Starstruk Logo MarkStarstruk

runtime-llamacpp-v0.1.1

huggingface/smolagentsruntime-llamacpp-v0.1.1Jun 21, 2026by LauraGPT

AI Summary

A lightweight runtime for running SenseVoice and Paraformer locally without Python, torch, or build requirements, featuring on-device ASR with high accuracy.

Key Highlights

  • No Python, build, or torch dependencies
  • Built-in FSMN-VAD for voice activity detection
  • High accuracy on Chinese and Cantonese (8.01% CER)
  • Cross-platform support (Linux, macOS, Windows)

New Features

  • Prebuilt binaries for SenseVoice and Paraformer
  • On-device ASR functionality

Full Release Notes

Prebuilt, self-contained binaries to run **SenseVoice** (and Paraformer / Fun-ASR-Nano) locally with the FunASR llama.cpp / GGUF runtime — built-in FSMN-VAD, whisper.cpp-style on-device ASR, strong on Chinese & Cantonese.

```bash
bash download-funasr-model.sh sensevoice ./gguf
llama-funasr-sensevoice -m ./gguf/SenseVoiceSmall-f16.gguf --vad ./gguf/fsmn-vad.gguf -a audio.wav
```

No Python, no build, no torch. Binaries for Linux (x64/arm64), macOS (arm64), Windows (x64). Docs: `runtime/llama.cpp/README.md`. CER (micro-avg): SenseVoice 8.01% vs whisper.cpp small 22%.