🛠️ Caramel: a speech-to-speech conversation model trained only on audio
Summary
Caramel is a speech-to-speech conversation model trained exclusively on audio recordings without any text. The open-source project runs fast enough on GPUs and Apple Silicon to enable live conversation.
Why it’s interesting
It is a text-free speech-to-speech model trained entirely on audio data that achieves fast generation speeds suitable for live conversation.
Source metrics: Points 3 · Comments 0
HN discussion · Project
Source: #HackerNews / Show HN
Summary
Caramel is a speech-to-speech conversation model trained exclusively on audio recordings without any text. The open-source project runs fast enough on GPUs and Apple Silicon to enable live conversation.
Why it’s interesting
It is a text-free speech-to-speech model trained entirely on audio data that achieves fast generation speeds suitable for live conversation.
Source metrics: Points 3 · Comments 0
HN discussion · Project
Source: #HackerNews / Show HN