---
title: SenseVoiceSmall-GGUF
canonical_url: "https://www.modelscope.cn/models/FunAudioLLM/SenseVoiceSmall-GGUF"
md_url: "https://www.modelscope.cn/models/FunAudioLLM/SenseVoiceSmall-GGUF.md"
repository: FunAudioLLM/SenseVoiceSmall-GGUF
chinese_name: "SenseVoice GGUF"
last_updated: 2026-06-21
license: "Apache License 2.0"
pipeline_tag: auto-speech-recognition
tasks:
  - auto-speech-recognition
library_name:
  - gguf
downloads: 3924
stars: 7
tags:
  - gguf
---

# SenseVoiceSmall-GGUF

> SenseVoiceSmall-GGUF - FunAudioLLM 在 ModelScope 开源的模型。SenseVoiceSmall-GGUF(SenseVoice GGUF)

FunAudioLLM/SenseVoiceSmall-GGUF 是 ModelScope 魔搭社区上的auto-speech-recognition模型，采用 Apache License 2.0 许可。

- **Repository**: FunAudioLLM/SenseVoiceSmall-GGUF
- **License**: Apache License 2.0
- **Tasks**: auto-speech-recognition
- **Tags**: gguf
- **Downloads**: 3924
- **Stars**: 7
- **Last updated**: 2026-06-21

Source: https://www.modelscope.cn/models/FunAudioLLM/SenseVoiceSmall-GGUF

---

# SenseVoiceSmall-GGUF(SenseVoice GGUF)

SenseVoiceSmall(SAN-M 编码器 + CTC)的 GGUF 量化版,多语种 ASR + 语种/情感/事件标签,中文/粤语强。用于**零 Python、CPU/边缘**的 [FunASR llama.cpp 运行时](https://github.com/modelscope/FunASR/tree/main/runtime/llama.cpp)——像 whisper.cpp 一样的单文件二进制,中文 CER 约为 whisper.cpp 的 1/3。

## 安装 / 运行(无需 Python、无需编译)
1. 从 GitHub Releases 下预编译二进制(Linux / macOS / Windows):https://github.com/modelscope/FunASR/releases (tag `runtime-llamacpp-v*`)
2. 一站式下载与 quickstart:https://www.funasr.com/llama-cpp
3. 拉模型并运行:
```bash
bash download-funasr-model.sh sensevoice ./gguf
llama-funasr-sensevoice -m ./gguf/sensevoice-small-q8.gguf --vad ./gguf/fsmn-vad.gguf -a audio.wav
```

## 文件
| 文件 | 大小 |
|---|---|
| `sensevoice-small-f16.gguf` | 470 MB |
| `sensevoice-small-q8.gguf` | 254 MB |

## ⭐ 开源,觉得有用请点 Star 支持
- FunASR: https://github.com/modelscope/FunASR
- SenseVoice: https://github.com/FunAudioLLM/SenseVoice
- Fun-ASR: https://github.com/FunAudioLLM/Fun-ASR
- FunClip: https://github.com/modelscope/FunClip