|
|
--- |
|
|
tags: |
|
|
- espnet |
|
|
- audio |
|
|
- automatic-speech-recognition |
|
|
language: noinfo |
|
|
datasets: |
|
|
- mr_openslr64 |
|
|
license: cc-by-4.0 |
|
|
--- |
|
|
|
|
|
## ESPnet2 ASR model |
|
|
|
|
|
### `espnet/marathi_openslr64` |
|
|
|
|
|
This model was trained by Sujay Suresh Kumar using mr_openslr64 recipe in [espnet](https://github.com/espnet/espnet/). |
|
|
|
|
|
### Demo: How to use in ESPnet2 |
|
|
|
|
|
```bash |
|
|
cd espnet |
|
|
git checkout 91325a1e58ca0b13494b94bf79b186b095fe0b58 |
|
|
pip install -e . |
|
|
cd egs2/mr_openslr64/asr1 |
|
|
./run.sh --skip_data_prep false --skip_train true --download_model espnet/marathi_openslr64 |
|
|
``` |
|
|
|
|
|
<!-- Generated by scripts/utils/show_asr_result.sh --> |
|
|
# RESULTS |
|
|
## Environments |
|
|
- date: `Mon Mar 21 16:06:03 UTC 2022` |
|
|
- python version: `3.9.7 (default, Sep 16 2021, 13:09:58) [GCC 7.5.0]` |
|
|
- espnet version: `espnet 0.10.7a1` |
|
|
- pytorch version: `pytorch 1.11.0+cu102` |
|
|
- Git hash: `91325a1e58ca0b13494b94bf79b186b095fe0b58` |
|
|
- Commit date: `Mon Mar 21 00:40:52 2022 +0000` |
|
|
|
|
|
## asr_train_asr_conformer_xlsr_raw_bpe150_sp |
|
|
### WER |
|
|
|
|
|
|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err| |
|
|
|---|---|---|---|---|---|---|---|---| |
|
|
|decode_asr_batch_size1_asr_model_valid.acc.ave/marathi_test|299|3625|72.9|22.5|4.7|1.7|28.9|88.6| |
|
|
|
|
|
### CER |
|
|
|
|
|
|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err| |
|
|
|---|---|---|---|---|---|---|---|---| |
|
|
|decode_asr_batch_size1_asr_model_valid.acc.ave/marathi_test|299|20557|91.4|3.1|5.5|1.9|10.5|88.6| |
|
|
|
|
|
### TER |
|
|
|
|
|
|dataset|Snt|Wrd|Corr|Sub|Del|Ins|Err|S.Err| |
|
|
|---|---|---|---|---|---|---|---|---| |
|
|
|decode_asr_batch_size1_asr_model_valid.acc.ave/marathi_test|299|13562|86.5|6.3|7.1|1.4|14.9|88.6| |
|
|
|
|
|
## ASR config |
|
|
|
|
|
<details><summary>expand</summary> |
|
|
|
|
|
``` |
|
|
config: conf/tuning/train_asr_conformer_xlsr.yaml |
|
|
print_config: false |
|
|
log_level: INFO |
|
|
dry_run: false |
|
|
iterator_type: sequence |
|
|
output_dir: exp/asr_train_asr_conformer_xlsr_raw_bpe150_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: false |
|
|
sharded_ddp: false |
|
|
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: 5 |
|
|
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: 10000 |
|
|
valid_batch_bins: null |
|
|
train_shape_file: |
|
|
- exp/asr_stats_raw_bpe150_sp/train/speech_shape |
|
|
- exp/asr_stats_raw_bpe150_sp/train/text_shape.bpe |
|
|
valid_shape_file: |
|
|
- exp/asr_stats_raw_bpe150_sp/valid/speech_shape |
|
|
- exp/asr_stats_raw_bpe150_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/marathi_train_sp/wav.scp |
|
|
- speech |
|
|
- sound |
|
|
- - dump/raw/marathi_train_sp/text |
|
|
- text |
|
|
- text |
|
|
valid_data_path_and_name_and_type: |
|
|
- - dump/raw/marathi_dev/wav.scp |
|
|
- speech |
|
|
- sound |
|
|
- - dump/raw/marathi_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: 0.0005 |
|
|
scheduler: warmuplr |
|
|
scheduler_conf: |
|
|
warmup_steps: 20000 |
|
|
token_list: |
|
|
- <blank> |
|
|
- <unk> |
|
|
- ▁ |
|
|
- ा |
|
|
- ी |
|
|
- े |
|
|
- त |
|
|
- र |
|
|
- ं |
|
|
- न |
|
|
- क |
|
|
- ् |
|
|
- व |
|
|
- ि |
|
|
- ल |
|
|
- ▁म |
|
|
- स |
|
|
- ो |
|
|
- श |
|
|
- द |
|
|
- च |
|
|
- म |
|
|
- ▁अ |
|
|
- ▁आ |
|
|
- ण |
|
|
- ु |
|
|
- ला |
|
|
- ह |
|
|
- ▁आहे |
|
|
- य |
|
|
- ▁स |
|
|
- ग |
|
|
- ▁ह |
|
|
- ्या |
|
|
- चा |
