Commit
•
92a7c8a
1
Parent(s):
a0e83b7
Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,163 @@
|
|
1 |
---
|
2 |
license: cc-by-nc-4.0
|
|
|
|
|
3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: cc-by-nc-4.0
|
3 |
+
tags:
|
4 |
+
- mms
|
5 |
---
|
6 |
+
|
7 |
+
# Massively Multilingual Speech (MMS) - Common Crawl Language Models
|
8 |
+
|
9 |
+
This repository consists of the n-gram language models trained on Common Crawl data ([Conneau et al. 2020b](https://aclanthology.org/2020.acl-main.747/), [NLLB_Team et al. 2022](https://arxiv.org/abs/2207.04672)) using [KenLM library](https://github.com/kpu/kenlm).
|
10 |
+
|
11 |
+
## Table Of Content
|
12 |
+
|
13 |
+
- [Example](#example)
|
14 |
+
- [Supported Languages](#supported-languages)
|
15 |
+
- [Model details](#model-details)
|
16 |
+
- [Additional links](#additional-links)
|
17 |
+
|
18 |
+
## Example
|
19 |
+
|
20 |
+
```py
|
21 |
+
|
22 |
+
TODO
|
23 |
+
```
|
24 |
+
|
25 |
+
## Supported Languages
|
26 |
+
|
27 |
+
We support language models in 102 languages. Unclick the following to toogle all supported languages of this checkpoint in [ISO 639-3 code](https://en.wikipedia.org/wiki/ISO_639-3).
|
28 |
+
You can find more details about the languages and their ISO 639-3 codes in the [MMS Language Coverage Overview](https://dl.fbaipublicfiles.com/mms/misc/language_coverage_mms.html).
|
29 |
+
<details>
|
30 |
+
<summary>Click to toggle</summary>
|
31 |
+
|
32 |
+
- afr
|
33 |
+
- amh
|
34 |
+
- ara
|
35 |
+
- asm
|
36 |
+
- ast
|
37 |
+
- azj
|
38 |
+
- bel
|
39 |
+
- ben
|
40 |
+
- bos
|
41 |
+
- bul
|
42 |
+
- cat
|
43 |
+
- ceb
|
44 |
+
- ces
|
45 |
+
- ckb
|
46 |
+
- cmn
|
47 |
+
- cym
|
48 |
+
- dan
|
49 |
+
- deu
|
50 |
+
- ell
|
51 |
+
- eng
|
52 |
+
- est
|
53 |
+
- fas
|
54 |
+
- fin
|
55 |
+
- fra
|
56 |
+
- ful
|
57 |
+
- gle
|
58 |
+
- glg
|
59 |
+
- guj
|
60 |
+
- hau
|
61 |
+
- heb
|
62 |
+
- hin
|
63 |
+
- hrv
|
64 |
+
- hun
|
65 |
+
- hye
|
66 |
+
- ibo
|
67 |
+
- ind
|
68 |
+
- isl
|
69 |
+
- ita
|
70 |
+
- jav
|
71 |
+
- jpn
|
72 |
+
- kam
|
73 |
+
- kan
|
74 |
+
- kat
|
75 |
+
- kaz
|
76 |
+
- kea
|
77 |
+
- khm
|
78 |
+
- kir
|
79 |
+
- kor
|
80 |
+
- lao
|
81 |
+
- lav
|
82 |
+
- lin
|
83 |
+
- lit
|
84 |
+
- ltz
|
85 |
+
- lug
|
86 |
+
- luo
|
87 |
+
- mal
|
88 |
+
- mar
|
89 |
+
- mkd
|
90 |
+
- mlt
|
91 |
+
- mon
|
92 |
+
- mri
|
93 |
+
- mya
|
94 |
+
- nld
|
95 |
+
- nob
|
96 |
+
- npi
|
97 |
+
- nso
|
98 |
+
- nya
|
99 |
+
- oci
|
100 |
+
- orm
|
101 |
+
- ory
|
102 |
+
- pan
|
103 |
+
- pol
|
104 |
+
- por
|
105 |
+
- pus
|
106 |
+
- ron
|
107 |
+
- rus
|
108 |
+
- slk
|
109 |
+
- slv
|
110 |
+
- sna
|
111 |
+
- snd
|
112 |
+
- som
|
113 |
+
- spa
|
114 |
+
- srp
|
115 |
+
- swe
|
116 |
+
- swh
|
117 |
+
- tam
|
118 |
+
- tel
|
119 |
+
- tgk
|
120 |
+
- tgl
|
121 |
+
- tha
|
122 |
+
- tur
|
123 |
+
- ukr
|
124 |
+
- umb
|
125 |
+
- urd
|
126 |
+
- uzb
|
127 |
+
- vie
|
128 |
+
- wol
|
129 |
+
- xho
|
130 |
+
- yor
|
131 |
+
- yue
|
132 |
+
- zlm
|
133 |
+
- zul
|
134 |
+
</details>
|
135 |
+
|
136 |
+
## Model details
|
137 |
+
|
138 |
+
- **Developed by:** Vineel Pratap et al.
|
139 |
+
- **Model type:** Multi-Lingual Automatic Speech Recognition model
|
140 |
+
- **Language(s):** 126 languages, see [supported languages](#supported-languages)
|
141 |
+
- **License:** CC-BY-NC 4.0 license
|
142 |
+
- **Num parameters**: 1 billion
|
143 |
+
- **Audio sampling rate**: 16,000 kHz
|
144 |
+
- **Cite as:**
|
145 |
+
|
146 |
+
@article{pratap2023mms,
|
147 |
+
title={Scaling Speech Technology to 1,000+ Languages},
|
148 |
+
author={Vineel Pratap and Andros Tjandra and Bowen Shi and Paden Tomasello and Arun Babu and Sayani Kundu and Ali Elkahky and Zhaoheng Ni and Apoorv Vyas and Maryam Fazel-Zarandi and Alexei Baevski and Yossi Adi and Xiaohui Zhang and Wei-Ning Hsu and Alexis Conneau and Michael Auli},
|
149 |
+
journal={arXiv},
|
150 |
+
year={2023}
|
151 |
+
}
|
152 |
+
|
153 |
+
## Additional Links
|
154 |
+
|
155 |
+
- [Blog post](https://ai.facebook.com/blog/multilingual-model-speech-recognition/)
|
156 |
+
- [Transformers documentation](https://huggingface.co/docs/transformers/main/en/model_doc/mms).
|
157 |
+
- [Paper](https://arxiv.org/abs/2305.13516)
|
158 |
+
- [GitHub Repository](https://github.com/facebookresearch/fairseq/tree/main/examples/mms#asr)
|
159 |
+
- [Other **MMS** checkpoints](https://huggingface.co/models?other=mms)
|
160 |
+
- MMS base checkpoints:
|
161 |
+
- [facebook/mms-1b](https://huggingface.co/facebook/mms-1b)
|
162 |
+
- [facebook/mms-300m](https://huggingface.co/facebook/mms-300m)
|
163 |
+
- [Official Space](https://huggingface.co/spaces/facebook/MMS)
|