++model=SenseVoiceSmall在modelscope-funasr这么写参数为啥会报错?The request model: SenseVoiceSmall does not exist
详细如下:
zhangzk@zhangzk:~$ funasr ++model=SenseVoiceSmall ++vad_model=fsmn-vad ++punc_model=ct-punc ++input=./mp3/1-0.mp3
You are using the latest version of funasr-1.1.0
[2024-07-15 11:07:08,419][root][INFO] - download models from model hub: ms
2024-07-15 11:07:08,797 - modelscope - ERROR - The request model: SenseVoiceSmall does not exist!
Download: SenseVoiceSmall failed!: The request model: SenseVoiceSmall does not exist!
Error executing job with overrides: ['++model=SenseVoiceSmall', '++vad_model=fsmn-vad', '++punc_model=ct-punc', '++input=./mp3/1-0.mp3']
有执行如下两个来升级:
pip install -U funasr
pip3 install -U modelscope huggingface_hub
当前环境如下命令可以执行成功:
funasr ++model=paraformer-zh ++vad_model=fsmn-vad ++punc_model=ct-punc ++input=./mp3/1-0.mp3
你是不是遗漏了,iic/,另外,sensevoice自带标点,不要再额外用punc模型。此回答整理自钉群“modelscope-funasr社区交流”