ModelCross-lingual Speech Representations
XLSR
A cross-lingual self-supervised speech model proposed by Meta. XLSR is pretrained on large multilingual speech corpora through self-supervised learning, supporting speech representation learning across tens to thousands of languages.
Timeline (last 90 days)
Oct 2
Multilingual self-supervised models such as XLSR and MMS extend the self-supervised approach to dozens or even thousands of languages.
Unverified50%