FunASR
FunASR copied to clipboard
Which romanization Persian model use
Hello
Persian model https://modelscope.cn/models/iic/speech_UniASR_asr_2pass-fa-16k-common-vocab1257-pytorch-online/ version 1.0.1 uses some kind of roman letters in the output. It is not clear how to turn them back to unicode. Please suggest.
Older version 1.0.0 uses unicode but not compatible with latest funasr.