MahmoudAshraf97/whisper-diarization

Additional punctuation support

hayata-yamamoto opened this issue · 2 comments

hayata-yamamoto commented

Do you have some plan of adding new supporting languages for punctuation? ex.) Japanese, chinese and so on

MahmoudAshraf97 commented

This totally depends on availability of the models, I haven't found any models with more languages supported other than the one used.
A brute force solution for this is to use an LLM to restore the punctuation, I tested this with large models but trying small models might be worth it

  • 👍1
hayata-yamamoto commented

OK, it's resolved, thanks