Open brucel11qwe opened 1 month ago
@brucel11qwe but would love a contribution for this model!
did you try using Chinese as language in the settings?
Yes, I have configured the settings and I'm using the openai/whisper-large-v3-turbo model. However, Whisper's support for Chinese language is quite poor, with almost all recognition results being incorrect. I suggest you try https://github.com/FunAudioLLM/SenseVoice - its performance is about 100 times better than Whisper for Chinese speech recognition
Yes, I have configured the settings and I'm using the openai/whisper-large-v3-turbo model. However, Whisper's support for Chinese language is quite poor, with almost all recognition results being incorrect. I suggest you try https://github.com/FunAudioLLM/SenseVoice - its performance is about 100 times better than Whisper for Chinese speech recognition
can you send a PR for it?
you can use ONNX runtime that they support - check example we use for VAD to use ONNX in rust:
https://github.com/thewh1teagle/vad-rs/tree/main
🙏
"The effect of Whisper in Chinese speech recognition is very poor, with almost all recognitions being incorrect. I hope to add support for sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17." This translation conveys your concern about the poor performance of Whisper for Chinese speech recognition and your desire to incorporate a specific model (sherpa-onnx-sense-voice-zh-en-ja-ko-yue-2024-07-17) for better results.