search

Naozumi520 / Bert-VITS2-Cantonese-Yue

vits2 backbone with multilingual-bert, modified to support Cantonese
GNU Affero General Public License v3.0
5 stars 1 forks source link

text_normalize() remove all alphabet characters #3

Closed indiejoseph closed 6 months ago

indiejoseph commented 6 months ago

https://github.com/Naozumi520/Bert-VITS2-Cantonese-Yue/blob/36790b0f1f8824d91b0c834131eed8b8f1a01637/text/chinese.py#L109

Is it intended?

Naozumi520 commented 6 months ago

https://github.com/Naozumi520/Bert-VITS2-Cantonese-Yue/blob/36790b0f1f8824d91b0c834131eed8b8f1a01637/text/chinese.py#L109

Is it intended?

Yes, a cleaner can only process one language only, and in chinese.py any Chinese punctuation will be replaced with English one. If a sentence contain English words, the inference webui have "Mixed" language which can split the sentence and use different cleaner. But it is necessary to remove the alphabet characters for the Chinese cleaner.

indiejoseph commented 6 months ago

Thanks for clarify, I should dig deeper to know how it works first.