Show HN: Caramel: a speech-to-speech conversation model trained only on audio

4 points | by naklecha 11 hours ago

No comments yet.