Instructions to use vanch007/SenseVoiceSmall-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- MLX
How to use vanch007/SenseVoiceSmall-4bit with MLX:
# Download the model from the Hub pip install huggingface_hub[hf_xet] huggingface-cli download --local-dir SenseVoiceSmall-4bit vanch007/SenseVoiceSmall-4bit
- Notebooks
- Google Colab
- Kaggle
- Local Apps
- LM Studio
File size: 470 Bytes
b5365ba | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 | ---
library_name: mlx-audio
tags:
- mlx
- mlx-audio
- sensevoice
- speech-to-text
- speech
- 4bit
base_model:
- mlx-community/SenseVoiceSmall
---
# SenseVoiceSmall 4bit MLX
This repository is a local 4-bit MLX quantization of `mlx-community/SenseVoiceSmall` for VTranslator.
Quantization:
- bits: 4
- group_size: 64
- target repo: `vanch007/SenseVoiceSmall-4bit`
The app expects the standard `config.json`, `model*.safetensors`, `am.mvn`, and tokenizer model files.
|