search

modelscope / FunASR

A Fundamental End-to-End Speech Recognition Toolkit and Open Source SOTA Pretrained Models, Supporting Speech Recognition, Voice Activity Detection, Text Post-processing etc.
https://www.funasr.com
Other
6.15k stars 657 forks source link

FunASR版本的Paraformer模型什么什么时候能支持音频中的数字转译？ #2047

Open smengfei opened 1 month ago

smengfei commented 1 month ago

目前SeACoParaformer模型的的pytorch版，转译出的文字中，日期和数字均被转成了大写数字，--itn-dir thuduj12/fst_itn_zh属性仍不支持pytorch版使用