ESPnet2 ENH model

`Johnson-Lsx/Shaoxiong_Lin_dns_ins20_enh_enh_train_enh_dccrn_raw`

This model was trained by Shaoxiong Lin using dns_ins20 recipe in espnet.

Demo: How to use in ESPnet2

cd espnet
git checkout 4538462eb7dc6a6b858adcbd3a526fb8173d6f73
pip install -e .
cd egs2/dns_ins20/enh1
./run.sh --skip_data_prep false --skip_train true --download_model Johnson-Lsx/Shaoxiong_Lin_dns_ins20_enh_enh_train_enh_dccrn_raw

RESULTS

Environments

date: Thu Feb 10 23:11:40 CST 2022
python version: 3.8.12 (default, Oct 12 2021, 13:49:34) [GCC 7.5.0]
espnet version: espnet 0.10.5a1
pytorch version: pytorch 1.9.1
Git hash: 6f66283b9eed7b0d5e5643feb18d8f60118a4afc
- Commit date: Mon Dec 13 15:30:29 2021 +0800

enh_train_enh_dccrn_batch_size_raw

config: ./conf/tuning/train_enh_dccrn_batch_size.yaml

dataset	STOI	SAR	SDR
enhanced_cv_synthetic	0.98	24.69	24.69
enhanced_tt_synthetic_no_reverb	0.96	17.69	17.69
enhanced_tt_synthetic_with_reverb	0.81	10.45	10.45

ENH config

expand

config: ./conf/tuning/train_enh_dccrn_batch_size.yaml
print_config: false
log_level: INFO
dry_run: false
iterator_type: chunk
output_dir: exp/enh_train_enh_dccrn_batch_size_raw
ngpu: 1
seed: 0
num_workers: 4
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: 46366
dist_launcher: null
multiprocessing_distributed: true
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: 10
val_scheduler_criterion:
- valid
- loss
early_stopping_criterion:
- valid
- loss
- min
best_model_criterion:
-   - valid
    - si_snr
    - max
-   - valid
    - loss
    - min
keep_nbest_models: 1
nbest_averaging_interval: 0
grad_clip: 5.0
grad_clip_type: 2.0
grad_noise: false
accum_grad: 1
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: []
num_iters_per_epoch: null
batch_size: 32
valid_batch_size: null
batch_bins: 1000000
valid_batch_bins: null
train_shape_file:
- exp/enh_stats_16k/train/speech_mix_shape
- exp/enh_stats_16k/train/speech_ref1_shape
- exp/enh_stats_16k/train/noise_ref1_shape
valid_shape_file:
- exp/enh_stats_16k/valid/speech_mix_shape
- exp/enh_stats_16k/valid/speech_ref1_shape
- exp/enh_stats_16k/valid/noise_ref1_shape
batch_type: folded
valid_batch_type: null
fold_length:
- 80000
- 80000
- 80000
sort_in_batch: descending
sort_batch: descending
multiple_iterator: false
chunk_length: 64000
chunk_shift_ratio: 0.5
num_cache_chunks: 1024
train_data_path_and_name_and_type:
-   - dump/raw/tr_synthetic/wav.scp
    - speech_mix
    - sound
-   - dump/raw/tr_synthetic/spk1.scp
    - speech_ref1
    - sound
-   - dump/raw/tr_synthetic/noise1.scp
    - noise_ref1
    - sound
valid_data_path_and_name_and_type:
-   - dump/raw/cv_synthetic/wav.scp
    - speech_mix
    - sound
-   - dump/raw/cv_synthetic/spk1.scp
    - speech_ref1
    - sound
-   - dump/raw/cv_synthetic/noise1.scp
    - noise_ref1
    - sound
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
    eps: 1.0e-08
    weight_decay: 1.0e-07
scheduler: reducelronplateau
scheduler_conf:
    mode: min
    factor: 0.7
    patience: 1
init: null
model_conf:
    loss_type: si_snr
criterions: 
  # The first criterion
  - name: si_snr 
    conf:
      eps: 1.0e-7
    # the wrapper for the current criterion
    # for single-talker case, we simplely use fixed_order wrapper
    wrapper: fixed_order
    wrapper_conf:
      weight: 1.0
use_preprocessor: false
encoder: stft
encoder_conf:
    n_fft: 512
    win_length: 400
    hop_length: 100
separator: dccrn
separator_conf: {}
decoder: stft
decoder_conf:
    n_fft: 512
    win_length: 400
    hop_length: 100
required:
- output_dir
version: 0.10.5a1
distributed: true

Citing ESPnet

@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}
}


@inproceedings{ESPnet-SE,
  author = {Chenda Li and Jing Shi and Wangyou Zhang and Aswin Shanmugam Subramanian and Xuankai Chang and 
  Naoyuki Kamo and Moto Hira and Tomoki Hayashi and Christoph B{"{o}}ddeker and Zhuo Chen and Shinji Watanabe},
  title = {ESPnet-SE: End-To-End Speech Enhancement and Separation Toolkit Designed for {ASR} Integration},
  booktitle = {{IEEE} Spoken Language Technology Workshop, {SLT} 2021, Shenzhen, China, January 19-22, 2021},
  pages = {785--792},
  publisher = {{IEEE}},
  year = {2021},
  url = {https://doi.org/10.1109/SLT48900.2021.9383615},
  doi = {10.1109/SLT48900.2021.9383615},
  timestamp = {Mon, 12 Apr 2021 17:08:59 +0200},
  biburl = {https://dblp.org/rec/conf/slt/Li0ZSCKHHBC021.bib},
  bibsource = {dblp computer science bibliography, https://dblp.org}
}

or arXiv:

@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}
}