Commit
•
9da4213
1
Parent(s):
e04352b
Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,169 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
tags:
|
| 3 |
+
- espnet
|
| 4 |
+
- audio
|
| 5 |
+
- self-supervised-learning
|
| 6 |
+
- speech-recognition
|
| 7 |
+
multilinguality:
|
| 8 |
+
- multilingual
|
| 9 |
+
task_categories:
|
| 10 |
+
- automatic-speech-recognition
|
| 11 |
+
language:
|
| 12 |
+
- afr
|
| 13 |
+
- amh
|
| 14 |
+
- ara
|
| 15 |
+
- asm
|
| 16 |
+
- ast
|
| 17 |
+
- azj
|
| 18 |
+
- bel
|
| 19 |
+
- ben
|
| 20 |
+
- bos
|
| 21 |
+
- cat
|
| 22 |
+
- ceb
|
| 23 |
+
- cmn
|
| 24 |
+
- ces
|
| 25 |
+
- cym
|
| 26 |
+
- dan
|
| 27 |
+
- deu
|
| 28 |
+
- ell
|
| 29 |
+
- eng
|
| 30 |
+
- spa
|
| 31 |
+
- est
|
| 32 |
+
- fas
|
| 33 |
+
- ful
|
| 34 |
+
- fin
|
| 35 |
+
- tgl
|
| 36 |
+
- fra
|
| 37 |
+
- gle
|
| 38 |
+
- glg
|
| 39 |
+
- guj
|
| 40 |
+
- hau
|
| 41 |
+
- heb
|
| 42 |
+
- hin
|
| 43 |
+
- hrv
|
| 44 |
+
- hun
|
| 45 |
+
- hye
|
| 46 |
+
- ind
|
| 47 |
+
- ibo
|
| 48 |
+
- isl
|
| 49 |
+
- ita
|
| 50 |
+
- jpn
|
| 51 |
+
- jav
|
| 52 |
+
- kat
|
| 53 |
+
- kam
|
| 54 |
+
- kea
|
| 55 |
+
- kaz
|
| 56 |
+
- khm
|
| 57 |
+
- kan
|
| 58 |
+
- kor
|
| 59 |
+
- ckb
|
| 60 |
+
- kir
|
| 61 |
+
- ltz
|
| 62 |
+
- lug
|
| 63 |
+
- lin
|
| 64 |
+
- lao
|
| 65 |
+
- lit
|
| 66 |
+
- luo
|
| 67 |
+
- lav
|
| 68 |
+
- mri
|
| 69 |
+
- mkd
|
| 70 |
+
- mal
|
| 71 |
+
- mon
|
| 72 |
+
- mar
|
| 73 |
+
- msa
|
| 74 |
+
- mlt
|
| 75 |
+
- mya
|
| 76 |
+
- nob
|
| 77 |
+
- npi
|
| 78 |
+
- nld
|
| 79 |
+
- nso
|
| 80 |
+
- nya
|
| 81 |
+
- oci
|
| 82 |
+
- orm
|
| 83 |
+
- ory
|
| 84 |
+
- pan
|
| 85 |
+
- pol
|
| 86 |
+
- pus
|
| 87 |
+
- por
|
| 88 |
+
- ron
|
| 89 |
+
- rus
|
| 90 |
+
- bul
|
| 91 |
+
- snd
|
| 92 |
+
- slk
|
| 93 |
+
- slv
|
| 94 |
+
- sna
|
| 95 |
+
- som
|
| 96 |
+
- srp
|
| 97 |
+
- swe
|
| 98 |
+
- swh
|
| 99 |
+
- tam
|
| 100 |
+
- tel
|
| 101 |
+
- tgk
|
| 102 |
+
- tha
|
| 103 |
+
- tur
|
| 104 |
+
- ukr
|
| 105 |
+
- umb
|
| 106 |
+
- urd
|
| 107 |
+
- uzb
|
| 108 |
+
- vie
|
| 109 |
+
- wol
|
| 110 |
+
- xho
|
| 111 |
+
- yor
|
| 112 |
+
- yue
|
| 113 |
+
- zul
|
| 114 |
+
datasets:
|
| 115 |
+
- fleurs
|
| 116 |
+
- babel
|
| 117 |
+
- voxpopuli
|
| 118 |
+
- commonvoice
|
| 119 |
+
license: cc-by-4.0
|
| 120 |
+
---
|
| 121 |
+
|
| 122 |
+
## WavLabLM-EK 40k
|
| 123 |
+
|
| 124 |
+
[Paper](https://arxiv.org/abs/2309.15317)
|
| 125 |
+
|
| 126 |
+
This model was trained by [William Chen](https://wanchichen.github.io/) using ESPNet2's SSL recipe in [espnet](https://github.com/espnet/espnet/).
|
| 127 |
+
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 English data for the quantization, making it especially strong for European languages.
|
| 128 |
+
|
| 129 |
+
|
| 130 |
+
```BibTex
|
| 131 |
+
@misc{chen2023joint,
|
| 132 |
+
title={Joint Prediction and Denoising for Large-scale Multilingual Self-supervised Learning},
|
| 133 |
+
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},
|
| 134 |
+
year={2023},
|
| 135 |
+
eprint={2309.15317},
|
| 136 |
+
archivePrefix={arXiv},
|
| 137 |
+
primaryClass={cs.CL}
|
| 138 |
+
}
|
| 139 |
+
```
|
| 140 |
+
|
| 141 |
+
|
| 142 |
+
### Citing ESPnet
|
| 143 |
+
|
| 144 |
+
```BibTex
|
| 145 |
+
@inproceedings{watanabe2018espnet,
|
| 146 |
+
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},
|
| 147 |
+
title={{ESPnet}: End-to-End Speech Processing Toolkit},
|
| 148 |
+
year={2018},
|
| 149 |
+
booktitle={Proceedings of Interspeech},
|
| 150 |
+
pages={2207--2211},
|
| 151 |
+
doi={10.21437/Interspeech.2018-1456},
|
| 152 |
+
url={http://dx.doi.org/10.21437/Interspeech.2018-1456}
|
| 153 |
+
}
|
| 154 |
+
|
| 155 |
+
|
| 156 |
+
```
|
| 157 |
+
|
| 158 |
+
or arXiv:
|
| 159 |
+
|
| 160 |
+
```bibtex
|
| 161 |
+
@misc{watanabe2018espnet,
|
| 162 |
+
title={ESPnet: End-to-End Speech Processing Toolkit},
|
| 163 |
+
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},
|
| 164 |
+
year={2018},
|
| 165 |
+
eprint={1804.00015},
|
| 166 |
+
archivePrefix={arXiv},
|
| 167 |
+
primaryClass={cs.CL}
|
| 168 |
+
}
|
| 169 |
+
```
|