Modelpublic

espnet/WavLabLM-MK-40k

sourceHugging Facecc-by-4.0updated 4d agoView on Hugging Face

0likes16downloads

Model Card

WavLabLM-MK 40k

This model was trained by William Chen using ESPNet2's SSL recipe in espnet. WavLabLM is an self-supervised audio encoder pre-trained on 40,000 hours of multilingual data across 136 languages. This specific variant, WavLabLM-MK, uses a K-means model trained on multilingual data for the quantization. It is used as the base model for WavLabLM-MS, which is better tuned for non-European languages.

BibTex

@misc{chen2023joint,
      title={Joint Prediction and Denoising for Large-scale Multilingual Self-supervised Learning}, 
      author={William Chen and Jiatong Shi and Brian Yan and Dan Berrebbi and Wangyou Zhang and Yifan Peng and Xuankai Chang and Soumi Maiti and Shinji Watanabe},
      year={2023},
      eprint={2309.15317},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

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}
}