FunASR icon indicating copy to clipboard operation
FunASR copied to clipboard

paraformer-en识别结果存在问题

Open yxzzhang opened this issue 1 year ago • 5 comments

Notice: In order to resolve issues more efficiently, please raise issue following the template. (注意:为了更加高效率解决您遇到的问题,请按照模板提问,补充细节)

🐛 Bug

paraformer-en识别结果完全不对,可能是词表存在问题。如下图所示,识别结果中出现de等非英文字符。 image

已按如下步骤更新过funasr,还是存在这一问题。 image

若是词表存在问题,tokens.json好像是在模型文件中,是否应该更新paraformer-en模型的版本?

To Reproduce

Steps to reproduce the behavior (always include the command you ran): image

  1. Run python test_paraformer.py

Environment

  • OS (Linux):
  • FunASR Version (1.1.2):
  • paraformer-en Version (v2.0.4):
  • ModelScope Version (1.12.0):
  • PyTorch Version (2.3.0):
  • How you installed funasr (source):
  • Python version: 3.10.13
  • GPU (A800)
  • CUDA/cuDNN version (cuda12.1):

yxzzhang avatar Jul 22 '24 09:07 yxzzhang