Add FunASR — offline speech recognition for Raspberry Pi

[LauraGPT]

Note

License and capability clarification (2026-07-14): FunASR is a toolkit, not a single checkpoint. The FunASR and SenseVoice repository source code is MIT; model weights follow each model card. SenseVoiceSmall supports Chinese, Cantonese, English, Japanese, and Korean, and its weights use the linked FunASR Model Open Source License Agreement. Fun-ASR-Nano-2512 is Apache-2.0. Language coverage, punctuation, and performance depend on the selected model and runtime configuration.

Hi! Great collection of Raspberry Pi tools.

Would you consider adding FunASR to the speech/audio tools section?

FunASR on Raspberry Pi

FunASR provides ONNX models that run offline on Raspberry Pi:

• Paraformer — High-accuracy ASR, ONNX format, runs on ARM
• SenseVoice-Small — Ultra-fast ASR (~70ms/10s audio on GPU, still fast on CPU)
• FSMN-VAD — Lightweight voice activity detection
• 50+ languages — Including CJK
• Fully offline — No API calls needed
• Apache 2.0

Running on Raspberry Pi

Via ONNX runtime or sherpa-onnx (C++ runtime optimized for embedded):

pip install funasr onnxruntime
# Or use sherpa-onnx for C++ runtime: https://github.com/k2-fsa/sherpa-onnx

References

• FunASR: https://github.com/modelscope/FunASR (16K+ stars)
• ONNX runtime: https://github.com/modelscope/FunASR/tree/main/runtime/onnxruntime
• SenseVoice: https://github.com/FunAudioLLM/SenseVoice (8K+ stars)

[LauraGPT]

Hi! Following up — FunASR supports ONNX runtime and can run efficiently on edge devices including Raspberry Pi. Happy to provide benchmarks or help with integration! 🙏