https://github.com/Audio-WestlakeU/FullSubNet

PyTorch implementation of "FullSubNet: A Full-Band and Sub-Band Fusion Model for Real-Time Single-Channel Speech Enhancement."
https://github.com/Audio-WestlakeU/FullSubNet

audio band denoising full-band narrow-band noise-reduction paper pretrained-model pytorch reproducible-research single-channel speech speech-enhancement speech-processing speech-separation sub-band

Last synced: 3 months ago
JSON representation

PyTorch implementation of "FullSubNet: A Full-Band and Sub-Band Fusion Model for Real-Time Single-Channel Speech Enhancement."

• Host: GitHub
• URL: https://github.com/Audio-WestlakeU/FullSubNet
• Owner: Audio-WestlakeU
• License: mit
• Created: 2020-12-18T09:05:36.000Z (almost 4 years ago)
• Default Branch: main
• Last Pushed: 2023-08-19T07:01:09.000Z (about 1 year ago)
• Last Synced: 2024-05-16T12:58:25.965Z (6 months ago)
• Topics: audio, band, denoising, full-band, narrow-band, noise-reduction, paper, pretrained-model, pytorch, reproducible-research, single-channel, speech, speech-enhancement, speech-processing, speech-separation, sub-band
• Language: Python
• Homepage: https://fullsubnet.readthedocs.io/en/latest/
• Size: 892 KB
• Stars: 508
• Watchers: 10
• Forks: 148
• Open Issues: 37
• Metadata Files:
  • Readme: README.md
  • Changelog: CHANGELOG.md
  • Contributing: CONTRIBUTING.md
  • License: LICENSE
  • Code of conduct: CODE_OF_CONDUCT.md

Awesome Lists containing this project

README

        


    


        FullSubNet

    

    


    Full-Band and Sub-Band Fusion Model for Real-Time Single-Channel Speech Enhancement

    

    

    

    

        

    

    

    

    



## Guides

The documentation is hosted on [Read the Docs](https://fullsubnet.readthedocs.io/). Check the documentation for **how to train and test models**.

- Improved FullSubNet: Further reduces computational costs and enables high sampling rate data processing, e.g., 48 KHz and 24 KHz.

  - ❇️ [Model Architecture](https://github.com/Audio-WestlakeU/FullSubNet/blob/main/recipes/dns_interspeech_2020/improved_fullsubnet/model.py)

- 📰 [FullSubNet: A Full-Band and Sub-Band Fusion Model for Real-Time Single-Channel Speech Enhancement, ICASSP 2021](https://arxiv.org/abs/2010.15508)

  - 📸 [Demo (Audio Clips)](https://www.haoxiangsnr.com/publications/3)

  - 🎏 [Model Checkpoints](https://github.com/haoxiangsnr/FullSubNet/releases)

  - ❇️ [Model Architecture](https://github.com/haoxiangsnr/FullSubNet/blob/fast_fullsubnet/recipes/dns_interspeech_2020/fullsubnet/model.py)

- 📰 [Fast FullSubNet: Accelerate Full-band and Sub-band Fusion Model for Single-channel Speech Enhancement](https://arxiv.org/abs/2212.09019)

  - ❇️ [Model Architecture](https://github.com/haoxiangsnr/FullSubNet/blob/fast_fullsubnet/recipes/dns_interspeech_2020/fast_fullsubnet/model.py)

  - 📸 [Demo (Audio Clips)](https://www.haoxiangsnr.com/publications/3)

- cIRM-based Fullband baseline model (described in the original FullSubNet paper)

  - ❇️ [Model Architecture](https://github.com/haoxiangsnr/FullSubNet/blob/fast_fullsubnet/recipes/dns_interspeech_2020/fullband_baseline/model.py)

## Citation

If you use this code for your research, please consider citeing:

```text

@INPROCEEDINGS{hao2020fullsubnet,

    author={Hao, Xiang and Su, Xiangdong and Horaud, Radu and Li, Xiaofei},

    booktitle={ICASSP 2021 - 2021 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)},

    title={Fullsubnet: A Full-Band and Sub-Band Fusion Model for Real-Time Single-Channel Speech Enhancement},

    year={2021},

    pages={6633-6637},

    doi={10.1109/ICASSP39728.2021.9414177}

}

```

## License

This repository Under the [MIT license](LICENSE).