Meta

Non-autoregressive generative model for speech with in-context learning, style transfer, and editing.

Unified model for speech-to-speech, speech-to-text, and text-to-speech translation across 100 languages.

Massively Multilingual Speech models covering 1,100+ languages for ASR and TTS.

Open-source large language model family with state-of-the-art performance and permissive licensing.

Industry-standard open-source ML framework used by researchers and engineers worldwide.