Kyutai
5 models in Artifacts Logshuggingface.co/kyutai
Kyutai STT 2.6B English06 Jun 2025
2.6BCC-BY-4.00 · 30d
The French Kyutai team continues to release open audio models. This speech-to-text model is their latest addition.
Helium-1-2b30 Apr 2025
2BCC-BY-SA-4.06.9K · 30d
2B LLM, targeting all 24 EU languages.
Hibiki-M MLX04 Feb 2025
1.7BCC-BY-4.047 · 30d
Another speech-to-speech model from Kyutai, this time for simultaneous translation from French to English. It is based on their previous model, moshi.
MoshiVis27 Jan 2025
7BCC-BY-4.00 · 30d
An image-capable extension of the text-to-speech moshi model, which uses PaliGemma 2 as its encoder.
moshiko-pytorch-bf1611 Sep 2024
253.7K · 30d
A fun high-speed speech-to-speech model: