zhifu gao
zhifu gao
You could delete the `spk_model=os.path.join(model_dir, 'speech_campplus_sv_zh-cn_16k-common'),` and try it again.
We would fix it soon.
I have tested the 16k model: ```text 'text': '中华人民共和国，你们一直莫名其妙的八七七四不明确电话卡销售踢皮球啥状态啊的诈骗注销揭阳市两篇核验完参与参与一点就是点进去以后，这个新庄新南新村辛庄西南新村。 ``` We suggest you to use the 16k model.
How you install it?
`pip install -U funasr modelscope`
Please offer details to reproduce, only use the code of demo.
uniasr模型不在维护，如果想使用，请降低版本： ```shell funasr-0.8.8 modelscope-1.10.0 ``` 同时模型，删除，让它自动下载，不指定任何model_revision
@@ means the token is subword. You could concat them via: replace('@@ ', '')
Please install modelscope-1.10.0
You could find it in v0.8.8 branch: https://github.com/alibaba-damo-academy/FunASR/tree/v0.8.8/egs/callhome/TOLD