---
license: cc-by-4.0
language:
  - en
  - fr
library_name: moshi
tags:
  - audio
  - automatic-speech-recognition
---

# Moshi Streaming Speech-to-Text (Quantized)

This is a quantized version of Kyutai’s [stt-1b-en_fr](https://huggingface.co/kyutai/stt-1b-en_fr) model. The original model is a 1B parameter streaming speech-to-text model for English and French. This fork contains the same model, quantized to Q8_0 and Q4_K GGUF formats for reduced memory usage and faster inference.