|
|
- ▁प |
|
|
- ड |
|
|
- ▁क |
|
|
- प |
|
|
- ट |
|
|
- ▁ब |
|
|
- ज |
|
|
- र् |
|
|
- ्र |
|
|
- ▁? |
|
|
- ▁ज |
|
|
- ब |
|
|
- ून |
|
|
- वा |
|
|
- ▁एक |
|
|
- ▁या |
|
|
- ळ |
|
|
- ात |
|
|
- ख |
|
|
- ध |
|
|
- ▁ति |
|
|
- ठ |
|
|
- ल्या |
|
|
- ले |
|
|
- ू |
|
|
- ▁तुम्हाला |
|
|
- ां |
|
|
- ार |
|
|
- घ |
|
|
- ची |
|
|
- ▁अस |
|
|
- थ |
|
|
- ▁का |
|
|
- ने |
|
|
- णि |
|
|
- ॅ |
|
|
- ▁त |
|
|
- ▁परवा |
|
|
- ▁ते |
|
|
- ली |
|
|
- ▁गेल |
|
|
- ळा |
|
|
- ष |
|
|
- ▁कर |
|
|
- . |
|
|
- च्या |
|
|
- ▁न |
|
|
- वर |
|
|
- ▁त्या |
|
|
- ▁प्र |
|
|
- ▁करू |
|
|
- ▁ग |
|
|
- ्ट |
|
|
- ई |
|
|
- झ |
|
|
- ▁फ |
|
|
- ाय |
|
|
- क्ष |
|
|
- ▁काय |
|
|
- पूर |
|
|
- ▁होती |
|
|
- मध |
|
|
- ▁तिथ |
|
|
- ▁काही |
|
|
- ए |
|
|
- ▁वि |
|
|
- ▁दोन |
|
|
- ▁महिन्या |
|
|
- व्हा |
|
|
- तील |
|
|
- जार |
|
|
- ▁नाही |
|
|
- ँ |
|
|
- ▁पुत |
|
|
- ॉ |
|
|
- ▁झाला |
|
|
- ▁दिसल |
|
|
- ▁साल |
|
|
- ▁रस्त्यावर |
|
|
- स्त |
|
|
- जवळ |
|
|
- न्म |
|
|
- मध्य |
|
|
- ऊ |
|
|
- ▁इथे |
|
|
- ▁तुमच |
|
|
- ▁शकते |
|
|
- मान |
|
|
- ▁उद् |
|
|
- फ |
|
|
- ै |
|
|
- ढ |
|
|
- ',' |
|
|
- इ |
|
|
- ौ |
|
|
- |
|
|
- ृ |
|
|
- ओ |
|
|
- ः |
|
|
- ॲ |
|
|
- आ |
|
|
- '-' |
|
|
- ञ |
|
|
- औ |
|
|
- '!' |
|
|
- ऑ |
|
|
- ऱ |
|
|
- ऐ |
|
|
- छ |
|
|
- उ |
|
|
- '?' |
|
|
- भ |
|
|
- अ |
|
|
- ऋ |
|
|
- <sos/eos> |
|
|
init: xavier_uniform |
|
|
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/token_list/bpe_unigram150/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_xlsr |
|
|
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: conformer |
|
|
encoder_conf: |
|
|
output_size: 512 |
|
|
attention_heads: 4 |
|
|
linear_units: 1024 |
|
|
num_blocks: 3 |
|
|
dropout_rate: 0.3 |
|
|
positional_dropout_rate: 0.3 |
|
|
attention_dropout_rate: 0.3 |
|
|
input_layer: conv2d |
|
|
normalize_before: true |
|
|
macaron_style: false |
|
|
pos_enc_layer_type: rel_pos |
|
|
selfattention_layer_type: rel_selfattn |
|
|
activation_type: swish |
|
|
use_cnn_module: true |
|
|
cnn_module_kernel: 17 |
|
|
postencoder: null |
|
|
postencoder_conf: {} |
|
|
decoder: transformer |
|
|
decoder_conf: |
|
|
attention_heads: 4 |
|
|
linear_units: 1024 |
|
|
num_blocks: 3 |
|
|
dropout_rate: 0.3 |
|
|
positional_dropout_rate: 0.3 |
|
|
self_attention_dropout_rate: 0.3 |
|
|
src_attention_dropout_rate: 0.3 |
|
|
required: |
|
|
- output_dir |
|
|
- token_list |
|
|
version: 0.10.7a1 |
|
|
distributed: false |
|
|
``` |
|
|
|
|
|
</details> |
|
|
|
|
|
|
|
|
|
|
|
### Citing ESPnet |
|
|
|
|
|
```BibTex |
|
|
@inproceedings{watanabe2018espnet, |
|
|
author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson Yalta and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai}, |
|
|
title={{ESPnet}: End-to-End Speech Processing Toolkit}, |
|
|
year={2018}, |
|
|
booktitle={Proceedings of Interspeech}, |
|
|
pages={2207--2211}, |
|
|
doi={10.21437/Interspeech.2018-1456}, |
|
|
url={http://dx.doi.org/10.21437/Interspeech.2018-1456} |
|
|
} |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
``` |
|
|
|
|
|
or arXiv: |
|
|
|
|
|
```bibtex |
|
|
@misc{watanabe2018espnet, |
|
|
title={ESPnet: End-to-End Speech Processing Toolkit}, |
|
|
author={Shinji Watanabe and Takaaki Hori and Shigeki Karita and Tomoki Hayashi and Jiro Nishitoba and Yuya Unno and Nelson Yalta and Jahn Heymann and Matthew Wiesner and Nanxin Chen and Adithya Renduchintala and Tsubasa Ochiai}, |
|
|
year={2018}, |
|
|
eprint={1804.00015}, |
|
|
archivePrefix={arXiv}, |
|
|
primaryClass={cs.CL} |
|
|
} |
|
|
``` |
|
|
|
