min
best_model_criterion:
-   - valid
    - loss
    - min
keep_nbest_models: 10
grad_clip: 5.0
grad_clip_type: 2.0
grad_noise: false
accum_grad: 2
no_forward_run: false
resume: true
train_dtype: float32
use_amp: false
log_interval: null
use_tensorboard: true
use_wandb: false
wandb_project: null
wandb_id: null
detect_anomaly: false
pretrain_path: null
init_param: []
ignore_init_mismatch: false
freeze_param: []
num_iters_per_epoch: null
batch_size: 20
valid_batch_size: null
batch_bins: 350000
valid_batch_bins: null
train_shape_file:
- exp/lm_stats_en_char/train/text_shape.char
valid_shape_file:
- exp/lm_stats_en_char/valid/text_shape.char
batch_type: numel
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/lm_train.txt
    - text
    - text
valid_data_path_and_name_and_type:
-   - dump/raw/test_dev93/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:
- 
- 
- 
- E
- T
- A
- N
- I
- O
- S
- R
- H
- L
- D
- C
- U
- M
- P
- F
- G
- Y
- W
- B
- V
- K
- .
- X
- ''''
- J
- Q
- Z
- 
- ','
- '-'
- '"'
- '*'
- ':'
- (
- )
- '?'
- '!'
- '&'
- ;
- '1'
- '2'
- '0'
- /
- $
- '{'
- '}'
- '8'
- '9'
- '6'
- '3'
- '5'
- '7'
- '4'
- '~'
- '`'
- _
- <*IN*>
- <*MR.*>
- \
- ^
- 
init: null
model_conf:
    ignore_id: 0
use_preprocessor: true
token_type: char
bpemodel: null
non_linguistic_symbols: data/nlsyms.txt
cleaner: null
g2p: null
lm: transformer
lm_conf:
    pos_enc: null
    embed_unit: 128
    att_unit: 512
    head: 8
    unit: 2048
    layer: 16
    dropout_rate: 0.1
required:
- output_dir
- token_list
version: 0.9.9
distributed: false

Files

asr_train_asr_conformer_s3prlfrontend_hubert_raw_en_char_valid.acc.best.zip

Files (1.6 GB)

Additional details

Related works

Is supplement to
https://github.com/espnet/espnet (URL)